Miscitation Register
This register is rendered directly from the references[].reference_review blocks curated in the gene-review YAML (see the Reference schema class). It is the structured, queryable counterpart to the worked cases on the parent project page.
19746 adjudicated reference(s) across 2459 of 5049 reviewed gene files.
- Flagged as a citation problem (WRONG_IDENTIFIER / MISCITED / DISPUTED / LOW_QUALITY): 815 (4.1%)
- Not yet checked (UNVERIFIED): 805
By correctness
| Correctness | Count | Share |
|---|---|---|
| WRONG_IDENTIFIER | 36 | 0.2% |
| MISCITED | 325 | 1.6% |
| DISPUTED | 282 | 1.4% |
| LOW_QUALITY | 172 | 0.9% |
| UNVERIFIED | 805 | 4.1% |
| VERIFIED | 18121 | 91.8% |
By relevance
| Relevance | Count |
|---|---|
| HIGH | 10003 |
| MEDIUM | 5523 |
| LOW | 4137 |
| NONE | 81 |
By organism
Organisms with at least one adjudicated reference, ranked by the number of flagged citation problems.
| Organism | Adjudicated | Flagged | WRONG_IDENTIFIER | MISCITED | DISPUTED | LOW_QUALITY | UNVERIFIED |
|---|---|---|---|---|---|---|---|
| human | 15760 | 627 | 26 | 252 | 210 | 139 | 465 |
| PSEPK | 926 | 49 | 0 | 27 | 5 | 17 | 157 |
| yeast | 530 | 24 | 4 | 4 | 13 | 3 | 49 |
| DICDI | 458 | 19 | 1 | 9 | 9 | 0 | 17 |
| SCHPO | 207 | 14 | 1 | 3 | 9 | 1 | 6 |
| worm | 398 | 9 | 0 | 0 | 6 | 3 | 11 |
| mouse | 61 | 9 | 0 | 8 | 1 | 0 | 16 |
| DROME | 197 | 8 | 1 | 5 | 2 | 0 | 19 |
| ARATH | 347 | 7 | 2 | 1 | 4 | 0 | 17 |
| HETGA | 123 | 7 | 0 | 1 | 1 | 5 | 1 |
| rat | 28 | 7 | 0 | 4 | 2 | 1 | 0 |
| POPTR | 47 | 4 | 0 | 0 | 4 | 0 | 0 |
| NICAT | 9 | 4 | 0 | 0 | 4 | 0 | 2 |
| XENTR | 15 | 3 | 0 | 1 | 0 | 2 | 0 |
| DANRE | 16 | 2 | 0 | 0 | 2 | 0 | 0 |
| DESVH | 10 | 2 | 0 | 1 | 1 | 0 | 1 |
| SPHPI | 6 | 2 | 0 | 2 | 0 | 0 | 0 |
| 9POAL | 5 | 2 | 0 | 0 | 2 | 0 | 0 |
| NOVAD | 5 | 2 | 0 | 2 | 0 | 0 | 0 |
| PYROR | 4 | 2 | 1 | 0 | 1 | 0 | 0 |
| ECOLI | 116 | 1 | 0 | 1 | 0 | 0 | 9 |
| EMENI | 54 | 1 | 0 | 1 | 0 | 0 | 0 |
| SACEN | 36 | 1 | 0 | 0 | 1 | 0 | 13 |
| WHEAT | 16 | 1 | 0 | 0 | 1 | 0 | 0 |
| NEUCR | 13 | 1 | 0 | 1 | 0 | 0 | 0 |
| BACSU | 11 | 1 | 0 | 0 | 1 | 0 | 0 |
| BRUMA | 8 | 1 | 0 | 0 | 0 | 1 | 0 |
| MAIZE | 7 | 1 | 0 | 1 | 0 | 0 | 0 |
| ANOGA | 6 | 1 | 0 | 0 | 1 | 0 | 0 |
| CANAL | 6 | 1 | 0 | 0 | 1 | 0 | 0 |
| DAPPU | 5 | 1 | 0 | 0 | 1 | 0 | 1 |
| MYCMD | 2 | 1 | 0 | 1 | 0 | 0 | 0 |
| DESRO | 25 | 0 | 0 | 0 | 0 | 0 | 8 |
| NOSS1 | 25 | 0 | 0 | 0 | 0 | 0 | 0 |
| NOSP7 | 20 | 0 | 0 | 0 | 0 | 0 | 0 |
| FELCA | 19 | 0 | 0 | 0 | 0 | 0 | 0 |
| FISTH | 19 | 0 | 0 | 0 | 0 | 0 | 0 |
| TRIV2 | 19 | 0 | 0 | 0 | 0 | 0 | 0 |
| CYLST | 17 | 0 | 0 | 0 | 0 | 0 | 0 |
| CAEBR | 16 | 0 | 0 | 0 | 0 | 0 | 2 |
| ECOLX | 14 | 0 | 0 | 0 | 0 | 0 | 2 |
| PSEAE | 12 | 0 | 0 | 0 | 0 | 0 | 0 |
| ECO8N | 10 | 0 | 0 | 0 | 0 | 0 | 0 |
| HORSE | 10 | 0 | 0 | 0 | 0 | 0 | 0 |
| ORYSJ | 8 | 0 | 0 | 0 | 0 | 0 | 0 |
| DERPT | 7 | 0 | 0 | 0 | 0 | 0 | 0 |
| STRPY | 7 | 0 | 0 | 0 | 0 | 0 | 0 |
| CANLF | 6 | 0 | 0 | 0 | 0 | 0 | 0 |
| AQUCT | 5 | 0 | 0 | 0 | 0 | 0 | 1 |
| FUSNU | 5 | 0 | 0 | 0 | 0 | 0 | 0 |
| STAAU | 5 | 0 | 0 | 0 | 0 | 0 | 0 |
| THLAR | 5 | 0 | 0 | 0 | 0 | 0 | 0 |
| 9BACI | 4 | 0 | 0 | 0 | 0 | 0 | 0 |
| METEA | 4 | 0 | 0 | 0 | 0 | 0 | 2 |
| PANTR | 4 | 0 | 0 | 0 | 0 | 0 | 0 |
| SORBI | 4 | 0 | 0 | 0 | 0 | 0 | 0 |
| SOYBN | 4 | 0 | 0 | 0 | 0 | 0 | 0 |
| STRCO | 4 | 0 | 0 | 0 | 0 | 0 | 0 |
| AILME | 3 | 0 | 0 | 0 | 0 | 0 | 0 |
| ASPRC | 3 | 0 | 0 | 0 | 0 | 0 | 0 |
| BETPN | 3 | 0 | 0 | 0 | 0 | 0 | 0 |
| OCTVU | 3 | 0 | 0 | 0 | 0 | 0 | 2 |
| PARTE | 3 | 0 | 0 | 0 | 0 | 0 | 2 |
| 9AVES | 2 | 0 | 0 | 0 | 0 | 0 | 0 |
| CHLRE | 2 | 0 | 0 | 0 | 0 | 0 | 0 |
| DOROP | 2 | 0 | 0 | 0 | 0 | 0 | 2 |
| ENTCL | 2 | 0 | 0 | 0 | 0 | 0 | 0 |
| HYPJE | 2 | 0 | 0 | 0 | 0 | 0 | 0 |
| AEDAE | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| BOVIN | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| CAEEL | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| COLLI | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| METAC | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| PICST | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| PSEAI | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| STAAT | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| STECR | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| XENLA | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
| XENNA | 1 | 0 | 0 | 0 | 0 | 0 | 0 |
Flagged references
Every reference whose correctness is not VERIFIED and not UNVERIFIED, i.e. every reference a reviewer has positively judged to be a citation or soundness problem. Notes are truncated at 400 characters; the full text is in the source YAML and in reports/miscitations.tsv.
WRONG_IDENTIFIER (36)
| Organism | Gene | Reference | Title | Rel. | Notes |
|---|---|---|---|---|---|
| ARATH | GL1 | PMID:3793867 | Cytotoxic and enterotoxic activities of Campylobacter jejuni are not specified by tetracycline resistance plasmids pMAK… | NONE | The cached abstract explicitly assays Campylobacter plasmids in Vero and Chinese hamster ovary cells. The intended Arabidopsis paper is unknown; biological assessment remains UNDECIDED. |
| ARATH | WIP1 | PMID:20579133 | A laboratory evaluation of the physical and mechanical properties of selected root canal sealers. | NONE | Resolves (PubMed record cached) to 'A laboratory evaluation of the physical and mechanical properties of selected root canal sealers' (Int Endod J 2010), unrelated to WIP1. Live QuickGO (2026-09-27) still lists the IntAct WIP1-RANGAP1 IPI with this PMID. IntAct PSICQUIC shows the source record IMEx IM-19345 labelled 'Xu et al. (2007)', which suggests the WIP discovery paper (PMID:17600715, Xu, Me… |
| DICDI | gbpC | PMID:18673369 | The Legionella pneumophila phosphatidylinositol-4 phosphate-binding type IV substrate SidC recruits endoplasmic reticul… | NONE | This PMID is a Legionella pneumophila SidC effector study with no relation to GbpC or cGMP binding. It is cited in GOA as the reference for GbpC cGMP binding (IDA) and regulation of chemotaxis (IMP); those functions are genuine but the identifier is mis-attributed. The correct GbpC references are PMID:12011437, PMID:18703517 and PMID:15827084. |
| DROME | insc | PMID:10973066 | Genetics of heart development. | NONE | Cited (TAS) for insc sensory organ development, but this is a review of cardiac/heart development in zebrafish and Drosophila; the abstract does not mention inscuteable or sensory organs. Appears to be a wrong/mis-applied citation for this annotation; the sensory-organ-development role itself is real but should be sourced to a primary pI/SOP study (e.g. PMID:12526793). |
| PYROR | PoMZ_10221 | file:PYROR/PoMZ_10221/PoMZ_10221-deep-research-openai.md | Deep research on PoMZ_10221 function | NONE | Report describes a mitochondrial dihydroorotate dehydrogenase/PoPYR4, whereas F8U970 is the AEH41994.1 UDP-rhamnose epimerase/reductase characterized in PMID:22102281. Do not use its pyrimidine or localization claims. |
| SCHPO | sel0 | PMID:9700395 | Education and research: where are we going? | NONE | This PMID resolves to a 1998 Australian Veterinary Journal editorial ("Education and research: where are we going?", Niethe G) that has no connection to SelO, AMPylation, mitochondria or S. pombe. It appears in GOA as the reference for one ISO AMPylase-activity annotation (WITH/FROM UniProtKB:P31040, human SDHA). It is a spurious/erroneous identifier in the GOA source and cannot support any annot… |
| human | ACTB | PMID:17340523 | Separation and Quantification of Some Alkaloids from Fumaria parviflora by Capillary Isotachophoresis1. | NONE | PubMed/primary page identity checked in this audit. The local cache is abstract-only. The PMID and title resolve to the unrelated Fumaria alkaloid paper, not a transcription experiment. Original annotation source retained; intended identifier unknown, so no replacement is guessed. |
| human | ADPRH | PMID:12464675 | Restoration of LDL receptor function in cells from patients with autosomal recessive hypercholesterolemia by retroviral… | NONE | A gene-symbol collision, not a finding about this gene. The "ARH1" of this paper is the autosomal recessive hypercholesterolaemia adaptor protein LDLRAP1, which the paper maps to chromosome 1p36; ADPRH is at 3q22.1 and has no role in LDL receptor internalisation. The affinage record lists it as an ADPRH finding, "establishing ARH1 as a functional adaptor required for LDL receptor-dependent LDL in… |
| human | APOO | PMID:26217776 | Mass spectrometry analysis of K63-ubiquitinated targets in response to oxidative stress. | NONE | Recorded here only to document a citation defect in the machine-generated deep research record, and used to support nothing. The affinage report cites PMID:26217776 twice for MIC26 crista-junction and MIC26/MIC27-antagonism findings and labels it Biochimica et biophysica acta; the identifier in fact resolves to a Data Brief paper on K63-ubiquitinated targets in oxidative stress, with no relation… |
| human | BCL2 | PMID:36599 | [Isolation of a strain of Streptococcus pneumoniae multiresistant to antibiotics]. | NONE | Fetched identifier resolves to a 1979 Streptococcus pneumoniae antibiotic-resistance publication. Preserve original GOA references; no verified canonical replacement was found. |
| human | BRIP1 | PMID:14504288 | Cadmium induces nuclear export of Bach1, a transcriptional repressor of heme oxygenase-1 gene. | NONE | This paper concerns the UNRELATED bZIP/BTB transcription factor BACH1 (BTB and CNC homology 1; UniProt O14867), a transcriptional repressor of heme oxygenase-1, NOT the BRIP1/FANCJ helicase (Q9BX63). The two genes share the historical "BACH1" alias. GOA annotations sourced from this PMID (nucleus, cytoplasm, and regulation of transcription by RNA polymerase II) are mis-mapped; the transcription-r… |
| human | ELOVL1 | PMID:10970790 | Cloning of a human cDNA encoding a novel enzyme involved in the elongation of long-chain polyunsaturated fatty acids. | LOW | The cached abstract describes cloning of "HELO1", a 299-aa human elongase on chromosome 6 that elongates long-chain POLYUNSATURATED fatty acids (gamma-linolenic to dihomo-gamma-linolenic, etc.) - i.e. ELOVL5, not ELOVL1 (279 aa, chromosome 1, saturated/monounsaturated VLCFAs). The Reactome GO:0009922 EXP annotation to ELOVL1 that cites this PMID appears to rest on a mis-attributed reference; the… |
| human | ELOVL3 | PMID:10970790 | Cloning of a human cDNA encoding a novel enzyme involved in the elongation of long-chain polyunsaturated fatty acids. | LOW | The abstract describes cloning of HELO1 (=ELOVL5), a paralog on chromosome 6 that elongates long-chain PUFAs (C18:3->C20:3, the ELOVL5 signature), not ELOVL3 (chromosome 10). It is used as a Reactome EXP source for GO:0009922 on ELOVL3; the GO term (fatty acid elongase activity) is correct for ELOVL3 but this specific paper is a likely wrong-paper attribution. ELOVL3 elongase activity is independ… |
| human | GMPS | PMID:8081953 | Enhanced in vitro growth of glomerular cells derived from rats with immune-mediated mesangial injury. | NONE | Cited by the MGI GO:0003921 (GMP synthase activity) IDA annotation, but this PMID resolves to a rat glomerular mesangial-cell growth paper (Harendza et al. 1993, Exp Nephrol) with no relation to GMP synthase. Wrong/transposed identifier; the intended reference is almost certainly PMID:8089153 (human GMP synthetase cloning). The annotated function is correct and independently supported, so the ann… |
| human | HADHA | PMID:22586271 | Acyl coenzyme A thioesterase Them5/Acot15 is involved in cardiolipin remodeling and fatty liver development. | NONE | Full text available and confirms this paper is entirely about Them5/Acot15 with no mention of HADHA. The GO:0052816 (long-chain fatty acyl-CoA hydrolase) IDA on HADHA citing this paper is a wrong-paper attribution; annotation marked for REMOVE. |
| human | KDSR | PMID:15364918 | Lateral diffusion of inositol 1,4,5-trisphosphate receptor type 1 is regulated by actin filaments and 4.1N in neuronal… | NONE | This PMID is a study of IP3R1 lateral diffusion in neurons and does not concern KDSR, yet it is cited as the MGI IDA source for three KDSR annotations (ER localization, 3-keto-sphinganine metabolic process, and reductase activity). Appears to be a wrong-PMID/mis-attributed citation. The affected annotations are nonetheless biologically correct and are independently supported by PMID:15328338 and… |
| human | LOXL1 | PMID:37602378 | HELZ2: a new, interferon-regulated, human 3'-5' exoribonuclease of the RNB family is expressed from a non-canonical ini… | NONE | PubMed and the full cached article were checked. PMID:37602378 is exclusively a HELZ2 RNA-helicase/exoribonuclease paper and contains no LOXL1 evidence. This is a well-formed but biologically mismatched citation; annotation adjudication should nevertheless respect curator-deference rules and avoid inferring how the mismatch arose. |
| human | MYH9 | PMID:2732579 | [Two sisters with pseudoidiopathic hypoparathyroidism presenting extensive intracranial calcification]. | NONE | PMID:2732579 resolves to a 1989 Japanese case report on pseudoidiopathic hypoparathyroidism with no relation to MYH9. The GOA rows citing it (protein binding with RAB3A, lysosome localization, regulated exocytosis, regulation of plasma membrane repair) exactly match PMID:27325790 (Rab3a/Slp4-a/NMHC-IIA complex), which is also cited by GOA with the same RAB3A partner; the cited id is a truncation… |
| human | NAA10 | PMID:25732826 | An organellar Nα-acetyltransferase, Naa60, acetylates cytosolic N termini of transmembrane proteins and maintains Golgi… | NONE | This PMID resolves to the Aksnes et al. Naa60 study (a different N-terminal acetyltransferase, NAA60); it does not specifically characterize NAA10. Title corrected to verbatim PubMed; the citation supports at most generic NAT-localization background and the underlying annotation is a candidate for removal. |
| human | NAA40 | PMID:25732826 | An organellar Nα-acetyltransferase, Naa60, acetylates cytosolic N termini of transmembrane proteins and maintains Golgi… | NONE | This PMID resolves to the Aksnes et al. Naa60 study (a different NAT, NAA60); it does not characterize NAA40. Title corrected to verbatim PubMed; weak background only and a candidate for removal. |
| human | NDUFA8 | PMID:23209302 | KIF14 negatively regulates Rap1a-Radil signaling during breast cancer progression. | NONE | Full text read; this paper concerns KIF14/Radil/Rap1a signaling in breast cancer and does not mention NDUFA8 or mitochondrial Complex I. It is the citation behind the GO:0044877 (protein-containing complex binding) IDA for NDUFA8 but does not support it and appears to be a wrong-paper citation. The annotation was marked as over-annotated (not removed, per policy on experimental annotations). |
| human | NLRP3 | PMID:1189953 | [Profanities and the profane person]. | NONE | This identifier resolves to '[Profanities and the profane person].', Alva Quinones J, Acta Psiquiatr Psicol Am Lat 1975 - a Spanish-language psychiatry abstract with no connection to NLRP3. It is cited in GOA as the IDA support for GO:0060090 and GO:0030674 and is a digit-dropped form of PMID:31189953 (Sharif et al., Nature 2019, NEK7-licensed activation of the NLRP3 inflammasome), which supports… |
| human | NPLOC4 | PMID:39329031 | Study of Clinical Characteristics of Intellectual Disability in Morocco. | NONE | PMID:39329031 is a clinical study of intellectual disability in Morocco and does not concern the VCP-NPL4-UFD1 complex. It is mis-attached to ComplexPortal complex/process annotations. The underlying complex membership is correct, but this citation is a wrong identifier and should be replaced. |
| human | NT5C1A | PMID:599155 | Regulation of the fatty acid composition of alkyl ether phospholipid in Ehrlich ascites tumor cells. The substrate spec… | NONE | PMID:599155 (Waku & Nakazawa 1977, J Biochem) is an Ehrlich-ascites acyltransferase paper unrelated to NT5C1A. It is a transcription error in GOA for PMID:7599155 (Tavenier 1995, "Kinetics of adenylate metabolism in human and rat myocardium"), the paper UniProt actually cites (ECO:0000305|PubMed:7599155) for 5'-nucleotidase activity and cytoplasmic location. Annotations citing PMID:599155 are mar… |
| human | P2RX7 | Reactome:R-HSA-139855 | P2X1-mediated entry of Ca++ from plasma | LOW | This Reactome event is titled "P2X1-mediated entry of Ca++ from plasma" yet is used as a TAS source for a P2RX7 plasma-membrane annotation. The localisation it supports is correct, but the event describes a different P2X subtype. |
| human | SERP1 | PMID:23264731 | MTR120/KIAA1383, a novel microtubule-associated protein, promotes microtubule stability and ensures cytokinesis. | NONE | Paper is entirely about MTR120/KIAA1383, not SERP1/RAMP4; the cytoplasmic-microtubule IDA annotation on SERP1 is a wrong-gene mis-attribution and is recommended for removal. |
| human | SIAH1 | PMID:11863358 | An anthropoid-specific locus of orphan C to U RNA-editing enzymes on chromosome 22. | NONE | The cited paper (APOBEC RNA-editing enzymes) has no relation to SIAH1 - the cached abstract contains zero mention of SIAH1/seven-in-absentia. The zinc ion binding annotation citing this PMID appears to use a wrong identifier; SIAH1 zinc binding is independently supported by PMID:16085652. |
| human | SRPRB | PMID:23264731 | MTR120/KIAA1383, a novel microtubule-associated protein, promotes microtubule stability and ensures cytokinesis. | NONE | Full text (now cached, full_text_available) is entirely about MTR120/KIAA1383 and never mentions or assays SRPRB (no occurrence of SRPRB / SR-beta / SRP receptor). The GO:0005881 cytoplasmic-microtubule IDA citing this paper is a wrong-gene mis-attribution and has been REMOVED, consistent with the sibling SERP1 review. |
| human | STAT2 | PMID:34521819 | Unresolvable identifier - no record in NCBI PubMed (likely a GOA data error) | LOW | Verified via NCBI eutils esummary that PMID:34521819 returns "cannot get document summary" - the identifier does not resolve to any PubMed record and is most likely an error in the GOA source. The entry is retained only because a GOA annotation (existing_annotations[32], a generic uninformative protein binding IPI already marked MARK_AS_OVER_ANNOTATED) references it, so it cannot be deleted witho… |
| human | UFD1 | PMID:39329031 | Study of Clinical Characteristics of Intellectual Disability in Morocco. | NONE | PMID:39329031 is a clinical study of intellectual disability in Morocco and does not concern the VCP-NPL4-UFD1 complex. It is mis-attached to ComplexPortal complex/process annotations. The underlying complex membership is correct, but this citation is a wrong identifier and should be replaced. |
| human | UPF1 | PMID:17469741 | Results of the determination of serum markers in patients with malignant melanoma. | NONE | Cited as a UPF1 protein binding (IPI) source but the resolved title is a melanoma serum-marker study unrelated to UPF1; appears to be an incorrect identifier in the GOA/IntAct record. |
| human | UPF2 | PMID:17469741 | Results of the determination of serum markers in patients with malignant melanoma. | NONE | Cited as a UPF2 protein binding (IPI) source but the resolved title is a melanoma serum-marker study unrelated to UPF2; appears to be an incorrect identifier in the GOA/IntAct record. |
| yeast | COX17 | PMID:8078902 | Isolation of a human cDNA for heme A:farnesyltransferase by functional complementation of a yeast cox10 mutant. | NONE | PubMed-verified: the record is the COX10 (heme A:farnesyltransferase) cloning paper, not a COX17 study. It sources both the GO:0018343 protein farnesylation row and the GO:0005739 row on COX17, assigned by MGI against a S. cerevisiae accession. GO:0018343 is wrong even for COX10, since heme A:farnesyltransferase farnesylates heme rather than protein. Recommend retraction by the assigning group. |
| yeast | TIM10 | PMID:19037698 | The Dutch multicenter experience of the endo-sponge treatment for anastomotic leakage after colorectal surgery. | NONE | The cached full text is an unrelated colorectal-surgery article. The GO term is independently correct for TIM10, but this identifier cannot support it and may be a transposition of PMID:19037098. |
| yeast | TIM10 | PMID:23267104 | Proteome-wide protein interaction measurements of bacterial proteins of unknown function. | NONE | Full-text inspection found an E. coli-only interaction study with no yeast, Tim10, or Tim12 mention; it cannot support the TIM10-TIM12 IPI row. |
| yeast | TIM9 | PMID:19037698 | The Dutch multicenter experience of the endo-sponge treatment for anastomotic leakage after colorectal surgery. | NONE | The cached full text is an unrelated colorectal-surgery paper and cannot support TIM9 IMS localization. The term is independently correct; PMID:19037098 is the plausible transposed identifier. |
MISCITED (325)
| Organism | Gene | Reference | Title | Rel. | Notes |
|---|---|---|---|---|---|
| ARATH | DREB2A | PMID:18266923 | Identification of a novel cis-regulatory element for UV-B-induced transcription in Arabidopsis. | LOW | Abstract concerns the ANAC13 UV-B cis-elements (MRE/UVBox) and does not mention DREB2A; it does not directly support a DREB2A UV-B function. The GO:0010224 IEP annotation citing it is weakly supported (expression-based). |
| DESVH | aprA | file:DESVH/Q72DT2/Q72DT2-hypotheses/function-hypothesis-go-0000104/falcon.md | Existing falcon AprA function-hypothesis report | HIGH | Corroborates APS-specific catalytic chemistry and QmoABC/AprAB interaction evidence. Its proposed GO:0033748 is incorrect: QuickGO resolves that identifier to hydrogenase (acceptor) activity; retain GO:0009973. Reports of fold classification do not by themselves prove a wrong graft node. |
| DICDI | carA | PMID:37866633 | The Legionella autoinducer LAI-1 is delivered by outer membrane vesicles to promote interbacterial and interkingdom sig… | LOW | Full text does not mention cAR1; the GO:0140295 pathogen-derived receptor ligand activity annotation to cAR1 appears to be a curation/mapping error and is recommended for removal. |
| DICDI | chlA | PMID:36252029 | Yellow polyketide pigment suppresses premature hatching in social amoeba. | LOW | Cited for the dibenzofuran metabolic process (GO:0018893) annotation, but the cached full text concerns the PKS5-derived yellow pigment and premature hatching and does not mention ChlA, DIF-1, or dibenzofuran; the citation does not appear to support that annotation. |
| DICDI | cotB | PMID:7713325 | Protein kinase A is a positive regulator of spore coat gene transcription in Dictyostelium. | MEDIUM | Paper is correctly identified but does not support DNA binding by the SP70 product; the DNA-binding activity it describes is a separate trans-acting factor regulating cotB transcription. |
| DICDI | ctnA | PMID:25858552 | Response of Dictyostelium discoideum to UV-C and involvement of poly (ADP-ribose) polymerase. | LOW | ctnA appears only as an expression marker reduced by UV-C; the paper does not show CtnA functions in the UV-C response, so the response to UV-C annotation over-interprets this readout. |
| DICDI | ctnA | PMID:37921687 | Collective signalling drives rapid jumping between cell states. | LOW | ctnA is a cAMP-responsive expression marker here; the paper does not establish CtnA as a participant in the adenylate cyclase-activating cAMP receptor signaling pathway. |
| DICDI | gpaB | PMID:9647646 | G protein beta subunit-null mutants are impaired in phagocytosis and chemotaxis due to inappropriate regulation of the… | MEDIUM | This is a Gβ-null study; it explicitly reports that Gα2-null mutants are NOT impaired in phagocytosis, so it does not support a Gα2 phagocytosis annotation (the phagocytosis phenotype is Gβ-specific). |
| DICDI | pkaR | PMID:14695060 | Surrogate hosts: protozoa and invertebrates as models for studying pathogen-host interactions. | LOW | This review of protozoan/invertebrate surrogate hosts does not document a pkaR sporulation phenotype and does not appear to support the GO:0030435 sporulation annotation it is cited for; likely a mis-attributed reference. |
| DICDI | srfA | PMID:25887420 | Leaps and lulls in the developmental transcriptome of Dictyostelium discoideum. | LOW | Genome-wide developmental transcriptome study that does not specifically discuss srfA in the cached full text. It is cited as the IEP source for a culmination annotation, but srfA cannot be verified from this reference. |
| DICDI | stlB | PMID:36252029 | Yellow polyketide pigment suppresses premature hatching in social amoeba. | LOW | This paper is about PKS5/dictyodenes and does not characterize dibenzofuran metabolism by StlB; it references StlB only for DIF-1 synthesis. It does not support the dibenzofuran metabolic process annotation. |
| DROME | caps | PMID:12508275 | Size isn't everything. | LOW | Abstract-only Bioessays review on tissue size/shape control; the abstract does not mention caps or cell migration. It is the TAS basis for GO:0016477 but is only background/contextual for caps - the real support for a migration role comes from the tracheal study (PMID:16764850). |
| DROME | caps | PMID:12717815 | Pattern formation in the Drosophila wing: The development of the veins. | LOW | Abstract-only Bioessays review on wing-vein patterning; the abstract does not mention caps. Cited as NAS support for GO:0007155 (cell adhesion); the general adhesion function is correct for caps but is not directly supported by this review - a weak/likely-inappropriate NAS citation, adjudicated identically in the trn review, which carries the same GOA row. |
| DROME | hairy | PMID:12593813 | Tube morphogenesis: no pipe dream in Drosophila. | LOW | Current Biology dispatch on tube lumen size and vesicle transport; does not describe a Hairy molecular function. Basis of the over-annotated membrane organization TAS for Hairy. |
| DROME | trn | PMID:12508275 | Size isn't everything. | LOW | Short BioEssays commentary on tissue size/shape and cell affinity; the abstract does not mention trn or cell migration. It is the TAS basis for GO:0016477 but is only background/contextual for trn - the real support for a migration role comes from the tracheal study (PMID:16764850). |
| DROME | trn | PMID:12717815 | Pattern formation in the Drosophila wing: The development of the veins. | LOW | Wing-vein-patterning review whose abstract does not discuss trn or cell adhesion; a weak/likely-inappropriate NAS citation for GO:0007155 cell adhesion, although the adhesion annotation itself is biologically correct and supported elsewhere. |
| ECOLI | DnaJ | PMID:9103205 | Crystal structure of the nucleotide exchange factor GrpE bound to the ATPase domain of the molecular chaperone DnaK. | NONE | The cached record is abstract-only, but its title and abstract explicitly describe a GrpE-DnaK structure and contain no DnaJ. It does not support any of the five CAFA-assigned DnaJ annotations that cite it. |
| EMENI | wA | PMID:7050088 | Purification and characterization of the conidial laccase of Aspergillus nidulans. | MEDIUM | The identifier resolves to the intended paper, so this is MISCITED rather than WRONG_IDENTIFIER - the paper characterizes the conidial LACCASE (the yA activity), not the wA PKS, so it does not support the GO:0052716 annotation made on wA, which is a cross-attribution. |
| HETGA | Cd44 | PMID:37006235 | Galectin-9 promotes natural killer cells activity via interaction with CD44. | LOW | Fetched and read for this review. It is the human source behind the signaling receptor activity and NK T cell activation rows. The identifier resolves to the intended paper and the CD44 finding is real - the galectin-9-CD44 interaction is shown in mouse and then confirmed in human NK cells - but the paper does not support GO:0051132. It studies natural killer cells throughout, and the strings NKT… |
| MAIZE | A0A804UIX9 | file:MAIZE/A0A804UIX9/A0A804UIX9-deep-research-falcon.md | Deep research report for A0A804UIX9 | MEDIUM | The falcon deep research report incorrectly interprets the PGG domain as being characteristic of the PGIP (polygalacturonase-inhibiting protein) family. PGIPs are completely extracellular LRR proteins lacking transmembrane domains, whereas A0A804UIX9 has six predicted transmembrane helices and is classified by PANTHER in the CASKIN family (PTHR24177), not in a PGIP family. The citations about PGI… |
| MYCMD | RCO1 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | LOW | IBA propagated carotenoid dioxygenase activity and carotene catabolic process to Rco1, a resveratrol cleaver experimentally shown to lack carotenoid activity - the same substrate-class over-annotation as cao-1. |
| NEUCR | cao-1 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | LOW | The phylogenetic (IBA) inferences derived carotenoid dioxygenase activity and carotene catabolic process, both of which are experimentally refuted for CAO-1. Useful only as the source of the annotations being corrected here. |
| NOVAD | Saro_0802 | GO_REF:0000118 | TreeGrafter-generated GO annotations | LOW | TreeGrafter propagated carotenoid dioxygenase activity and carotene catabolic process onto NOV1, a stilbene cleavage oxygenase - a substrate-class over-annotation from the mixed CCO family. The InterPro2GO dioxygenase term is fine; the two carotenoid terms are the error. |
| NOVAD | Saro_2809 | GO_REF:0000118 | TreeGrafter-generated GO annotations | LOW | TreeGrafter propagated carotenoid dioxygenase activity and carotene catabolic process to NOV2, a stilbene cleavage oxygenase - the same substrate-class over-annotation seen on NOV1, LSD-III, and cao-1. |
| PSEPK | PP_0375 | file:PSEPK/PP_0375/PP_0375-deep-research-openscientist.md | OpenScientist gene research for PP_0375 (Q88QV9) | MEDIUM | The report retrieves the correct S9-family identity, PQQ-locus context, and Pseudomonas PqqH/PqqM homologs, but it overstates the PA1990 extracellular-PQQ phenotype as proof that synthesis is intact and assigns PP_0375 a distinct excretion role. It also reports exact AlphaFold, STRING, and pairwise-alignment measurements without retaining their underlying inputs or analyses. Only the independentl… |
| PSEPK | PP_0431 | file:PSEPK/PP_0431/PP_0431-deep-research-openscientist.md | OpenScientist research report for PSEPK PP_0431 | HIGH | The report correctly identifies the HemJ-family reaction, membrane architecture, and lack of direct KT2440 enzymology. It overextends ortholog evidence by calling PP_0431 essential and the sole KT2440 protoporphyrinogen oxidase, and it treats oxygen independence, the native electron acceptor, and lateral-transfer history as established for this protein. Those target-specific conclusions were not… |
| PSEPK | PP_1969 | file:PSEPK/PP_1969/PP_1969-deep-research-openscientist.md | OpenScientist gene research for PP_1969 | HIGH | The report is miscited for target-level curation: it declares the divergent Q88LG4 protein a canonical GTP 3',8'-cyclase without a target assay, mislabels PP_1294 (moaE) as moaB and PP_3457 (mobA) as mogA, and presents unsaved motif/residue analysis as verified. None of its catalytic, core-function, locus, or residue claims are propagated; only its explicit statement that no PP_1969-specific bioc… |
| PSEPK | PP_2482 | file:PSEPK/PP_2482/PP_2482-deep-research-openscientist.md | OpenScientist gene research for PP_2482 | HIGH | The report is miscited for target-level curation: it declares Q88K11 the essential, nonredundant canonical GTP 3',8'-cyclase despite the distinct reviewed canonical KT2440 MoaA Q88E69 and the unresolved three-paralog relationship. It also presents unsaved residue, motif, neighborhood, and target-sequence analyses as decisive. None of its catalytic, essentiality, or residue-level claims are propag… |
| PSEPK | PP_2483 | file:PSEPK/PP_2483/PP_2483-deep-research-openscientist.md | OpenScientist gene research for PP_2483 | HIGH | The report is unreliable for target-level curation. It promotes PP_2482 to canonical MoaA despite the curated three-paralog uncertainty, treats indirect neighborhood and orthologous-enzyme evidence as decisive support for dedicated delivery to one KT2440 client, and reports sequence values that differ from the saved local analysis. None of those paralog, client, or client claims are propagated. I… |
| PSEPK | PP_4230 | file:PSEPK/PP_4230/PP_4230-deep-research-openscientist.md | OpenScientist gene research for PP_4230 | HIGH | The report is miscited for target-level curation. It treats a heterologous Cellulosimicrobium CsXodCBA production experiment in KT2440 as experimental evidence for the native PP_4230/PP_4231 locus, misidentifies PP_2482 as the MobA-like member of the PP_2483 cluster, and presents unsaved motif and AlphaFold analyses as target-specific support. None of those client, residue, structure, or locus cl… |
| PSEPK | fcs | file:PSEPK/fcs/fcs-hypotheses/function-hypothesis-go-0031956/falcon.md | Existing falcon Fcs function-hypothesis report | HIGH | Supports hydroxycinnamate activation with gene-context, related-strain mutant and homolog enzymology evidence. Its proposed 4-coumarate-CoA ligase term is substrate-specific rather than a general parent of feruloyl-CoA synthetase; retain exact ferulate term GO:0050563 already present. The text also incorrectly calls benzoate/vanillate non-aromatic; that phrase is not adopted. |
| PSEPK | fcs | file:PSEPK/fcs/fcs-hypotheses/function-hypothesis-go-0031956/openscientist.md | Blinded OpenScientist function-assignment report (TreeGrafter audit) | HIGH | Substrate and exact-strain biochemical conclusion is supported by PMID:12764569, but the report misattributes this paper to Jimenez 2002; correct authors are Plaggenborg et al. 2003. It explicitly states the PANTHER tree was not examined, so precise graft-error assertions remain hypotheses. |
| PSEPK | fliI | file:PSEPK/fliI/fliI-deep-research-openscientist.md | OpenScientist functional report for PSEPK FliI | HIGH | The report correctly states that FliI consumes ATP and is not an ATP synthase. It also repeats EC 7.1.2.2 and calls the historical "flagellum-specific ATP synthase" name consistent, an internal contradiction. Only the explicit ATPase/refutation statement is used. |
| PSEPK | hemA | file:PSEPK/hemA/hemA-deep-research-openscientist.md | OpenScientist research report for PSEPK hemA | HIGH | The report correctly identifies Q88PW6 and recovers direct species-level evidence for the P. putida C5 route. However, it presents an unsaved target-sequence alignment and exact residue mapping as confirmed analysis, attributes a P. aeruginosa complementation and operon result to KT2440, and promotes E. coli/Salmonella regulation and channeling results to the P. putida protein. Only the core reac… |
| PSEPK | hemB | file:PSEPK/hemB/hemB-deep-research-openscientist.md | OpenScientist research report for PSEPK hemB | HIGH | The report recovers the correct reaction and a zinc-site prediction that is independently present in UniProt. It does not account for the KT2440 hemBB paralog, calls a P. aeruginosa protein the close ortholog without resolving paralogy, and presents unsaved global alignments, exact residue maps, and ortholog-derived metal, oligomer, and inhibitor behavior as target-specific conclusions. Only the… |
| PSEPK | hemBB | file:PSEPK/hemBB/hemBB-deep-research-openscientist.md | OpenScientist research report for PSEPK hemBB | HIGH | The report recovers the predicted ALAD reaction and Mg-site architecture, which are independently present in UniProt. It incorrectly calls PP_3322 the single KT2440 hemB/PBGS gene and omits the Q88IT6/PP_2913 paralog, so it cannot adjudicate physiological dominance. It also supplies unsaved sequence analysis and transfers oligomeric state, metal behavior, essentiality, and antibacterial-target cl… |
| PSEPK | hemC | file:PSEPK/hemC/hemC-deep-research-openscientist.md | OpenScientist research report for PSEPK hemC | HIGH | The report correctly identifies Q88RE5 and describes conserved HMBS chemistry that is independently present in UniProt. It also presents an unsaved 66.6% alignment and exact residue map, transfers essentiality from Francisella and Salmonella, infers HemC-HemD channeling from adjacency, generalizes membrane-associated HemH despite the KT2440 HemH record, and gives an unsupported claim that the dip… |
| PSEPK | hemD | file:PSEPK/hemD/hemD-deep-research-openscientist.md | OpenScientist research report for PSEPK hemD | HIGH | The report correctly identifies hydroxymethylbilane cyclization and ring-D inversion. It presents unsaved hydropathy and sequence analyses, infers monomeric and cytoplasmic status, treats gene adjacency as an operon-level physiological argument, and transfers detailed mechanism and essentiality from orthologs. Its pathway sketch also omits the KT2440 HemJ-family PP_0431 step. Those target-specifi… |
| PSEPK | hemE | file:PSEPK/hemE/hemE-deep-research-openscientist.md | OpenScientist research report for PSEPK hemE | HIGH | The report correctly identifies the four-decarboxylation reaction, cytoplasmic assignment, and absence of direct KT2440 experiments. It presents an unsaved human-target alignment and exact residue map, transfers kinetics, substrate breadth, mechanism, dimer behavior, and essentiality from other organisms, and its KT2440 pathway sketch omits the HemJ-family PP_0431 step. Those target-specific conc… |
| PSEPK | hemF | file:PSEPK/hemF/hemF-deep-research-openscientist.md | OpenScientist research report for PSEPK hemF | HIGH | The report correctly identifies RHEA:18257, cytoplasmic localization, and the HemF/HemN oxygen-regime question. It presents unsaved sequence analysis and exact residue mapping, transfers human structure, mechanism, and dimer details, calls the target cofactor-free despite its UniProt-predicted divalent-metal sites, and assumes HemF dominance during aerobic KT2440 growth. Its pathway sketch also o… |
| PSEPK | hemH | file:PSEPK/hemH/hemH-deep-research-openscientist.md | OpenScientist research report for PSEPK hemH | HIGH | The report correctly recovers the ferrochelatase reaction and the lack of direct KT2440 experiments. It presents unsaved target-sequence, alignment, and AlphaFold analyses; proposes an unverified [2Fe-2S] cluster; transfers inner-membrane association, metal specificity, mechanism, and essentiality from orthologs; and omits HemJ from its KT2440 pathway sketch. Those claims were not retained. |
| PSEPK | hemL | file:PSEPK/hemL/hemL-deep-research-openscientist.md | OpenScientist research report for PSEPK hemL | HIGH | The report correctly identifies Q88DP0 and the organism-level C5 route. It also supplies an unsaved 63.4% pairwise alignment and exact target residue mapping, transfers oligomeric and structural details from other HemL proteins, and infers KT2440 essentiality without a target-organism experiment. Those claims were not promoted; the curated review relies on UniProt for the reaction, PLP, homodimer… |
| PSEPK | hemN | file:PSEPK/hemN/hemN-deep-research-openscientist.md | OpenScientist research report for PSEPK hemN | HIGH | The report correctly identifies the radical-SAM coproporphyrinogen-III dehydrogenase reaction and [4Fe-4S] cofactor. It presents unarchived sequence-to-residue mapping, transfers E. coli structure and mechanism, transfers P. aeruginosa Anr/Dnr regulation, and treats the uncharacterized PP_3781 family protein as a redundant oxidase. Its pathway sketch also omits the KT2440 HemJ-family PP_0431 step… |
| PSEPK | moaC | file:PSEPK/moaC/moaC-deep-research-openscientist.md | OpenScientist gene research for moaC | HIGH | The conserved Q88NC0 reaction agrees with the target UniProt entry, but the report mislabels PP_3457 (mobA) as mogA, overstates PP_1292 as experimentally established nonredundant/obligatory, and presents unsaved target alignment, residue, coordinate, and operon analyses as evidence. The exact reaction and explicit lack of a KT2440 assay are retained; the erroneous or unsaved organism-specific cla… |
| PSEPK | mobA | file:PSEPK/mobA/mobA-deep-research-openscientist.md | OpenScientist gene research for mobA | HIGH | The one-MGD reaction agrees with reviewed target UniProt, and the report explicitly acknowledges that Q88HA3 lacks direct target-specific study. However, it assigns moaA to PP_2123 (actually moeA) and a moeA-family gene to PP_1294 (actually moaE), conflates MobA-catalyzed Mo-MPT guanylylation with target-specific bis-MGD assembly and delivery, and presents unsaved target alignment, motif, neighbo… |
| PSEPK | pqqB | file:PSEPK/pqqB/pqqB-deep-research-openscientist.md | OpenScientist gene research for pqqB (Q88QV5) | HIGH | The report correctly retrieves direct support for PqqB as a non-heme hydroxylase and the target-specific structural study. It nevertheless presents a modeled cross-linked diamino-acid substrate, AHQQ product, and PqqF/PqqG-to-PqqB order as established native chemistry. The direct study used substrate analogs, while the QM/MM work and pathway literature do not resolve the native KT2440 substrate,… |
| PSEPK | pqqC | file:PSEPK/pqqC/pqqC-deep-research-openscientist.md | OpenScientist gene research for pqqC (Q88QV6) | HIGH | The report retrieves the correct terminal PqqC reaction class and ortholog structural literature, but it writes the net reaction as AHQQ + O2 producing PQQ + H2O2. The curated Rhea/UniProt reaction consumes three O2 and produces two H2O2, two H2O, and a proton. It also presents ortholog-derived structure and downstream quinoprotein physiology as target-specific conclusions. The report was not use… |
| PSEPK | pqqD1 | file:PSEPK/pqqD1/pqqD1-deep-research-openscientist.md | OpenScientist research report for pqqD1 (Q88QV7, P. putida KT2440) | MEDIUM | The report retrieves conserved PqqD chaperone literature correctly, but assigns pqqE to PP_0378; PP_0378 is pqqC and KT2440 pqqE is PP_0376. The locus error and unsaved alignment claims were not propagated. |
| PSEPK | pqqD2 | file:PSEPK/pqqD2/pqqD2-deep-research-openscientist.md | OpenScientist research report for pqqD2 (Q88JG8, P. putida KT2440) | MEDIUM | The report retrieves the conserved PqqD chaperone literature and correct locus neighborhood, but presents Ped-system specialization as the best interpretation without target-specific expression, genetic, or binding evidence. Its exact alignment and STRING-derived quantitative claims are not retained as reproducible artifacts and were not propagated. |
| PSEPK | pqqF | file:PSEPK/pqqF/pqqF-deep-research-openscientist.md | OpenScientist research report for pqqF (Q88QV3, P. putida KT2440) | MEDIUM | The report correctly retrieves the M16 metallopeptidase family, KT2440 locus, and PQQ-pathway context, but it incorrectly transfers the two-component Methylorubrum M16B PqqF/PqqG architecture to the 766-residue single-chain KT2440 PqqF. It also treats the unrelated S9-family PP_0375 pqqG candidate as though it were the homologous M16B partner. Those target-specific conclusions were rejected. |
| PSEPK | ptxD | file:PSEPK/ptxD/ptxD-deep-research-openscientist.md | OpenScientist retrieval report for Q88HI1 | MEDIUM | Useful for retrieving authentic phosphite-dehydrogenase literature, but its Q88HI1 conclusion transfers PtxD activity from other proteins without direct target evidence and misses the direct PP_3376 kguD genetics. |
| SCHPO | SPBC3H7.04 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | MEDIUM | IBA cytoplasm propagated from PANTHER family PTHR43595, whose characterized members are cytoplasmic/matrix Fe/Mn-SODs (SOD2, SodA). The mitoribosomal mS42 branch is inside the mitochondrion, not in the cytoplasm — the phylogenetic annotation applies to the SOD side of the family and mis-propagates onto the mitoribosomal subfamily. |
| SCHPO | cmp7 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | MEDIUM | The IPR005024 (SNF7 family) → GO:0007034 vacuolar transport InterPro2GO mapping is the source of an inappropriate IEA on cmp7. The mapping is correctly applied at the InterPro family level but is not biologically appropriate for cmp7, which acts at the nuclear envelope, not in vacuolar transport. Upstream tracker geneontology/go-annotation#6458. |
| SCHPO | rad3 | PMID:18180284 | Minichromosome maintenance proteins interact with checkpoint and recombination proteins to promote s-phase genome stabi… | MEDIUM | The complete PMC main article was checked. It supports Rad3-dependent replication-checkpoint phenotypes but contains no Rad3 localization assay and no mention of the nucleolus; it therefore does not support the GO:0005730 IDA annotation. Separately hosted supplemental material remains unverified. |
| SPHPI | Q53353 | GO_REF:0000118 | TreeGrafter-generated GO annotations | LOW | TreeGrafter propagated carotenoid dioxygenase activity and carotene catabolic process onto LSD-I, a genuine lignostilbene dioxygenase - directly contradicting the gene's own experimental (IDA) annotations. |
| SPHPI | lsdB | GO_REF:0000118 | TreeGrafter-generated GO annotations | LOW | TreeGrafter propagated carotenoid dioxygenase activity and carotene catabolic process onto LSD-III, a genuine lignostilbene dioxygenase - directly contradicting the gene's own experimental (IDA) annotations. A clean example of automated carotenoid over-annotation of the stilbene-cleaving clade. |
| XENTR | LOC101732819 | file:XENTR/A0A8J1IYX6/A0A8J1IYX6-deep-research-falcon.md | Deep research summary for LOC101732819 in Xenopus tropicalis | LOW | The research report was generated from the RefSeq SubName and describes mammalian PKCdelta rather than this protein. A0A8J1IYX6 lacks the C1/C2 regulatory domains diagnostic of PKC family kinases (UniProt features show only a disordered N-terminus, one kinase domain, and an AGC-kinase C-terminal domain), and both the PANTHER family assignment (PTHR24351, ribosomal protein S6 kinase) and the IBA W… |
| human | A1BG | file:human/A1BG/A1BG-deep-research-affinage.md | Affinage mechanistic annotation for A1BG (human) | MEDIUM | The entire report was read. Its positive biochemical leads were checked against primary papers; its lack of receptor/GH discussion was previously misused as corroboration of REMOVE. That unsupported negative use is withdrawn. |
| human | AACS | Reactome:R-HSA-77111 | Synthesis of Ketone Bodies | MEDIUM | The Reactome pathway record correctly describes mitochondrial ketogenesis, but it does not support assigning GO:0046951 to cytosolic AACS, which consumes acetoacetate. This is bad parent-pathway propagation; replace it with GO:0046952 ketone body catabolic process. |
| human | AADAC | PMID:11481320 | Characterization of the rodent genes for arylacetamide deacetylase, a putative microsomal lipase, and evidence for tran… | MEDIUM | The abstract is explicitly about mouse and rat genes and proposes, rather than directly measures, a microsomal-lipase role. It is historical support only; later experimental work provides the substrate-level evidence. It cannot directly support the original human TAS lipase tuple, although the annotation-level MODIFY to DAG lipase is justified independently. |
| human | AADAC | PMID:17936933 | Human carboxylesterases and their role in xenobiotic and endobiotic metabolism. | MEDIUM | This is a review article, not a primary experimental report, and therefore cannot itself justify an IDA evidence code for AADAC. It remains relevant background and is cited by UniProt for function; independent primary and activity-based evidence supports AADAC serine-hydrolase activity. |
| human | AARS1 | PMID:7654687 | Human alanyl-tRNA synthetase: conservation in evolution of catalytic core and microhelix recognition. | HIGH | PubMed identity and abstract verified. The study supports human AARS1 tRNA binding, aminoacylation, acceptor-helix recognition, and monomeric behavior of the purified enzyme. Later humanized-yeast co-expression evidence (full-text PMID:40491354) limits generalization to all cellular contexts, without refuting the purified preparation. It does not establish pre-tRNA processing, so its source GO:00… |
| human | ABI3BP | PMID:20551380 | Proteomics characterization of extracellular space components in the human aorta. | MEDIUM | Cached title matches and full text is available, and the paper legitimately supports the two cellular-component rows it is cited for. Marked MISCITED for its third use only: the same dataset is also cited as RCA evidence for the molecular function GO:0005201 extracellular matrix structural constituent, which a fractionation-based protein inventory cannot establish. ABI3BP itself appears in the su… |
| human | ABI3BP | PMID:27559042 | Glycoproteomics Reveals Decorin Peptides With Anti-Myostatin Activity in Human Atrial Fibrillation. | LOW | Cached title matches and full text is available; the paper properly supports the cellular-component row. Marked MISCITED for its second use: it is also cited as RCA evidence for GO:0005201 extracellular matrix structural constituent, and a glycoproteomic inventory whose actual finding concerns decorin peptides cannot establish a molecular function for ABI3BP. ABI3BP does not appear in the body te… |
| human | ABI3BP | PMID:28675934 | Characterization of the Extracellular Matrix of Normal and Diseased Tissues Using Proteomics. | LOW | Cached title matches and full text is available. It legitimately supports the HDA cellular-component row. Marked MISCITED for its use as RCA evidence for the molecular function GO:0005201: a methods and software paper illustrated on two tissue pairs is the weakest of the three citations behind that term and addresses neither ABI3BP nor matrix mechanics. |
| human | ABRA | PMID:20940261 | Costars, a Dictyostelium protein similar to the C-terminal domain of STARS, regulates the actin cytoskeleton and motili… | LOW | Cached title matches and the paper is sound; the problem is the use made of it. The affinage record cites it for the claim that ABRA's actin-cytoskeleton-regulating activity is conserved through the C-terminal Costars domain. The rescue construct is the small standalone Costars protein, whose human counterpart is ABRACL (Q9P1F3, 81 aa) - and ABRACL and ABRA are the only two reviewed human protein… |
| human | ABRACL | PMID:37759737 | Characterization of the Abracl-Expressing Cell Populations in the Embryonic Mammalian Telencephalon. | MEDIUM | Cached title matches and the paper itself is sound - a careful expression and co-labelling study of mouse and cat embryonic forebrain. It is marked MISCITED for the use made of it elsewhere, not for anything it does wrong: UniProt cites it as the evidence for Cell projection, cilium on human ABRACL, and it contains no such observation. It reports somebody else's proteomics result in its Discussio… |
| human | ACAA1 | PMID:18281296 | Contribution of peroxisome-specific isoform of Lon protease in sorting PTS1 proteins to peroxisomes. | LOW | Abstract-only in cache. Cited by GOA as a "protein binding" IPI for ACAA1, but the paper's finding is essentially negative (pLon does not process thiolase); the bare protein-binding annotation is uninformative/over-annotated. |
| human | ACAT1 | PMID:32944968 | Cholesterol 25-Hydroxylase inhibits SARS-CoV-2 and other coronaviruses by depleting membrane cholesterol. | HIGH | The intended paper, human Calu-3 experiments and reported hairpin were verified in full cached text. The paper is miscited for cholesterol O-acyltransferase activity and ER activity of P24752: its cellular assays do not establish either gene-specific assignment. Broad GPP maps the stated hairpin to ACAT1 NM_000019.4/gene 38; this genuine reagent conflict is retained, so the assessment does not as… |
| human | ACBD3 | file:human/ACBD3/ACBD3-deep-research-affinage.md | Affinage mechanistic annotation for ACBD3 (human) | MEDIUM | Gates passed with pairwise = win, and its breadth is genuinely useful - it synthesises 30+ primary papers, though what this review took to be its most valuable contribution - the absence of acyl-CoA - turned out to be its most damaging. But it states that PI4KB is recruited 'through its GOLD domain', and that is wrong: UniProt is explicit that the Q domain (241-308) binds PI4KB and TBC1D22A/B whi… |
| human | ACRV1 | PMID:21252238 | TDP-43 is a transcriptional repressor: the testis-specific mouse acrv1 gene is a TDP-43 target in vivo. | LOW | The identifier and title are correct and the paper is a solid study, but it is the wrong support for ACRV1 involved_in spermatogenesis. It establishes that the mouse acrv1 gene is a TDP-43 target whose transcription is restricted to round spermatids; it assays no property of the ACRV1 protein. Cached copy is abstract-only, but the abstract is unambiguous about the paper's subject. Flagged MISCITE… |
| human | ACTG2 | PMID:10633868 | Antisense oligodeoxynucleotide complementary to smooth muscle alpha-actin inhibits endothelial-mesenchymal transformati… | LOW | Sole source of six ACTG2 ISS rows, via chicken ACTA2 (P08023). The paper is about alpha-smooth-muscle actin in chick atrioventricular cushion endothelial-to-mesenchymal transformation. It is correctly cited for chicken ACTA2 and correctly used for human ACTA2; it is miscited as evidence about human ACTG2, whose chicken orthologue (P63270) was available and unused. Its GO:0010628 IDA is separately… |
| human | ACTG2 | PMID:21294755 | Leptin and acetaldehyde synergistically promotes αSMA expression in hepatic stellate cells by an interleukin 6-dependen… | LOW | Source of rat Actg2's IEP annotations for GO:0071354 and GO:1905641, propagated to human ACTG2. The measured protein is alpha-SMA in hepatic stellate cells, and the reference annotates six entities as a marker panel with GO:1905641 on all six. Miscited as evidence about enteric gamma-actin. Abstract only, so the possible rat-side symbol conflation is raised as a question rather than asserted. |
| human | ACTG2 | PMID:28320086 | Antifibrotic effect of diethylcarbamazine combined with hesperidin against ethanol induced liver fibrosis in rats. | LOW | Source of rat Actg2's IEP annotation for GO:0045471. A pharmacology study whose actin readout is a-SMA in liver. Same conflation concern as PMID:21294755. Abstract only. |
| human | ACTG2 | PMID:8006065 | Actin isoform compartments in chicken gizzard smooth muscle cells. | MEDIUM | Genuinely relevant biology - chicken gizzard is enteric smooth muscle and the antibody used was pan-muscle-actin, so gamma-enteric actin is among what was labelled - but the GO term derived from it is wrong. The paper says labelling was found around the myosin filaments, which cannot support located_in GO:0032982 myosin filament. Basis for the MODIFY to GO:0030485. Abstract only. |
| human | ACTL7A | PMID:10373328 | Cloning, mapping, and expression of two novel actin genes, actin-like-7A (ACTL7A) and actin-like-7B (ACTL7B), from the… | MEDIUM | The paper itself is sound and correctly identified: it is the original cloning, mapping and expression description of ACTL7A and ACTL7B. It is miscited as the basis of the TAS annotations GO:0005200 structural constituent of cytoskeleton and GO:0005856 cytoskeleton. It reports cDNA selection, direct genomic sequencing, linkage mapping in mouse and Northern expression analysis, with no biochemical… |
| human | ACTR10 | Reactome:R-HSA-6798751 | Exocytosis of azurophil granule lumen proteins | LOW | The reaction itself is sound; ACTR10's inclusion in its azurophil granule lumen protein set is not. ACTR10 has no signal peptide or transmembrane segment and no route to a granule lumen; it is also the only one of the 11 canonical dynactin subunits annotated to azurophil granule lumen at all, and the only one carrying any annotation from Reactome's Neutrophil degranulation route. |
| human | ACTR10 | Reactome:R-HSA-6800434 | Exocytosis of ficolin-rich granule lumen proteins | LOW | Same issue as R-HSA-6798751: a proteomics-derived granule protein set that places a cytosolic dynactin subunit in a secretory granule lumen and in the extracellular region. |
| human | ACTR1B | Reactome:R-HSA-6798748 | Exocytosis of secretory granule lumen proteins | LOW | Exocytosis of secretory granule lumen proteins. The reaction is real; ACTR1B's participation in it is not. Membership traces to neutrophil granule-fraction proteomics and places a cytosolic, signal-peptide-less, dynactin-bound protein in a granule lumen and then in the extracellular region. ACTR1A, in the same complex at fifteen times the abundance, is absent from the pathway, which is the decisi… |
| human | ACTR1B | Reactome:R-HSA-6800434 | Exocytosis of ficolin-rich granule lumen proteins | LOW | Exocytosis of ficolin-rich granule lumen proteins; same problem and same reasoning as R-HSA-6798748. |
| human | ACTR5 | PMID:25016522 | The mammalian INO80 chromatin remodeling complex is required for replication stress recovery. | LOW | Full text cached and checked: it contains zero occurrences of "Arp5" or "ACTR5"; the knockdowns were of INO80 and ARP8. ComplexPortal cites it as IMP for two ACTR5 rows via complex-level projection. Marked MISCITED for this gene in the strict sense that the paper does not support an ARP5-specific inference, not because the complex-level claim is wrong. |
| human | ACTR8 | GO_REF:0000107 | Automatic transfer of experimentally verified manual GO annotation data to orthologs using Ensembl Compara | HIGH | The method reference is correctly cited, but its precondition is not met for the five rows that use it here. It promises transfer of experimentally verified manual annotation; the mouse Actr8 source annotations are ComplexPortal projections of a complex-level phenotype onto all 15 INO80 subunits, from an Ino80 conditional knockout in which Actr8 was not manipulated. Verified by QuickGO reference=… |
| human | ACTR8 | PMID:23979016 | The mINO80 chromatin remodeling complex is required for efficient telomere replication and maintenance of genome stabil… | MEDIUM | A sound paper, cited for the wrong subject. Its experiment is an Ino80 conditional knockout, and Actr8 is not named in the abstract; yet via a ComplexPortal projection to all 15 subunits and then Ensembl Compara it is the sole basis for five human ACTR8 rows, including regulation of embryonic development and positive regulation of telomere maintenance in response to DNA damage. MISCITED describes… |
| human | ADAMTSL3 | PMID:18474779 | Antiangiogenic characteristics of astrocytes from optic nerve heads with primary open-angle glaucoma. | LOW | Cited by the affinage deep-research record as showing an antiangiogenic role for ADAMTSL-3. It shows no such thing: ADAMTSL-3 is one of two transcripts raised in glaucoma astrocytes, and the coculture tube-formation assay compares whole astrocyte populations with no ADAMTSL3 gain or loss of function. Correlation presented as function. |
| human | ADGRA1 | PMID:12565841 | There exist at least 30 human G-protein-coupled receptors with long Ser/Thr-rich N-termini. | MEDIUM | Correctly identified, and the source of the durable structural fact that GPR123 alone in the family lacks a GPS domain. MISCITED for the GO:0004930 NAS row that cites it: the paper is a search of human genome databases followed by phylogenetic clustering, with no functional experiment on any of the six receptors it reports. |
| human | ADGRA1 | PMID:15203201 | The human and mouse repertoire of the adhesion family of G-protein-coupled receptors. | LOW | A repertoire catalogue with EST expression charts and no perturbation of any receptor. Querying QuickGO by reference shows it carries 78 annotations across 27 distinct entities - GO:0016020 on 27, GO:0007186 on 26, GO:0004930 on 25, all TAS, all assigned by GDB - i.e. it annotates essentially the entire human adhesion-GPCR family with the same three terms. The terms happen to be defensible for AD… |
| human | ADGRA1 | PMID:17212699 | The evolutionary history and tissue mapping of GPR123: specific CNS expression pattern predominantly in thalamic nuclei… | MEDIUM | Genuine, correctly identified paper, and its expression mapping and its prediction of a C-terminal PDZ-binding motif are used here as sound. MISCITED refers specifically to the two GOA rows that rest on it: GO:0004930 and GO:0007165 by NAS. The paper is an in-situ hybridisation and real-time PCR survey containing no functional or biochemical assay, and its only functional statement is hedged as w… |
| human | ADGRA2 | PMID:15203201 | The human and mouse repertoire of the adhesion family of G-protein-coupled receptors. | LOW | PubMed-verified and a sound paper, but miscited as TAS support for ADGRA2's molecular function. Its abstract describes a genome-mining, phylogenetics and EST-expression survey and reports no functional assay on any receptor; the record is abstract-only, so this review makes no claim about its full text. GDB used it to place one generic triple (GO:0016020, GO:0007186, GO:0004930) on 27 human adhes… |
| human | ADIPOQ | PMID:10095105 | Organization of the gene for gelatin-binding protein (GBP28). | LOW | A gene-structure paper (FISH mapping to 3q27, three exons, no TATA box) cited as TAS for GO:0006091 generation of precursor metabolites and energy. It performs no metabolic experiment. This is its only annotation in all of GOA. |
| human | ADIPOQ | PMID:17327472 | Effects of adiponectin on the renal sympathetic nerve activity and blood pressure in rats. | MEDIUM | Supports GO:0045776 negative regulation of blood pressure directly. It is flagged because it also carries GO:0010906 regulation of glucose metabolic process, which the abstract mentions only as background, and GO:0050805 negative regulation of synaptic transmission, whereas the measurement is renal sympathetic nerve activity. Full text was not available, so the curator's reading of the glucose ro… |
| human | ADIPOQ | PMID:19109165 | The peroxisome proliferator-activated receptor gamma agonist rosiglitazone ameliorates murine lupus by induction of adi… | LOW | The source of mouse Adipoq's GO:0071466 IDA, which reaches human ADIPOQ by ISS and IEA. The paper shows that rosiglitazone INDUCES adiponectin; adiponectin is the output of the xenobiotic response, not a participant in it. |
| human | ADIPOQ | PMID:19855092 | Sialic acid modification of adiponectin is not required for multimerization or secretion but determines half-life in ci… | HIGH | The identifier resolves to the intended paper and the paper is sound. It is recorded as MISCITED because the annotation it supports, GO:0033691 sialic acid binding, inverts the paper's finding: the study characterises adiponectin as the CARRIER of sialylated O-linked glycans, not as a sialic-acid-binding protein. UniProt encodes the same reference correctly, as CARBOHYD features at Thr-21/Thr-22… |
| human | ADIPOQ | PMID:28675934 | Characterization of the Extracellular Matrix of Normal and Diseased Tissues Using Proteomics. | MEDIUM | A legitimate ECM proteomics survey, correctly supporting the GO:0031012 HDA detection row. Flagged because the same reference also drives a bulk RCA block assigning GO:0005201 extracellular matrix structural constituent to 41 entities, a structural claim the proteomics cannot support for a circulating hormone. |
| human | ADIPOQ | PMID:36399478 | MatrisomeDB 2.0: 2023 updates to the ECM-protein knowledge database. | LOW | MatrisomeDB 2.0 is a database-update paper, not a study of adiponectin. In GOA it supplies GO:0140149 to 272 distinct entities, i.e. it is a bulk import rather than an author statement about this gene. |
| human | ADNP | file:human/ADNP/ADNP-deep-research-affinage.md | Affinage mechanistic annotation for ADNP (human) | MEDIUM | The genuine precomputed report was read as a source of leads. Its ADNP-CHD4-BRG1 composition for ChAHP is incorrect; primary evidence defines ADNP-CHD4-HP1. Its broad mechanistic generalizations are not imported without primary support. The report and all machine sources remain unchanged. The source8/9 recovery closes all 21 previously absent provider PMIDs and six broad Reactome citations. Recov… |
| human | ADO | GO_REF:0000033 | Annotation inferences using phylogenetic trees | LOW | PANTHER/PAN-GO phylogenetic (IBA) inference of mitochondrial localization. This contradicts the curated cytosolic localization in UniProt/Reactome and the known cytosolic substrates of ADO; treated as an over-annotation. |
| human | ADO | PMID:34800366 | Quantitative high-confidence human mitochondrial proteome and its dynamics in cellular context. | LOW | High-throughput mitochondrial proteome dataset providing the HTP mitochondrion annotation. ADO is not discussed in the cached full text (listed only in supplementary data); the mitochondrial signal conflicts with the curated cytosolic localization and is most likely co-purification. |
| human | ADPRS | PMID:24191052 | ADP-ribosyl-acceptor hydrolase 3 regulates poly (ADP-ribose) degradation and cell death during oxidative stress. | HIGH | The mouse Arh3 oxidative-stress paper. Correctly cited by MGI, but it is the origin of two term-choice defects that propagate into human ADPRS: the stressor is hydrogen peroxide throughout, yet the annotation made from it is GO:0071451 cellular response to SUPEROXIDE; and the death process it characterises is parthanatos (PARP1 - free PAR - AIF release), which is annotated as regulation of necrop… |
| human | ADPRS | PMID:33769608 | Molecular Tools for the Study of ADP-Ribosylation: A Unified and Versatile Method to Synthesise Native Mono-ADP-Ribosyl… | HIGH | PubMed-verified, full text read. The paper is correctly cited as evidence that ARH3 hydrolyses mono-ADP-ribosyl-serine and -threonine linkages, but GOA uses it to support GO:0004649 poly(ADP-ribose) glycohydrolase activity and the study contains no poly(ADP-ribose) experiment of any kind - only synthetic MARylated Ser/Thr/Cys peptides. Flagged MISCITED for that specific row; the paper itself is s… |
| human | AFF3 | PMID:20444755 | Investigation of rheumatoid arthritis susceptibility genes identifies association of AFF3 and CD226 variants with respo… | LOW | Correctly identified and internally consistent, but it does not support the annotation it is cited for. It is a pharmacogenetic association study - 18 tag SNPs genotyped in RA patients receiving TNF-blocking drugs, with change in DAS28 as the outcome - and it contains no TNF stimulus and no experiment on the AFF3 gene product. PubMed records an ErratumIn reference to Ann Rheum Dis 2011;70(8):1519… |
| human | AGFG1 | PMID:16765935 | Changes in intranuclear chromatin architecture induce bipolar nuclear localization of histone variant H1T2 in male hapl… | MEDIUM | This is the sole reference behind mouse Agfg1's GO:0007289 IMP, from which the human IBA descends, and its result for hrb is a NEGATIVE: H1T2 polar localisation is unaltered in hrb-null spermatids, which the authors use to argue that acrosome development is not required for nuclear polarity. The mouse spermatid nuclear phenotype (nucleopodes) is reported in PMID:14724135 instead, so the term is d… |
| human | AGFG1 | PMID:7634337 | Identification of a novel cellular cofactor for the Rev/Rex class of retroviral regulatory proteins. | HIGH | A sound paper cited for two things it does not establish. Its own abstract has Rev binding the RNA and Rab/AGFG1 binding Rev's protein activation domain, so it cannot support GO:0003723 RNA binding; and the RNA it concerns is RRE-containing viral RNA, not host mRNA, so GO:0006406 mRNA export from nucleus reads more broadly than the evidence. The nucleus localisation it also carries is correct. Th… |
| human | AGFG1 | PMID:7637788 | A human nucleoporin-like protein that specifically interacts with HIV Rev. | HIGH | The abstract's nucleoporin claim is explicitly one of sequence homology, yet it is the sole support for GO:0005643 nuclear pore, a term whose definition asserts membership of the nuclear pore complex. A primary paper states outright that it is not certain whether AGFG1 is an authentic NPC constituent, and UniProt, curating this same paper, records Nucleus rather than nuclear envelope or nuclear p… |
| human | AGGF1 | PMID:15905966 | Biomedicine and diseases: the Klippel-Trenaunay syndrome, vascular anomalies and vascular morphogenesis. | LOW | The identifier and title are right and the paper is real, but it does not support the annotation it was cited for. GO:0001570 is vasculogenesis; this review discusses angiogenesis and restates a primary paper that assayed angiogenesis. The term survives on zebrafish knockdown data published seven years later (PMID:23197652), which this reference could not have cited. |
| human | AGT | PMID:10406457 | Analysis of functional domains of angiotensin II type 2 receptor involved in apoptosis. | MEDIUM | Sound paper, over-extended in GOA. It is an AT2 intracellular-loop mutagenesis study in PC12 cells measuring apoptosis, ERK inhibition and SHP-1 activation, and it properly supports the AGTR2-binding, apoptosis and TRK rows. It cannot support GO:0008083 growth factor activity, which it contradicts, nor GO:0001974 blood vessel remodeling, which it does not measure. Abstract only, but the abstract… |
| human | AGT | PMID:17069818 | Angiotensin II type 1-receptor antagonism prevents type IIA secretory phospholipase A2-dependent lipid peroxidation. | LOW | Correct paper, wrong agent. The LDL modification it reports is the activity of sPLA2-IIA, confirmed by the paper's own inhibitor control; angiotensin II is two steps upstream, inducing the enzyme. Cited on AGT as NAS for LDL particle remodeling. |
| human | AGT | PMID:17416596 | Angiotensin II-induced sudden arrhythmic death and electrical remodeling. | HIGH | An unusually well-justified IGI source: dTGR rats carry both the human renin and the human angiotensinogen transgenes, and neither alone produces the phenotype, which is exactly what IGI encodes. Two of its four rows stand as annotated. A third understates the paper, which determines the direction of the matrix effect, so the neutral parent is replaced by the positive child. The fourth has the si… |
| human | AGT | PMID:18971559 | Angiotensin II upregulates acyl-CoA:cholesterol acyltransferase-1 via the angiotensin II Type 1 receptor in human monoc… | MEDIUM | Good receptor-subtype pharmacology in primary human macrophages, supporting the angiotensin-signalling, cholesterol and foam-cell rows. It cannot support the GO:0005576 localisation row also drawn from it: no localisation experiment on angiotensinogen appears in the study. |
| human | AGT | PMID:21183621 | Protein microarrays discover angiotensinogen and PRKRIP1 as novel targets for autoantibodies in chronic renal disease. | LOW | Cited as an IDA for GO:0003014 renal system process. The cached full text is complete and contains no renal-process experiment: it is an autoantibody biomarker screen, and the authors state they cannot tell whether the antibodies target angiotensinogen or angiotensin I, nor establish pathogenicity. Correct paper, correct title, wrong use. |
| human | AGT | PMID:3397061 | Molecular cloning of the mouse angiotensinogen gene. | MEDIUM | Cited by GOA as a TAS for GO:0004867 serine-type endopeptidase inhibitor activity, which it does not assert. Its own abstract says only that angiotensinogen is homologous to serpins and that the authors aligned a putative reactive centre - and then notes that the human site differs from the rodent one. A sequence-alignment observation in a mouse gene-cloning paper became a molecular function anno… |
| human | AGT | PMID:8513325 | A molecular variant of angiotensinogen associated with preeclampsia. | MEDIUM | A historically important genetic association study, cited on AGT as a TAS for the generic term 'cell-cell signaling', which it contains no experiment bearing on. The variant it identifies is also now understood to act through angiotensinogen concentration rather than function. |
| human | AGTRAP | PMID:10425188 | Increased vasoconstrictor response of the mouse lacking angiotensin II type 2 receptor. | HIGH | Relevant to this review only as the defect it documents. The paper is an Agtr2-knockout study and its abstract never mentions Agtrap, yet MGI attached two annotations on mouse Agtrap to it: GO:0004945 enables by IMP and GO:0008217 by IMP. Querying QuickGO by this reference returns 13 annotations over 3 entities (Agt, Agtr2, Agtrap), with Agtr2 correctly receiving GO:0004945 by IGI; the Agtrap row… |
| human | AHNAK | PMID:22940583 | Structure of an asymmetric ternary protein complex provides insight for membrane interaction. | HIGH | The paper is correctly identified and is a genuine, important AHNAK structure (3.5 A, S100A10/annexin A2/AHNAK ternary complex, ~30 nM). It is miscited for the annotation that uses it: GO:1905686 positive regulation of plasma membrane repair. Membrane repair appears only in the paper's framing sentence; no repair experiment is performed. The row is one of ten identical NAS annotations that Comple… |
| human | AHNAK | PMID:3126079 | The submembranous location of p11 and its interaction with the p36 substrate of pp60 src kinase in situ. | LOW | Osborn et al. 1988 concerns p11 and p36 (S100A10 and annexin A2) and contains no AHNAK experiment - it could not, since the AHNAK gene was first reported in 1992 (PMID:1608957). It supports GO:0005886 for its actual subjects and reaches AHNAK only as a ComplexPortal complex-to-subunit projection across eight entities. The plasma membrane term is correct for AHNAK on other evidence. |
| human | ALB | PMID:16245148 | Production of human serum albumin by sugar starvation induced promoter and rice cell culture. | HIGH | Identifier and original author-uploaded full Methods/Results verified. Mature human ALB coding sequence is expressed behind the starvation-inducible rice alphaAmy3 promoter/signal sequence in cultured rice cells. ALB is the produced output, not the performer of the nutrient-response mechanism. This reference is valid for heterologous production but does not support the assigned human ALB starvati… |
| human | ALDOA | PMID:17329259 | A hydrophobic pocket in the active site of glycolytic aldolase mediates interactions with Wiskott-Aldrich syndrome prot… | MEDIUM | Concerns WASP/actin-pathway binding at the aldolase active site, not tubulin; the tubulin-binding annotation citing this paper is not supported by it. |
| human | ALKBH1 | PMID:16860792 | Neural regeneration protein is a novel chemoattractive and neuronal survival-promoting factor. | MEDIUM | PubMed-verified (Gorba et al., Exp Cell Res 2006). Cited here not as evidence for ALKBH1 function but as the traced source of the mouse chemoattractant-activity annotation that was transferred to human ALKBH1 by Ensembl Compara. The paper is about NRP, a product of a different reading frame of the Alkbh1 locus, so it does not support a chemoattractant molecular function for the ALKBH1 dioxygenase. |
| human | ALPL | PMID:16210410 | Differential expression profiling of membrane proteins by quantitative proteomics in a human mesenchymal stem cell line… | LOW | The paper is sound, but it does not support the GO:0001649 osteoblast differentiation annotation drawn from it. It measures a 27-fold increase in alkaline phosphatase abundance during induced differentiation, which establishes ALPL as a differentiation marker and not as a participant in the differentiation process. The GO:0016020 membrane annotation is supported but far less specific than what is… |
| human | ALPP | PMID:2133555 | Two alkaline phosphatase genes are expressed during early development in the mouse embryo. | LOW | Cited in GOA as an IDA source for human ALPP alkaline phosphatase activity, but the abstract states plainly that the work is in mouse preimplantation embryos and that the method is reverse transcription PCR of transcripts. A transcript survey in another species cannot be direct evidence of the human enzyme's catalytic activity, and the gene it tracks is the mouse embryonic AP rather than a human… |
| human | AMPD1 | PMID:29079593 | Hypoxia modulates the purine salvage pathway and decreases red blood cell and supernatant levels of hypoxanthine during… | LOW | Full-text available. The paper concerns erythrocyte purine salvage/deamination during RBC storage; AMP deaminase acts on AMP (not guanine nucleotides), so the derived "GMP salvage" annotation does not reflect AMPD1's direct biochemistry (marked over-annotated, not removed per policy on experimental annotations). |
| human | AMPD2 | PMID:29079593 | Hypoxia modulates the purine salvage pathway and decreases red blood cell and supernatant levels of hypoxanthine during… | LOW | Study of erythrocyte purine metabolism centered on the RBC AMPD isoform (AMPD3) and the salvage/deamination network; cited for an AMPD2 GMP salvage IDA, but it does not establish AMPD2 in GMP salvage (AMP deaminase makes IMP, not GMP). |
| human | AP1S2 | PMID:23247405 | Cell type-specific Rab32 and Rab38 cooperate with the ubiquitous lysosome biogenesis machinery to synthesize specialize… | LOW | An extra-view commentary, cited by ComplexPortal for five terms on this gene. Two of those, platelet dense granule organization and melanosome assembly, are not supported by what the paper says: its only platelet sentence is an explicit prediction about Rab38, and its AP-1 experiments use pan-AP-1 reagents that cannot distinguish sigma isoforms. Reference projection: 178 annotations across 48 dis… |
| human | AP1S3 | PMID:23247405 | Cell type-specific Rab32 and Rab38 cooperate with the ubiquitous lysosome biogenesis machinery to synthesize specialize… | LOW | A review of Rab32/Rab38-directed lysosome-related organelle biogenesis, used by ComplexPortal as the citation for five annotations on the sigma-1C complex, including melanosome assembly and platelet dense granule organization. The paper is correctly identified and its AP-1-level statement is accurate, but it contains no experiment on any sigma subunit, let alone sigma-1C, so it does not support g… |
| human | AP2S1 | PMID:10753805 | Clathrin coat construction in endocytosis. | LOW | PubMed-verified as a real and correctly-identified review, but it is a poor basis for the three UniProt TAS rows it uniquely supplies to AP2S1 (it is the only reference in GOA that projects to exactly one entity for these terms). The cached record is abstract-only and its abstract concerns the clathrin cage, the mu2 subunit and the alpha-appendage; the sigma subunit is not mentioned. Marked MISCI… |
| human | AP2S1 | Reactome:R-HSA-2130486 | Uncoating of clathrin-coated vesicles and fusion with endosomes | LOW | Reactome reaction 'Uncoating of clathrin-coated vesicles and fusion with endosomes'. Participant-compartment import: it supplies AP2S1 with GO:0036020 endolysosome membrane (7 entities), and QuickGO shows it annotating 20 distinct gene products in total, i.e. every participant of the reaction rather than AP2S1 specifically. This reaction is one of the five that place AP2S1 in the endolysosome mem… |
| human | AP2S1 | Reactome:R-HSA-2130725 | Internalization of MHC II:Ii clathrin coated vesicle | LOW | Reactome reaction 'Internalization of MHC II:Ii clathrin coated vesicle'. Participant-compartment import: it supplies AP2S1 with GO:0005886 plasma membrane (20 entities); GO:0036020 endolysosome membrane (7 entities), and QuickGO shows it annotating 20 distinct gene products in total, i.e. every participant of the reaction rather than AP2S1 specifically. This reaction is one of the five that plac… |
| human | AP2S1 | Reactome:R-HSA-6784729 | PCSK9:LDLR:Clathrin-coated vesicle transport from plasma membrane to endolysosome | LOW | Reactome reaction 'PCSK9:LDLR:Clathrin-coated vesicle transport from plasma membrane to endolysosome'. Participant-compartment import: it supplies AP2S1 with GO:0005886 plasma membrane (8 entities); GO:0036020 endolysosome membrane (9 entities), and QuickGO shows it annotating 9 distinct gene products in total, i.e. every participant of the reaction rather than AP2S1 specifically. This reaction i… |
| human | AP2S1 | Reactome:R-HSA-6784738 | Degradation of PCSK9:LDLR:Clathrin-coated vesicle | LOW | Reactome reaction 'Degradation of PCSK9:LDLR:Clathrin-coated vesicle'. Participant-compartment import: it supplies AP2S1 with GO:0036020 endolysosome membrane (9 entities), and QuickGO shows it annotating 9 distinct gene products in total, i.e. every participant of the reaction rather than AP2S1 specifically. This reaction is one of the five that place AP2S1 in the endolysosome membrane by lettin… |
| human | AP2S1 | Reactome:R-HSA-8855130 | VLDLR:PCSK9:Clathrin-coated vesicle translocates from the plasma membrane to lysosomal membrane | LOW | Reactome reaction 'VLDLR:PCSK9:Clathrin-coated vesicle translocates from the plasma membrane to lysosomal membrane'. Participant-compartment import: it supplies AP2S1 with GO:0005886 plasma membrane (8 entities); GO:0036020 endolysosome membrane (7 entities), and QuickGO shows it annotating 9 distinct gene products in total, i.e. every participant of the reaction rather than AP2S1 specifically. T… |
| human | AP4M1 | Reactome:R-HSA-5229111 | AP4 transports APP from trans-Golgi network to endosome lumen | MEDIUM | The biology in the cached summary is right - mu4 recognizes the APP tyrosine signal and APP moves from the trans-Golgi network to endosomes. Two of the three GO terms derived from it are not: transmembrane protein transporter activity reads the reaction class as a translocase step, and endosome lumen assigns the adaptor the compartment of its cargo. |
| human | AP4S1 | Reactome:R-HSA-5229111 | AP4 transports APP from trans-Golgi network to endosome lumen | MEDIUM | The reaction itself is a reasonable model of APP transport, but as a source of a cellular-component annotation for AP4S1 it is miscited: its output entity is the AP4:APP complex placed in compartment 'endosome lumen', and the Reactome-to-GO mapping hands that compartment to all four AP-4 subunits, which are cytosolic coat proteins. Verified through the Reactome ContentService (it is a BlackBoxEve… |
| human | APOO | PMID:22261194 | Proteomics analysis of cardiac extracellular matrix remodeling in a porcine model of ischemia/reperfusion injury. | LOW | A porcine cardiac extracellular-matrix proteomics study. It does not concern APOO, and APOO is not discussed in it; it entered this gene's record only because pig APOO was among the proteins identified in the ECM fraction, which then became an HDA on F1SPZ9 and an ISS on the human protein. Flagged MISCITED in the sense that the paper does not support the claim made from it: a single detection in… |
| human | APOOL | Reactome:R-HSA-481007 | Exocytosis of platelet alpha granule contents | LOW | The Reactome reaction is a real and correctly described reaction; what is wrong is APOOL's presence in it. Reactome still lists Q6UXV4 among the 27 participants of Platelet degranulation (R-HSA-114608) as entities R-HSA-8863003 and R-HSA-8862979, on the strength of a 2004 platelet releasate proteomics screen (PMID:14630798). The same database now also carries Q6UXV4 as R-HSA-8949612 APOOL (MIC27)… |
| human | ARC | PMID:15537891 | Synaptic targeting by Alzheimer's-related amyloid beta oligomers. | LOW | An amyloid-beta oligomer paper, not an Arc paper: Arc appears as a readout of aberrant synaptic induction. It carries four ARUK-UCL TAS rows on the human gene plus the rat IGI behind GO:0061001. A reference-projection check (QuickGO by reference, paginated) returns 16 annotations over only four entities, so it is gene-specific rather than boilerplate - but its Arc statements are second-hand citat… |
| human | ARC | PMID:23936366 | Conditional knockout of tumor overexpressed gene in mouse neurons affects RNA granule assembly, granule translation, LT… | MEDIUM | The sole gene-level seed behind the GO:0071598 IBD, and the reason that row is downgraded. Full text verified. The paper is about TOG/Ckap5; Arc enters it as the A2RE reporter TRANSCRIPT used to count granules, with Arc protein measured only as nascent translation output near granules and as bulk immunostaining intensity. It is a real Arc observation but it is about Arc mRNA as granule cargo, so… |
| human | ARFGEF1 | PMID:10212200 | Purification and cloning of a brefeldin A-inhibited guanine nucleotide-exchange protein for ADP-ribosylation factors. | MEDIUM | The BIG1/BIG2 cloning paper. Correctly identified and a real primary source for the gene, but the GO:0006887 exocytosis TAS drawn from it is not supported - its statements about secretion are background about the ARF family and about brefeldin A, and its only nucleotide-exchange assay was performed on BIG2. |
| human | ARFGEF2 | PMID:15644318 | BIG1 is a binding partner of myosin IXb and regulates its Rho-GTPase activating protein activity. | MEDIUM | The paper is correctly identified and scientifically sound; it simply does not support a myosin-binding annotation on ARFGEF2. Its subject, its yeast two-hybrid screen and its antibodies are all BIG1/ARFGEF1, and across GOA it placed six annotations on human ARFGEF1 and none on human ARFGEF2; human ARFGEF2's row reaches it only through a rat intermediate. Coded MISCITED rather than WRONG_IDENTIFI… |
| human | ARFGEF2 | PMID:17276987 | The brefeldin A-inhibited guanine nucleotide-exchange protein, BIG2, regulates the constitutive release of TNFR1 exosom… | HIGH | Correctly identified and scientifically sound, but it is cited in GOA for GO:0032760 positive regulation of tumour necrosis factor production, and it measures release of the TNF receptor TNFR1, not production of TNF. Flagged as MISCITED for that specific annotation, not as a defect of the paper. |
| human | ARFGEF3 | PMID:14657013 | BIG-3, a novel WD-40 repeat protein, is expressed in the developing growth plate and accelerates chondrocyte differenti… | NONE | Not this gene. The affinage record presents this paper's chondrocyte differentiation results as ARFGEF3 findings, but its "BIG-3" is BMP-2-induced gene 3 kb, a 34 kDa protein with seven WD-40 repeats (PMID:11551928). ARFGEF3/BIG3 is 2177 residues with Sec7 and armadillo-type repeats and no WD-40 domain. A pure symbol collision. Nothing in this review rests on it; listed so the collision is record… |
| human | ARFGEF3 | PMID:15707593 | The effects of BIG-3 on osteoblast differentiation are not dependent upon endogenously produced BMPs. | NONE | Same symbol collision as PMID:14657013 and by the same authors: the abstract states plainly that BIG-3, BMP-2-induced gene 3 kb, encodes a WD-40 repeat protein. Not ARFGEF3. The Smad1 result the affinage record attributes to this gene belongs to a different protein. |
| human | ARG1 | PMID:28813417 | CMTM6 maintains the expression of PD-L1 and regulates anti-tumour immunity. | LOW | PubMed-verified paper, but it does not support a functional ARG1 annotation - ARG1 is only an MS interactor of CMTM6, and the resulting bare protein-binding term is uninformative. |
| human | ARGLU1 | PMID:25468996 | E-cadherin interactome complexity and robustness resolved by quantitative proteomics. | LOW | The paper is sound and correctly identified; the problem is the use made of it. It reports proteins in the vicinity of the E-cadherin cytoplasmic tail by proximity biotinylation, and that result has been converted into 272 cadherin binding annotations over 272 distinct gene products. For a nuclear protein such as ARGLU1 the inference from proximity to binding is not supported. Full text is cached… |
| human | ARHGAP11B | Reactome:R-HSA-8981637 | RHOA GAPs stimulate RHOA GTPase activity | LOW | The reaction is real and correctly describes RHOA GAPs as a class, and for genuine members such as ARHGAP11A the resulting annotation is right. Marked MISCITED for ARHGAP11B specifically: including this protein in the catalyst set asserts an activity that two independent IDA annotations on the same gene explicitly negate, and that UniProt's recommended protein name denies. The defect is set membe… |
| human | ARHGAP11B | Reactome:R-HSA-9012999 | RHO GTPase cycle | LOW | Same defect as R-HSA-8981637, one level up: ARHGAP11B is listed as a participant in the RHO GTPase cycle on the strength of its RhoGAP domain, and every participant receives GO:0051056. The gene has no measured role in small GTPase signalling. |
| human | ARHGEF18 | Reactome:R-HSA-205039 | p75NTR indirectly activates RAC and Cdc42 via a guanyl-nucleotide exchange factor | LOW | ARHGEF18 is returned among this reaction's participants only as a member of a generic GEFs catalyst set. The reaction's declared substrate is RAC1 and its name additionally invokes Cdc42, which every primary source agrees ARHGEF18 does not act on. Recorded as miscited in the sense that the protein is placed in a reaction it has not been shown to catalyse; the cytosol annotation derived from it is… |
| human | ARHGEF19 | Reactome:R-HSA-205039 | p75NTR indirectly activates RAC and Cdc42 via a guanyl-nucleotide exchange factor | LOW | The reaction itself is fine; using it as a source for ARHGEF19 is not. Its output is RAC1-GTP and human WGEF does not activate Rac1 (PMID:18256687, Fig. 2B), so the protein is attached here by set membership rather than by evidence. The derived GO annotation is only the cytosol compartment, which is why the row is marked over-annotated rather than removed. |
| human | ARSJ | PMID:16174644 | Sulfatases and sulfatase modifying factors: an exclusive and promiscuous relationship. | HIGH | PubMed identity and abstract were checked. The paper supports ARSJ sulfatase-family membership and conserved catalytic architecture, but the TAS use for the narrower GO arylsulfatase activity overstates the paper: no ARSJ aryl-substrate assay is reported in the accessible record. |
| human | AZIN1 | PMID:18508777 | Antizyme inhibitor 2 (AZIN2/ODCp) stimulates polyamine uptake in mammalian cells. | MEDIUM | PubMed-verified; abstract-only. The paper characterizes the paralog AZIN2/ODCp, not AZIN1. The IDA GO:1902269 (positive regulation of polyamine transmembrane transport) annotation on AZIN1 is sourced from an AZIN2 study, so for AZIN1 it is at best an orthology-inferred, non-core role rather than a directly demonstrated AZIN1 function. |
| human | BBS1 | PMID:22228099 | BBS proteins interact genetically with the IFT pathway to influence SHH-related phenotypes. | MEDIUM | Full text available. Demonstrates BBSome-dependent regulation of SMO/PTCH1 ciliary accumulation and reduced Shh response via genetic interaction with IFT. Does NOT show direct BBS1 binding to Patched or Smoothened; the patched-binding/smoothened-binding IPI annotations derived from it over-interpret a trafficking/regulatory readout as a molecular binding function. |
| human | BBS1 | PMID:22302990 | Direct role of Bardet-Biedl syndrome proteins in transcriptional regulation. | LOW | Paper centers on BBS7 nuclear/RNF2 (PcG) function and proposes a similar role for other BBS proteins. The BBS1-RNF2 RNA-Pol-II-TF-binding annotation extrapolates BBS7 findings to BBS1 and is not a core BBS1 function. |
| human | BBS5 | PMID:22302990 | Direct role of Bardet-Biedl syndrome proteins in transcriptional regulation. | LOW | Primary evidence concerns BBS7 interacting with the polycomb protein RNF2/RING2; a transcription-factor binding role for BBS5 is only inferred ("a similar role for other BBS proteins"). The GO:0061629 annotation to BBS5 is weakly supported and conflicts with the consensus cytoplasmic/ciliary localization. |
| human | BCL2 | PMID:12086670 | Bcl2 regulation by the melanocyte master regulator Mitf modulates lineage survival and melanoma cell viability. | HIGH | Correct and informative melanocyte-survival paper, but full-text Figure 3 identifies MITF, not BCL2 protein, in sequence-specific DNA-binding assays. See notes for full-text URL, ChIP and EMSA context. |
| human | BCL2 | PMID:16717086 | PP2A regulates BCL-2 phosphorylation and proteasome-mediated degradation at the endoplasmic reticulum. | HIGH | Full Methods/Results/Figure 3G establish BCL2 as ubiquitination substrate and PP2A-regulated survival effector. The paper is relevant, but cannot support BCL2 performing protein polyubiquitination. |
| human | BTD | PMID:7550325 | Mutational hotspot in the human biotinidase gene causes profound biotinidase deficiency. | MEDIUM | Correctly identifies a common BTD mutation causing profound biotinidase deficiency and the biotin-recycling function, but does NOT support a "central nervous system development" (GO:0007417) role; that annotation over-interprets the disorder's neurological phenotype. Relevant to BTD function/disease overall. |
| human | C1GALT1 | PMID:17228361 | Variants of C1GALT1 gene are associated with the genetic susceptibility to IgA nephropathy. | MEDIUM | The identifier resolves to the intended paper and the paper is sound, but it does not support either of the two annotations drawn from it. It is a genetic association study of susceptibility to an acquired adult glomerular disease, carrying an IMP evidence code, and it demonstrates neither participation in angiogenesis nor participation in kidney development. Flagged here because this is exactly… |
| human | CA8 | PMID:8977131 | Two point mutations convert a catalytically inactive carbonic anhydrase-related protein (CARP) to an active enzyme. | HIGH | PubMed-verified: FEBS Lett 1996;398(2-3):322-5, DOI 10.1016/s0014-5793(96)01263-x; abstract-only in cache. The identifier is correct and the paper is highly relevant, but it is cited by GOA (PINC, TAS) in support of GO:0004089 carbonate dehydratase activity, which is the opposite of what it reports: unmodified CARP is catalytically inactive and is not isolated as a zinc complex, and activity appe… |
| human | CERS2 | PMID:34080016 | Glucosylceramide and galactosylceramide, small glycosphingolipids with significant impact on health and disease. | LOW | General GlcCer/GalCer review; does not mention CERS2. Cited for the glycosphingolipid biosynthetic process (EXP) annotation, which over-reaches CERS2's role (it supplies the VLC-ceramide precursor, not the glycosylation step). |
| human | CERS2 | PMID:36170811 | De novo sphingolipid biosynthesis necessitates detoxification in cancer cells. | LOW | Cached full text is entirely about SPT/KDSR (first two de novo pathway enzymes) and never mentions CERS2 or ceramide synthase; cannot verify the CERS2-specific IDA annotations (GO:0046513, GO:0098554) sourced from it. |
| human | CKB | PMID:22926577 | Quantitative proteomic analysis of human substantia nigra in Alzheimer's disease, Huntington's disease and Multiple scl… | LOW | A disease-state differential-abundance proteomic survey of substantia nigra; it does not demonstrate a role for CKB in substantia nigra development. |
| human | COX6A2 | PMID:9284905 | Assignment of COX6A1 to 6p21 and a pseudogene (COX6A1P) to 1p31.1 by in situ hybridization and somatic cell hybrids. | LOW | Cited (ProtInc/PINC, TAS) for GO:0006091 on COX6A2, but the paper is a chromosomal-mapping study of the paralog COX6A1 and the COX6A1P pseudogene, not COX6A2. Wrong-gene citation for a broad energy-metabolism term. |
| human | CPS1 | PMID:18063578 | The layered structure of human mitochondrial DNA nucleoids. | LOW | Nucleoid co-purification study. CPS1 co-purifies as an abundant matrix protein, but the paper notes several metabolic proteins in native nucleoids do not cross-link to mtDNA; supports co-purification rather than genuine nucleoid localization. |
| human | CPS1 | PMID:20031578 | Novel associations of CPS1, MUT, NOX4, and DPEP1 with plasma homocysteine in a healthy population: a genome-wide evalua… | LOW | GWAS association of a common CPS1 SNP with plasma homocysteine; does not demonstrate that CPS1 acts in homocysteine metabolism or directly binds a modified amino acid. The IDA annotations attributed to it (homocysteine metabolic process; modified amino acid binding) are misapplied to this reference, though modified amino acid binding is independently correct (NAG). |
| human | CPT1A | PMID:21492153 | Analysis of proteomic changes induced upon cellular differentiation of the human intestinal cell line Caco-2. | LOW | CPT1A up-regulation is a correlative observation among lipid-metabolism proteins during Caco-2 differentiation; does not establish that CPT1A drives epithelial cell differentiation (GO:0030855). |
| human | CRISPLD1 | PMID:21254358 | Nonsyndromic cleft lip and palate: CRISPLD genes and the folate gene pathway connection. | MEDIUM | PubMed/PMC identity and full text were checked. This is the correct CRISPLD1 association paper, but it does not provide a CRISPLD1 perturbation phenotype supporting an IMP face-morphogenesis assertion; it reports a marginal SNP result, gene-gene interactions, and a need for functional studies. |
| human | CYP17A1 | PMID:2808364 | Deletion of a phenylalanine in the N-terminal region of human cytochrome P-450(17 alpha) results in partial combined 17… | MEDIUM | Correctly cited for the AH5 variant and (via UniProt) subcellular location, but it is the original_reference for a GO:0019825 "oxygen binding" annotation that the paper does not address; that annotation is marked over-annotated. |
| human | CYP27A1 | PMID:15465040 | Metabolism of vitamin D by human microsomal CYP2R1. | MEDIUM | Cited as support for CYP27A1 vitamin D 25-hydroxylase activity, but the paper's conclusion is that CYP2R1 (not CYP27A1) is the physiological 25-hydroxylase and that CYP27A1 hydroxylates vitamin D2 at C-24/C-27; supports the over-annotation flag rather than a core CYP27A1 function. |
| human | DGAT1 | Reactome:R-HSA-6799350 | Exocytosis of specific granule membrane proteins | LOW | Reactome neutrophil-degranulation pathway; used to place DGAT1 at the plasma membrane and specific granule membrane. These localizations do not reflect the physiological ER-membrane site of DGAT1 catalysis and drive over-annotated CC terms. |
| human | DPAGT1 | PMID:6289658 | Demonstration of the heterozygous state for I-cell disease and pseudo-Hurler polydystrophy by assay of N-acetylglucosam… | LOW | PubMed-verified, but the paper is about the DISTINCT lysosomal-enzyme GlcNAc-1-phosphotransferase (GNPTAB/GNPTG; I-cell disease / mucolipidosis II-III), not DPAGT1. It does not support a DPAGT1 function; used to support the GO:0003976 annotation that has been flagged as over-annotation (wrong enzyme). |
| human | DPM1 | PMID:21630459 | Proteomic characterization of the human sperm nucleus. | LOW | Sperm-nucleus proteome catalogue used to annotate DPM1 to nucleus; the nuclear localization is a likely co-purifying contaminant and contradicts established ER localization (basis for REMOVE). |
| human | DPYD | PMID:10410956 | Radiochemical assay for determination of dihydropyrimidinase activity using reversed-phase high-performance liquid chro… | LOW | Methods paper describing a radiochemical assay for dihydropyrimidinase (DHP, the downstream enzyme), not a direct DPD assay. Poor support for the GO:0017113 IDA it is attached to; process (thymine catabolism) assignment is acceptable. |
| human | DPYD | PMID:18075467 | Genetic regulation of dihydropyrimidinase and its possible implication in altered uracil catabolism. | LOW | Paper is about dihydropyrimidinase (DHP, gene DPYS) and a uracil breath test, not a direct DPD assay. The GO:0017113 IDA on this reference is a mis-fit; the uracil catabolic process assignment for DPD is acceptable. |
| human | DPYSL5 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | MEDIUM | Source of the two domain-based IEA hydrolase annotations (GO:0016787, GO:0016810). The InterPro2GO mapping fires on the metallo-hydrolase fold but DPYSL5 lacks the catalytic metal site; the activity annotations are not supported for this member. |
| human | EN1 | PMID:8094370 | Regional assignment of the human homeobox-containing gene EN1 to chromosome 2q13-q21. | LOW | The paper maps EN1 but does not provide evidence for the associated anatomical-structure-morphogenesis annotation. |
| human | FANCG | PMID:9806548 | The Fanconi anaemia group G gene FANCG is identical with XRCC9. | HIGH | Correct identifier for the FANCG=XRCC9 identification paper, but it is used in GOA to support a 'damaged DNA binding' molecular function that the paper does not claim; cited here only to document that over-annotation. |
| human | FANCI | PMID:18445686 | EML3 is a nuclear microtubule-binding protein required for the correct alignment of chromosomes in metaphase. | LOW | Correctly identified paper, but it is an EML3/mitotic-spindle proteomics study in which FANCI is only an incidental hit; it does not establish a genuine cytoplasmic function for FANCI. Basis for the over-annotated cytoplasm localization. |
| human | FANCI | PMID:19946888 | Defining the membrane proteome of NK cells. | NONE | Bulk NK-cell membrane-fraction proteomics; FANCI has no transmembrane region and no membrane role. The paper itself flags most identified proteins as non-membrane. Basis for the removed membrane annotation. |
| human | FASN | PMID:18455495 | Differential activation of recombinant human acetyl-CoA carboxylases 1 and 2 by citrate. | LOW | Reactome-linked EXP citation for FASN fatty acid synthase activity (GO:0004312). The cached abstract is about acetyl-CoA carboxylase (ACACA/ACACB) activation by citrate, the enzyme immediately upstream of FASN, and does not itself demonstrate FASN activity. Retained (experimental annotation, not removed) but the core FASN activity is better supported by PMID:8962082 and PMID:26851298. |
| human | FEN1 | PMID:19946888 | Defining the membrane proteome of NK cells. | NONE | PubMed-verified as a real NK-cell membrane proteome study, but it does not support membrane localization of FEN1. The authors state that only ~40% of identified proteins were plausible membrane proteins and that the rest were proteins transiently or adventitiously associated with the membrane fraction. Basis for the REMOVE on GO:0016020. |
| human | FEN1 | PMID:8007985 | Structural and functional conservation of the human homolog of the Schizosaccharomyces pombe rad2 gene, which is requir… | LOW | PubMed-verified; cached record is abstract-only (full_text_available: false). The abstract reports cloning of the human homolog of S. pombe rad2 and complementation of the UV sensitivity of a rad2 null mutant, and describes no damaged-DNA-binding experiment. The GO:0003684 REMOVE therefore rests primarily on the positive argument that FEN1 discriminates branch geometry rather than base damage, no… |
| human | FLG | PMID:17704059 | Role of ADAM-9 disintegrin-cysteine-rich domains in human keratinocyte migration. | LOW | Cited by GOA as TAS for FLG involvement in keratinocyte differentiation, but the paper is about ADAM-9 disintegrin-cysteine-rich domains, beta1 integrin engagement and keratinocyte migration. Filaggrin is not its subject and the abstract does not mention it. The annotated term is nonetheless correct for FLG on other evidence; the reference should be replaced with PMID:32165560. |
| human | FTO | PMID:26287746 | FTO Obesity Variant Circuitry and Adipocyte Browning in Humans. | LOW | Excellent paper, wrong gene for these annotations. It dissects the FTO locus and concludes the causal mechanism runs through IRX3 and IRX5; FTO protein is never perturbed. Used in GOA as the IMP source for two FTO biological-process annotations, which conflates locus with gene product. The annotations would be better re-homed to IRX3 and IRX5. |
| human | G6PD | PMID:22926577 | Quantitative proteomic analysis of human substantia nigra in Alzheimer's disease, Huntington's disease and Multiple scl… | LOW | Disease proteomics survey; detection/altered expression does not support a substantia nigra developmental role for G6PD (HEP annotation over-interpreted). |
| human | GALNS | PMID:8325655 | Mucopolysaccharidosis IV A: assignment of the human N-acetylgalactosamine-6-sulfate sulfatase (GALNS) gene to chromosom… | LOW | This is a gene-mapping (FISH/chromosome-assignment) abstract for GALNS. It was cited (TAS) to support GO:0003943 N-acetylgalactosamine-4-sulfatase activity, but GALNS is a 6-sulfatase (EC 3.1.6.4); the 4-sulfatase is ARSB. The paper does not support a 4-sulfatase activity. |
| human | GALT | PMID:20605918 | The Lyn kinase C-lobe mediates Golgi export of Lyn through conformation-dependent ACSL3 association. | NONE | Paper is about Lyn kinase/ACSL3 Golgi trafficking, not GALT. Cited as the basis of a GALT Golgi-apparatus IDA that is almost certainly mis-attributed; that localization is therefore flagged as MARK_AS_OVER_ANNOTATED (retained, not removed, pending full-text confirmation). |
| human | GAMT | PMID:8547310 | Cloning and sequence analysis of human guanidinoacetate N-methyltransferase cDNA. | MEDIUM | Correctly identifies and clones human GAMT cDNA, but is used to support a 'muscle contraction' (GO:0006936) annotation, which is a distal physiological consequence rather than a direct GAMT process. |
| human | GGT1 | PMID:15528406 | Hemoxygenase-2 is an oxygen sensor for a calcium-sensitive potassium channel. | NONE | Paper is about hemoxygenase-2 as an oxygen sensor for the BK (KCNMA1) channel and does not concern GGT1. Cited only as the source of a bare protein-binding IPI (partner Q12791 KCNMA1); does not support a functional GGT1 interaction. |
| human | GLB1 | PMID:11927518 | Endothelial cell senescence in human atherosclerosis: role of telomere in endothelial dysfunction. | LOW | Measures senescence-associated beta-galactosidase (SA-beta-gal) histochemical staining as a senescence marker; does not biochemically characterize GLB1 substrate specificity or localization. The IDA beta-galactosidase-activity, carbohydrate-metabolic-process, and cytoplasm annotations derived from it are over-annotations. |
| human | GLB1 | PMID:15498789 | Role of a novel EGF-like domain-containing gene NGX6 in cell adhesion modulation in nasopharyngeal carcinoma cells. | LOW | Paper concerns NGX6 and cell adhesion in nasopharyngeal carcinoma; does not establish a physiologically meaningful GLB1 protein interaction. Supports only a bare, uninformative protein-binding annotation. |
| human | GLB1 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | High-throughput binary (Y2H) interactome map; the reported GLB1 partners (GOLM1, SLC10A6, SLC30A2, SLC7A1) are likely non-physiological screen hits and support only a bare protein-binding annotation. |
| human | GLDC | PMID:2268343 | One of the two genomic copies of the glycine decarboxylase cDNA has been deleted at a 5' region in a patient with nonke… | LOW | Abstract-only cache; PubMed-verified. The paper describes a genomic deletion of the glycine decarboxylase gene in an NKH patient and does NOT support the electron transfer activity (GO:0009055) it is cited for; that MF annotation is an over-annotation/mis-assignment for the P protein. |
| human | GLOD4 | PMID:25468996 | E-cadherin interactome complexity and robustness resolved by quantitative proteomics. | LOW | Verified against PubMed as Guo Z et al., Sci Signal 2014;7(354):rs7, doi 10.1126/scisignal.2005473. The paper is real and correctly identified, but it does not support a cadherin binding molecular function for GLOD4: it is a BioID proximity-labelling survey that reported 561 proteins near the E-cadherin tail, and proximity within a labelling radius is not binding. Flagged MISCITED for that mismat… |
| human | GLUD1 | PMID:22926577 | Quantitative proteomic analysis of human substantia nigra in Alzheimer's disease, Huntington's disease and Multiple scl… | LOW | Proteomic abundance survey of substantia nigra in neurodegenerative disease. Detection/differential abundance does not support a role in substantia nigra development; the derived GO annotation is an over-interpretation. |
| human | GPAA1 | PMID:12052837 | Structural requirements for the recruitment of Gaa1 into a functional glycosylphosphatidylinositol transamidase complex. | MEDIUM | Concerns ER retention/sorting of GAA1 itself and complex assembly via its luminal domain; does NOT support 'protein retention in ER lumen' (GO:0006621) as a GPAA1 function acting on other proteins - that annotation is a term/paper mismatch. Relevant for localization/assembly. Abstract-only. |
| human | GPI | PMID:7435496 | The first stable variant of erythrocyte glucose-phosphate isomerase associated with severe hemolytic anemia. | MEDIUM | Describes a stable GPI variant causing hemolytic anemia with altered F6P/G6P kinetics; it does not support a role in hemostasis, so the TAS hemostasis annotation appears mis-mapped. |
| human | GRID1 | PMID:27276689 | mGlu1 receptor canonical signaling pathway contributes to the opening of the orphan GluD2 receptor. | MEDIUM | Neuropharmacology 2017. The paper is sound and the observation (Gq-PLC-PKC opening of a GluD channel) is real, but its direct recordings are of GluD2, and it has been used in GOA to give GRID1 an enables annotation to G protein-coupled receptor activity (GO:0099530). GluD is the effector channel in that experiment; the GPCR is mGlu1. Cited here for the pathway, not for the receptor typing. |
| human | GUSB | PMID:20028034 | Residues essential for plasminogen binding by the cation-independent mannose 6-phosphate receptor. | LOW | Full text verified. The paper is about the CI-MPR's plasminogen-binding site; human beta-glucuronidase is used only as a mannose-6-phosphate reference ligand purified on a CI-MPR affinity column. It supports M6P-receptor recognition (a trafficking event), not "signaling receptor binding" or a signaling function of GUSB; used here as basis for the over-annotation calls on the IPI binding terms. |
| human | HMGCS1 | PMID:7911016 | Molecular cloning and nucleotide sequence of complementary DNA for human hepatic cytosolic acetoacetyl-coenzyme A thiol… | LOW | This paper is about cytosolic acetoacetyl-CoA thiolase (ACAT2), the enzyme immediately UPSTREAM of HMGCS1 in the mevalonate pathway, not about HMGCS1 itself. It is used only as the reference for a curator IC (inferred from GO:0004421) placing HMGCS1 in cholesterol biosynthesis. The conclusion (HMGCS1 acts in cholesterol biosynthesis) is correct, but the reference is not about this gene. |
| human | HPRT1 | PMID:8643611 | Dopamine transporters are markedly reduced in Lesch-Nyhan disease in vivo. | MEDIUM | Documents reduced striatal DA transporters in LND patients (a downstream disease phenotype); does not establish an HPRT1 molecular function of regulating dopamine metabolism. Underlies an over-annotated GO:0045964 annotation. |
| human | INPP5D | PMID:9485206 | Inhibition of antigen-induced T cell response and antibody-induced NK cell cytotoxicity by NKG2A: association of NKG2A… | MEDIUM | Abstract explicitly says NKG2A phospho-ITIMs associate with SHP-1/SHP-2 but not SHIP; corresponding protein-binding annotation removed. |
| human | ITIH1 | PMID:2446322 | Isolation and characterization of cDNAs encoding the heavy chain of human inter-alpha-trypsin inhibitor (I alpha TI): u… | MEDIUM | PubMed identity and abstract were checked. The paper supports HC1 cDNA, liver expression, and multipolypeptide complex structure, but only predicts potential calcium-binding sites; it does not establish calcium ion binding. |
| human | ITIH3 | PMID:2465147 | Human plasma inter-alpha-trypsin inhibitor is encoded by four genes on three chromosomes. | MEDIUM | PubMed identity and abstract were checked. The paper establishes the H3 cDNA, liver RNA, and gene organization but does not demonstrate intrinsic endopeptidase-inhibitor activity in HC3; it is therefore not adequate support for the TAS molecular-function annotation. |
| human | ITIH4 | PMID:9756925 | Human inter-alpha-trypsin inhibitor heavy chain H3 gene. Genomic organization, promoter analysis, and gene linkage. | NONE | The PMID and title are internally correct, but the paper is an ITIH3 gene/promoter study. It mentions linkage to ITIH4 only and does not support a TAS assertion of ITIH4 endopeptidase inhibitor activity. |
| human | KDSR | PMID:34080016 | Glucosylceramide and galactosylceramide, small glycosphingolipids with significant impact on health and disease. | LOW | Review of glucosyl-/galactosylceramide biology; does not mention or assay KDSR. Cited for a glycosphingolipid biosynthesis annotation that is several enzymatic steps downstream of KDSR - a pathway-level over-annotation. |
| human | LNPEP | PMID:11062501 | Two new proteases in the MHC class I processing pathway. | LOW | The abstract explicitly names two other cytosolic enzymes. Reactome's attribution of an LNPEP aminopeptidase row to this paper is a wrong-entity mapping. |
| human | LNPEP | PMID:15691326 | Regulation of the human leukocyte-derived arginine aminopeptidase/endoplasmic reticulum-aminopeptidase 2 gene by interf… | LOW | The cited experiment is about ERAP2/L-RAP promoter regulation and mentions placental leucine aminopeptidase only as a related family member; it does not characterize LNPEP activity. |
| human | LNPEP | PMID:9668046 | Interferon-gamma can stimulate post-proteasomal trimming of the N terminus of an antigenic peptide by inducing leucine… | LOW | The abstract explicitly describes soluble HeLa-cell extracts and cytosolic LAP. Reactome's use for Q9UIQ6 is a historical name collision. |
| human | LNPEP | Reactome:R-HSA-1236977 | Endosomal/Vacuolar pathway | MEDIUM | LNPEP is supported in exogenous-antigen cross-presentation, but PMID:19498108 specifically demonstrates a proteasome-dependent IRAP route; TAP independence is not established for the LNPEP activity. |
| human | LNPEP | Reactome:R-HSA-983162 | Trimming of N-ter extended precursor fragments by cytosolic aminopeptidases | LOW | Q9UIQ6 inclusion reflects a historical LAP-name collision. Full-length LNPEP is membrane embedded and acts in endosomes rather than as one of these cytosolic enzymes. |
| human | LNPEP | Reactome:R-HSA-983168 | Antigen processing: Ubiquitination & Proteasome degradation | LOW | LNPEP trims downstream peptide precursors and is not ubiquitination machinery; the GO:0000209 projection is a pathway-participant overreach caused by the same cytosolic aminopeptidase mapping. |
| human | LNPK | PMID:12732147 | A global control region defines a chromosomal regulatory landscape containing the HoxD cluster. | LOW | Expression evidence does not establish membrane localization or causal limb development. |
| human | LONP1 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | A membrane-fraction proteomics detection does not establish resident membrane localization for the soluble mitochondrial-matrix enzyme. |
| human | LOX | PMID:22919265 | Yeast two-hybrid analysis of a human trabecular meshwork cDNA library identified EFEMP2 as a novel PITX2 interacting pr… | NONE | The cached full text identifies EFEMP2 as a PITX2-interacting protein in trabecular-meshwork experiments and does not assay or mention LOX. It cannot support a LOX protein-binding annotation. |
| human | LPCAT4 | Reactome:R-HSA-1483166 | Synthesis of PA | LOW | Generic PA-synthesis pathway is not sound LPCAT4 functional evidence; direct LPCAT4 assays in PMID:18458083 and PMID:41740885 found no appreciable LPA acylation. |
| human | LPCAT4 | Reactome:R-HSA-75885 | 1-acyl LPA is acylated to PA by AGPAT (LPAAT) | LOW | Generic AGPAT/LPAAT reaction does not explicitly identify LPCAT4 and conflicts with direct LPCAT4 assays showing no appreciable LPA acylation. |
| human | LRCH4 | PMID:16449650 | Identification and characterization of SAP25, a novel component of the mSin3 corepressor complex. | NONE | The PMID/title and full PMC article were checked externally; the project cache remains abstract-only, which is why full_text_unavailable is retained. This paper characterizes SAP25/Q8TEE9; searches of the full article for LRCH4, O75427, LRN and LRRN aliases found no LRCH4 evidence. The PML-body result belongs to SAP25, so this source does not support an LRCH4 nuclear localization or function. |
| human | LRCH4 | PMID:9799793 | Large-scale sequencing of two regions in human chromosome 7q22: analysis of 650 kb of genomic sequence around the EPO a… | MEDIUM | PubMed title and identifier are verified and the paper is relevant to discovery of the locus. However, it reports only weak sequence similarity to a neuronal protein and contains no developmental assay; it does not support the TAS claim that human LRCH4 is involved in nervous-system development. |
| human | LRP4 | PMID:18289866 | LRP promotes endocytosis and degradation, but not transcytosis, of the amyloid-beta peptide in a blood-brain barrier in… | NONE | This is a verified construct-name collision. “mLRP4” denotes a mini-receptor derived from LRP in a canine-cell BBB model, not human LRP4/O75096. |
| human | LRP4 | PMID:20093106 | The low-density lipoprotein receptor-related protein 10 is a negative regulator of the canonical Wnt/beta-catenin signa… | NONE | The abstract and title concern the distinct current LRP10 gene product, not LRP4/O75096. Because the cache is abstract-only, it cannot resolve how this citation became associated with LRP4, but it does not provide direct LRP4 support. |
| human | LRP5 | PMID:11029007 | LDL-receptor-related proteins in Wnt signal transduction. | LOW | PubMed title and abstract were verified. The experiments described in the accessible record concern LRP6 in Xenopus and Wnt/Fz assays; LRP5 appears only as a mammalian homolog of Arrow. The abstract therefore does not support a positive experimental LRP5 pathway annotation. |
| human | LRP5 | PMID:11433302 | Novel mechanism of Wnt signalling inhibition mediated by Dickkopf-1 interaction with LRP6/Arrow. | NONE | PubMed title and abstract were verified. The accessible evidence is explicitly about human DKK1 binding LRP6/Arrow, not LRP5, so it cannot verify the seeded LRP5-DKK1 IPI row. |
| human | LRP5 | PMID:12121999 | A novel set of Wnt-Frizzled fusion proteins identifies receptor components that activate beta -catenin-dependent signal… | MEDIUM | PubMed title and abstract were verified. The LRP5 result in this specific construct context is negative, so the paper is informative for receptor selectivity but does not support a positive LRP5 canonical-pathway IDA. |
| human | LRP5 | PMID:12857724 | Functional characterization of WNT7A signaling in PC12 cells: interaction with A FZD5 x LRP6 receptor complex and modul… | NONE | PubMed title and abstract were verified. This rat PC12 study explicitly assays FZD5-LRP6, not LRP5, and therefore does not support the seeded positive LRP5 transcription or canonical-Wnt annotations. |
| human | LRP5 | PMID:20093360 | Reconstitution of a frizzled8.Wnt3a.LRP6 signaling complex reveals multiple Wnt and Dkk1 binding sites on LRP6. | LOW | PubMed title and abstract were verified. The paper provides strong LRP6 structural-biochemical evidence and only family-level LRP5/6 background; it does not directly establish membership of human LRP5 in the assayed complex. |
| human | LRP5 | PMID:21471202 | Bone overgrowth-associated mutations in the LRP4 gene impair sclerostin facilitator function. | NONE | PubMed title and abstract were verified. This source explicitly studies LRP4 and cannot support the seeded LRP5-SOST IPI row; the distinct direct LRP5-SOST evidence is PMID:15908424. |
| human | LRP5 | Reactome:R-NUL-1458902 | frog CK1gamma phosphorylates LRP5/6 | NONE | Cached Reactome event verified. Despite its display-name shorthand, the event text is a frog LRP6 phosphorylation result and does not support plasma-membrane localization of human LRP5. |
| human | LRP6 | PMID:18721193 | LRP5 in premature adrenarche and in metabolic characteristics of prepubertal children. | NONE | The identifier and title are correct, but the paper studies LRP5 genetic associations and provides no LRP6 interaction evidence in the accessible record. It is miscited for an LRP6 IPI. |
| human | LRP6 | PMID:21471202 | Bone overgrowth-associated mutations in the LRP4 gene impair sclerostin facilitator function. | LOW | The identifier/title are correct, but the paper directly establishes LRP4-sclerostin binding and function, not LRP6 binding. It is miscited for the LRP6 IPI; direct LRP6-SOST evidence is instead in PMID:15778503 and PMID:15908424. |
| human | LRP6 | Reactome:R-HSA-4641236 | USP8 deubiquitinates FZD to potentiate WNT signaling | LOW | The cached Reactome event is about USP8-dependent FZD deubiquitination/recycling and does not mention LRP6. It is miscited for an LRP6 plasma-membrane location assertion. |
| human | LRP6 | Reactome:R-HSA-5340587 | RNF43 frameshift mutants show enhanced WNT siganling | LOW | The cached Reactome event describes RNF43-mutant effects on FZD abundance and Wnt signaling without identifying LRP6. It is miscited for an LRP6 plasma-membrane location assertion. |
| human | LRP6 | Reactome:R-NUL-1458871 | mFz8CRD binds LRP6N in vitro | MEDIUM | The cached event uses soluble mouse Fzd8 and human LRP6 ectodomain constructs in conditioned medium. It verifies a Wnt-dependent fragment association but is miscited for native extracellular localization of full-length LRP6. |
| human | LRP6 | Reactome:R-NUL-1458902 | frog CK1gamma phosphorylates LRP5/6 | MEDIUM | The cached event supports phosphorylation of LRP6 by Xenopus CK1gamma but does not establish the seeded human LRP6 plasma-membrane localization claim. |
| human | LRP6 | Reactome:R-NUL-209104 | Frog CKIgamma further phosphorylates Human LRP6 in the receptor complex | MEDIUM | The cached event states that frog CKIgamma phosphorylates human LRP6 in a receptor complex; it does not independently establish plasma-membrane localization. |
| human | MBL2 | PMID:2477488 | Lipopolysaccharide (LPS) binding protein opsonizes LPS-bearing particles for recognition by a novel receptor on macroph… | LOW | This paper is about lipopolysaccharide-binding protein (LBP), an acute-phase opsonin - NOT MBL2, and it does not assay MBL. The functions it is used to ground for MBL2 (opsonization, acute-phase response, defense response to bacterium, cell surface) are independently true of MBL but are correctly supported by other references; the D-mannose binding TAS attributed to this paper is not supported by… |
| human | MDH1 | PMID:8786100 | Molecular cloning and mapping of a human cDNA for cytosolic malate dehydrogenase (MDH1). | MEDIUM | cDNA cloning and chromosomal mapping (2p15) of cytosolic MDH1; supports the cytosol localization. Does NOT support the "malic enzyme activity" molecular function it is cited for (that MF is EC 1.1.1.40, a different enzyme). |
| human | MEA1 | PMID:2813404 | Male-enhanced antigen gene is phylogenetically conserved and expressed at late stages of spermatogenesis. | LOW | Verified against PubMed as Lau YF, Chan KM, Sparkes R, Proc Natl Acad Sci USA 1989 Nov;86(21):8462-6, doi 10.1073/pnas.86.21.8462; abstract only in the cache. The identifier and title are correct, but the paper does not support the two biological-process annotations drawn from it. It reports cloning, conservation and an expression pattern, and states its functional claim explicitly as a hypothesi… |
| human | MICU1 | PMID:9806765 | Isolation of cDNA clones coding for IgE autoantigens with serum IgE from atopic dermatitis patients. | LOW | PubMed-verified (J Invest Dermatol 1998) and correctly identified as the paper that cloned this cDNA, but it does not support the GO:0006952 defense response annotation drawn from it. MICU1/CBARA1 appears there only as one of four IgE autoantigens recovered from an expression library; IgE reactivity is a property of the patients' sera, not a function of the protein. |
| human | MMUT | PMID:20031578 | Novel associations of CPS1, MUT, NOX4, and DPEP1 with plasma homocysteine in a healthy population: a genome-wide evalua… | LOW | A GWAS associating a MUT-locus SNP with plasma homocysteine. The paper itself states MUT's catalytic activity is "hardly related to homocysteine"; it does not support direct involvement of MMUT in homocysteine metabolic process or binding of a modified amino acid, so the IDA annotations to GO:0050667 and GO:0072341 are over-interpretations. |
| human | MOCOS | GO_REF:0000104 | Electronic Gene Ontology annotations created by transferring manual GO annotations between related proteins based on sh… | LOW | UniRule sequence-feature transfer that yields GO:0016829 lyase activity for MOCOS. MOCOS is a sulfurtransferase (EC 2.8.1.9); the terminal reaction transfers sulfur to the Moco rather than performing a lyase (bond-cleaving, non-hydrolytic) reaction, so the lyase MF is an incorrect over-propagation for this protein. |
| human | MTHFD1 | PMID:3528153 | Purification and characterization of a mitochondrial isozyme of C1-tetrahydrofolate synthase from Saccharomyces cerevis… | LOW | This is the yeast MITOCHONDRIAL C1-THF synthase isozyme (the paralog of MTHFD2/MTHFD1L), and the paper explicitly states the cytoplasmic synthase is NOT mitochondrial. It does not support a mitochondrial location for human cytoplasmic MTHFD1. |
| human | MTHFR | PMID:20031578 | Novel associations of CPS1, MUT, NOX4, and DPEP1 with plasma homocysteine in a healthy population: a genome-wide evalua… | MEDIUM | A GWAS of plasma homocysteine that confirms the MTHFR locus association. It does NOT provide direct (IDA) evidence for MTHFR "modified amino acid binding" or a direct homocysteine-metabolic assay; the IDA annotations sourced to it are over-attributed. |
| human | MTRR | PMID:21071249 | Interaction between MMACHC and MMADHC, two human proteins participating in intracellular vitamin B₁₂ metabolism. | LOW | PubMed-verified paper, but it characterises the MMACHC-MMADHC interaction and does not establish a molecular carrier function for MTRR. Cited by Reactome for the GO:0140104 molecular carrier activity annotation on MTRR, which it does not support; the MTRR molecular carrier annotation is flagged as over-annotated. |
| human | MVK | PMID:16732551 | AA amyloidosis complicating hyperimmunoglobulinemia D with periodic fever syndrome: a report of two cases. | LOW | Cited to support GO:0050728 negative regulation of inflammatory response, but this is a two-patient AA amyloidosis case report and does not demonstrate a direct MK-mediated regulation of inflammation; the inflammatory phenotype is a downstream consequence of MKD. |
| human | MYO7A | PMID:11398101 | Mutations of the protocadherin gene PCDH15 cause Usher syndrome type 1F. | LOW | The identifier resolves to the intended paper by Ahmed et al. 2001 on PCDH15 and USH1F, but that paper is an experimental study of PCDH15, not of MYO7A; MYO7A appears only in background statements about known Usher genes. It is therefore a weak citation for the two MYO7A IMP annotations (GO:0007605 and GO:0050953) made against it. The underlying claims are correct and are properly supported by PM… |
| human | NAMPT | PMID:23001182 | Interactions among HCLS1, HAX1 and LEF-1 proteins are essential for G-CSF-triggered granulopoiesis. | LOW | Paper is about HCLS1/HAX1/LEF-1; NAMPT is only a pathway component. Does not support a direct positive-regulation-of-Pol-II-transcription MF/BP role for NAMPT. |
| human | NANP | GO_REF:0000041 | Gene Ontology annotation based on UniPathway vocabulary mapping | LOW | UniPathway mapping (UPA00630 = N-acetylneuraminate biosynthesis) was mapped to GO:0006045 N-acetylglucosamine biosynthetic process, which is the wrong process; the correct target is N-acetylneuraminate biosynthesis (GO:0046380). |
| human | NDUFA6 | PMID:9345899 | In situ hybridisation mapping of genomic clones for five human respiratory chain complex I genes. | LOW | This is a chromosomal gene-mapping (in situ hybridisation) study; it maps NDUFA6 to 21q22 but does not demonstrate any catalytic (NADH dehydrogenase) activity of NDUFA6. Cited as TAS for GO:0008137 (enables), which it does not support; used here to justify MARK_AS_OVER_ANNOTATED of that MF. |
| human | NDUFC2 | Reactome:R-HSA-6798739 | Exocytosis of azurophil granule membrane proteins | LOW | Neutrophil degranulation pathway that seeds plasma-membrane and azurophil-granule-membrane localizations for NDUFC2; these are implausible for a mitochondrial Complex I subunit (pathway bulk over-inclusion). |
| human | NDUFS3 | PMID:22926577 | Quantitative proteomic analysis of human substantia nigra in Alzheimer's disease, Huntington's disease and Multiple scl… | LOW | PubMed-verified identifier, but the paper is a differential-abundance proteomics survey; it does not support a role for NDUFS3 in substantia nigra development. The HEP substantia nigra development annotation is an over-annotation. |
| human | NLRP3 | PMID:15967716 | NLRs join TLRs as innate sensors of pathogens. | LOW | Meylan, Tschopp & Karin, a 2005 review of NLRs and TLRs. Correctly identified, and reasonable as TAS support for the general cytoplasm and defense-response annotations, but it is cited in GOA as the support for GO:0042834 peptidoglycan binding, a ligand-binding molecular function for which neither this review nor any primary study provides direct evidence. That is the basis for the REMOVE on that… |
| human | NQO2 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | LOW | The IBA propagated GO:0003955 (NAD(P)H dehydrogenase (quinone) activity, EC 1.6.5.2) to NQO2, but NQO2 uses NRH rather than NAD(P)H; the phylogenetic inference over-generalizes the family cofactor. The IBA cytosol annotation is correct. |
| human | NSUN2 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | MEDIUM | The tRNA-related IBAs on this gene are sound. Three others (rRNA C5 methyltransferase activity, rRNA processing, mitochondrial large ribosomal subunit assembly) name mouse Nsun4 as the donor in the GOA WITH/FROM column and appear to carry a paralogue-specific function onto NSUN2. |
| human | NSUN2 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | ARBA machine-learning annotations. tRNA methylation is right; mRNA processing is the wrong branch for what NSUN2 does to mRNA. |
| human | NSUN7 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | MEDIUM | The InterPro2GO mapping is correctly applied as a procedure, but IPR001678 fires on the NOL1/NOP2/Sun fold and cannot see that NSUN7's motif IV has lost the SAM-binding aspartate. The resulting methyltransferase-activity assertion is refuted by direct assay. |
| human | NT5C3A | PMID:9428647 | Pyrimidine nucleotidases from human erythrocyte possess phosphotransferase activities specific for pyrimidine nucleotid… | MEDIUM | PubMed-verified. Cited as support for GO:0000215 (tRNA 2'-phosphotransferase activity), but the paper describes a pyrimidine-nucleotide phosphotransferase, not a tRNA 2'-phosphotransferase; the GO:0000215 annotation is a mis-mapping. |
| human | P2RX7 | PMID:12849743 | Up-regulation of P2X2, P2X4 receptor and ischemic cell death: prevention by P2 antagonists. | NONE | PubMed-verified (Neuroscience 2003). The abstract reports up-regulation of P2X2 and P2X4, not P2X7, in ischemic neuronal death. Used in GOA as a NAS annotation for P2X7 response to ischemia. Full text not available here, so the annotation is flagged rather than removed, but this reference does not appear to be about P2X7. |
| human | P2RX7 | PMID:17299767 | Involvement of P2X4 and P2Y12 receptors in ATP-induced microglial chemotaxis. | LOW | PubMed-verified (Glia 2007). The paper is about P2X4 and P2Y12 in microglial chemotaxis; P2X7 is not its subject. It is used in GOA only as a TAS plasma-membrane localisation, which is harmless, but it is a poor choice of reference for this gene. |
| human | PAPSS1 | PMID:23207770 | Ethanol sulfation by the human cytosolic sulfotransferases: a systematic analysis. | LOW | PubMed-verified paper, but it studies SULT sulfotransferases and 35S-sulfate labeling; it is a weak/indirect source for the PAPSS1 APS-kinase and PAPS-biosynthesis IDA annotations attributed to it (CAFA-assigned). |
| human | PAPSS1 | PMID:9771708 | Mutations in orthologous genes in human spondyloepimetaphyseal dysplasia and the brachymorphic mouse. | LOW | PubMed-verified, but the paper is about PAPSS2/ATPSK2, not PAPSS1; it does not support a skeletal-development role specific to PAPSS1. |
| human | PCBD1 | GO_REF:0000107 | Automatic transfer of experimentally verified manual GO annotation data to orthologs using Ensembl Compara | LOW | The Ensembl Compara transfer from mouse Pcbd1 produced an incorrect GO:0004505 (phenylalanine 4-monooxygenase activity) annotation on PCBD1. That activity belongs to PAH, not PCBD1; this is an over-propagation artifact and the resulting annotation is recommended for removal. |
| human | PGD | PMID:3858849 | 6-Phosphogluconolactonase deficiency, a hereditary erythrocyte enzyme deficiency: possible interaction with glucose-6-p… | LOW | Cited to support the PGD oxidative-branch IDA, but the abstract concerns 6-phosphogluconolactonase (PGLS, EC 3.1.1.31), the second oxiPPP enzyme, and its interaction with G6PD deficiency, not PGD. Abstract-only; associated annotation left UNDECIDED rather than removed. |
| human | PHKG2 | PMID:8999860 | The regulatory Ser262 of microtubule-associated protein tau is phosphorylated by phosphorylase kinase. | LOW | Demonstrates in-vitro tau phosphorylation by rabbit skeletal muscle phosphorylase kinase (muscle PHKG1 holoenzyme), not the liver/testis PHKG2 isoform; supports an over-annotated, non-physiological activity for PHKG2. |
| human | PLCG2 | PMID:18784435 | Molecular profiling of isolated histological components of wilms tumor implicates a common role for the Wnt signaling p… | LOW | Cached abstract is an expression-profiling Wilms tumor paper and does not establish PLCG2 as a Wnt signaling component; used to remove the Wnt annotation. |
| human | PMM2 | PMID:9525984 | Carbohydrate-deficient glycoprotein syndrome type Ib. Phosphomannose isomerase deficiency and mannose therapy. | LOW | Cited as the IMP source for PMM2 GDP-mannose biosynthetic process, but the paper is about phosphomannose isomerase (PMI/MPI) deficiency / CDG-Ib and mannose therapy - a different enzyme catalyzing F6P<->Man6P and a different gene (MPI). The GO term is still correct for PMM2 and is strongly supported by other references, so the annotation was kept; the citation itself appears mis-attributed. Recom… |
| human | PNPO | PMID:34800366 | Quantitative high-confidence human mitochondrial proteome and its dynamics in cellular context. | LOW | Large-scale mitochondrial proteome study; used to support a mitochondrion localization for PNPO, but PNPO is a canonical cytosolic enzyme and the cached full text does not specifically establish PNPO as a mitochondrial resident. Likely reflects cytosolic contamination; annotation flagged as over-annotation. |
| human | POFUT1 | PMID:11698403 | Composition of Drosophila melanogaster proteome involved in fucosylated glycan metabolism. | LOW | PubMed-verified, but this is a Drosophila genome bioinformatics survey, not a study of human POFUT1 enzymatic activity. It is a weak/inappropriate basis for a TAS human peptide-O-fucosyltransferase-activity annotation; the same MF term is far better supported by human IDA (PMID:11524432, PMID:9023546, PMID:15653671). |
| human | PPCDC | PMID:26514574 | Complex stability and dynamic subunit interchange modulates the disparate activities of the yeast moonlighting proteins… | LOW | PubMed-verified, but this is a yeast (S. cerevisiae) study of the heterotrimeric Hal3/Vhs3/Cab3 PPCDC, not a direct assay of human PPCDC self-interaction. It informs the family-level oligomerization concept only; the human homotrimer (PMID:14501115) is the appropriate support for GO:0042802. |
| human | PRODH | PMID:21998747 | Structural and biochemical studies of human 4-hydroxy-2-oxoglutarate aldolase: implications for hydroxyproline metaboli… | LOW | Full-text paper on human HOGA (4-hydroxy-2-oxoglutarate aldolase). It describes the 4-hydroxyproline degradation pathway as beginning with hydroxyproline oxidase (HPOX) — the distinct paralog PRODH2/HYPDH — and does not implicate PRODH. The GO:0019470 (trans-4-hydroxy-L-proline catabolic process) annotation to PRODH from this paper is a paralog mis-attribution; marked over-annotated. |
| human | PSAP | PMID:23555801 | BANK1 and BLK act through phospholipase C gamma 2 in B-cell signaling. | NONE | Cited as support for a PSAP scaffold protein binding (GO:0097110) annotation, but the paper is about BANK1/BLK/PLCg2 in B-cell signalling and does not mention prosaposin, PSAP or saposins anywhere in the cached full text. The annotation cannot be verified from this reference and is likely a citation/interactor mismatch; the associated annotation is marked over-annotated. |
| human | PSPH | GO_REF:0000107 | Automatic transfer of experimentally verified manual GO annotation data to orthologs using Ensembl Compara | LOW | Basis for the three rat-ortholog IEA response-phenotype annotations (response to mechanical stimulus / nutrient levels / testosterone). The transferred terms are regulatory-context phenotypes of the rat ortholog, not molecular functions of human PSPH; flagged as over-annotations. |
| human | PTTG1 | PMID:15929994 | Human full-length Securin is a natively unfolded protein. | MEDIUM | PubMed identity and abstract were verified. The paper directly establishes intrinsic disorder but its accessible experimental summary does not establish molecular-function activator activity; direct activation support comes from PMID:11371342 and PMID:23798554. |
| human | RABGGTB | PMID:8380507 | Retinal degeneration in choroideremia: deficiency of rab geranylgeranyl transferase. | LOW | This paper is about component A (REP1/CHM), the choroideremia disease gene, and explicitly finds that component B (which contains RABGGTB) is NOT deficient. It is therefore the wrong reference to attribute "visual perception" and "protein modification process" to RABGGTB; the visual phenotype maps to CHM, not to this catalytic subunit. |
| human | RCE1 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | MEDIUM | InterPro2GO (IPR039731) pipeline reference. The CAAX-box protein processing mapping it produces is correct, but the metalloendopeptidase activity mapping from the same reference is the wrong catalytic class for this glutamate-type protease. |
| human | RCE1 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | HIGH | PAN-GO phylogenetic (IBA) reference. The ER-membrane and CAAX-box protein processing inferences are correct, but the metalloendopeptidase-activity inference propagated in the same tree assigns the wrong catalytic class. |
| human | RPL36A | PMID:26684695 | Expression of Muscle-Specific Ribosomal Protein L3-Like Impairs Myotube Growth. | NONE | PubMed identity and full text were verified. The experiments perturb RPL3L in mouse myoblasts and never assay RPL36A; the paper does not support assigning the RPL3L fusion phenotype to RPL36A. |
| human | RPL36A | PMID:34081545 | Knockdown of Muscle-Specific Ribosomal Protein L3-Like Enhances Muscle Function in Healthy and Dystrophic Mice. | NONE | PubMed identity and abstract were verified. This mouse study knocks down Rpl3l and does not assay RPL36A, so it cannot support an RPL36A striated-muscle-contraction annotation. |
| human | SECISBP2 | PMID:19467292 | SECIS-binding protein 2, a key player in selenoprotein synthesis, is an intrinsically disordered protein. | MEDIUM | Abstract-only; DisProt source (DP00420). Characterizes SBP2 as an intrinsically disordered protein with a folded RNA-binding domain. Does NOT support the GO:0003677 DNA binding annotation it is cited for; SBP2's nucleic-acid-binding activity is RNA (SECIS) binding. |
| human | SLC25A10 | PMID:21630459 | Proteomic characterization of the human sperm nucleus. | LOW | Sperm-nucleus proteome; underlies the nucleus HDA annotation. For a mitochondrial inner-membrane transporter this is a co-purification or contaminant artifact and does not support a genuine nuclear localization. |
| human | SLC25A22 | Reactome:R-HSA-442660 | SLC-mediated transport of neurotransmitters | LOW | Reactome node describes the SLC6 plasma-membrane neurotransmitter transporters, not SLC25A22's mitochondrial glutamate/H+ symport; used to support an over-annotated neurotransmitter-transport term (GO:0006836). |
| human | SLC25A26 | Reactome:R-HSA-549127 | SLC-mediated transport of organic cations | LOW | This Reactome pathway groups SLC22/OCT organic-cation transporters; it is a mismatched basis for the "amine transport" annotation on SAMC. The correct Reactome event for SAMC is R-HSA-8855062. |
| human | SLC25A3 | PMID:23209302 | KIF14 negatively regulates Rap1a-Radil signaling during breast cancer progression. | LOW | Full text available and searched; the paper is about KIF14 and the Rap1a- Radil pathway and does not mention SLC25A3, the phosphate carrier, or Q00325. It does not support a protein-complex-binding function for SLC25A3 in the cached (human-titled) record; the GOA IDA was assigned by MGI and presumably derives from a mouse Slc25a3 context not represented here. |
| human | SLC25A5 | GO_REF:0000108 | Automatic assignment of GO terms using logical inference, based on on inter-ontology links | LOW | Inter-ontology inference produced "adenine transport" from the incorrect "adenine transmembrane transporter activity"; ANT transports adenine nucleotides, not free adenine, so the inferred term has the wrong substrate. |
| human | SLC25A5 | PMID:19116139 | Capsaicin binds to prohibitin 2 and displaces it from the mitochondria to the nucleus. | LOW | This capsaicin/prohibitin-2 paper is cited (IMP) for ANT2 nucleotide-transport activity and adenine nucleotide transport, but it does not measure ANT2 transport; it supports only a prohibitin-2 co-purification (protein binding). |
| human | SLC25A5 | PMID:2168878 | The human fibroblast adenine nucleotide translocator gene. Molecular cloning and sequence. | MEDIUM | Genomic cloning/sequence of the fibroblast (ANT2) gene; no transport assay and no localization experiment, yet cited (TAS) for "adenine transmembrane transporter activity" (wrong substrate) and "plasma membrane" location. |
| human | SLC6A6 | PMID:26590417 | Establishment and Dysfunction of the Blood-Brain Barrier. | LOW | General BBB review; cached text does not specifically establish SLC6A6 as a blood-brain-barrier taurine transporter. Weak support for the NAS transport-across-BBB annotation. |
| human | SLC6A6 | PMID:30280653 | Blood-Brain Barrier: From Physiology to Disease and Back. | LOW | Comprehensive BBB physiology review; cached text does not specifically establish an SLC6A6 BBB transport role. Weak support for the NAS annotation. |
| human | SLC6A8 | PMID:8661155 | The genomic organization of a human creatine transporter (CRTR) gene located in Xq28. | LOW | PubMed-verified as a genomic-organization paper; cited as TAS for muscle contraction (GO:0006936), which it does not directly establish. The muscle-contraction link is an indirect physiological inference, kept as non-core. |
| human | SMPD4 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | LOW | InterPro (IPR024129) is mapped to GO:0050290 sphingomyelinase D activity, which is the wrong reaction for SMPD4 (a type-C sphingomyelinase, EC 3.1.4.12). The pipeline correctly identifies an SMPD4-family function, but the target GO term is incorrect. |
| human | SPHK1 | PMID:23106337 | Sphingosine 1-phosphate induces filopodia formation through S1PR2 activation of ERM proteins. | MEDIUM | PubMed-verified paper, but it does NOT support assigning "sphingosine-1-phosphate receptor activity" (GO:0038036) to SPHK1: the receptor is S1PR2. The SPHK1 GO:0038036 IMP citing this paper is a mis-assignment of receptor activity to the kinase. |
| human | SPHK2 | PMID:12391145 | Sphingosine kinase mediates vascular endothelial growth factor-induced activation of ras and mitogen-activated protein… | LOW | Basis of the NAS 'small GTPase binding' annotation to SPHK2, but the paper attributes the Ras/MAPK effect to SPHK1 (siRNA against SPK1, not SPK2, blocks it) and never assays SPHK2-GTPase binding. Does not support the SPHK2 annotation. |
| human | SPHK2 | PMID:23106337 | Sphingosine 1-phosphate induces filopodia formation through S1PR2 activation of ERM proteins. | LOW | Cited for SPHK2 'S1P receptor activity' and 'sphingosine biosynthetic process' IMP/IGI, but this is an S1P/S1PR2 signaling study attributing S1P generation mainly to SPHK1; SPHK2 is not a receptor and does not synthesize sphingosine. Does not support those SPHK2 annotations. |
| human | SPTLC1 | PMID:30242129 | Complex formation of sphingomyelin synthase 1 with glucosylceramide synthase increases sphingomyelin and decreases gluc… | LOW | Full text is about SMS1 (SGMS1) and GCS (UGCG); it does not mention SPTLC1 or serine palmitoyltransferase. Used to annotate SPTLC1 to sphingomyelin biosynthetic process, a downstream terminus not assayed here; the SPTLC1 annotation is over-annotated/miscited. |
| human | SPTSSA | PMID:30242129 | Complex formation of sphingomyelin synthase 1 with glucosylceramide synthase increases sphingomyelin and decreases gluc… | LOW | Full text does not mention SPTSSA/ssSPTa; the paper concerns the SMS1-GCS Golgi complex. The GOA sphingomyelin biosynthetic process IDA citing this paper is an over-annotation for SPTSSA. |
| human | ST6GAL1 | PMID:16237761 | Screening of hepatocyte proteins binding to F protein of hepatitis C virus by yeast two-hybrid system. | LOW | Verified against the cached full text: a yeast two-hybrid screen for hepatocyte proteins binding HCV F protein, in which "1 colony was sialyltransferase" among 36 colonies, with no ST6GAL1-specific validation. This is a generic, weakly supported protein-binding (IPI) annotation that does not inform ST6GAL1 function and is recommended for removal. |
| human | STAR | PMID:18403318 | Intracellular cholesterol transporter StarD4 binds free cholesterol and increases cholesteryl ester formation. | MEDIUM | Primarily a StarD4 (paralog) paper. It names StarD1/STAR only as the reference cholesterol-binding START protein, which supports the STAR cholesterol-binding IDA. Its bile-acid biosynthesis phenotype is a StarD4 property and does not support the STAR bile-acid regulation annotation (flagged over-annotated). |
| human | STING1 | PMID:18818105 | The adaptor protein MITA links virus-sensing receptors to IRF3 transcription factor activation. | HIGH | PubMed-verified: Zhong B et al., Immunity 2008;29(4):538-50 (the MITA paper) - the identifier and title are correct and the paper is a landmark for this gene. Flagged MISCITED because of what two of the annotations drawn from it assert, not because of the citation itself: GO:0003713 transcription coactivator activity and GO:0005741 mitochondrial outer membrane. The paper's own abstract says MITA… |
| human | SULT1A1 | PMID:8661000 | Human phenol sulfotransferase STP2 gene: molecular cloning, structural characterization, and chromosomal localization. | LOW | The PMID resolves to the intended paper, but the paper characterizes the paralogous gene and contains no enzymology, so it supports neither annotation drawn from it for SULT1A1. Two independent lines confirm this. (1) Gene identity, checkable against UniProt without the full text: this is reference 4 of P50226 (SULT1A2), whose gene-name line reads "Name=SULT1A2; Synonyms=STP2", while P50225 (SULT… |
| human | SULT1B1 | PMID:21492153 | Analysis of proteomic changes induced upon cellular differentiation of the human intestinal cell line Caco-2. | LOW | The paper is correctly identified, but it reports differential abundance of SULT1B1 between proliferating and differentiated Caco-2 cells. That is an expression correlation and does not support involved_in epithelial cell differentiation. |
| human | SULT1B1 | PMID:23207770 | Ethanol sulfation by the human cytosolic sulfotransferases: a systematic analysis. | MEDIUM | PubMed-verified (Biol Pharm Bull 2012;35:2180-5). Correct paper, but it is used to support an ethanol catabolic process annotation for SULT1B1 when its own summary names SULT1A1, SULT1A2, SULT1A3 and SULT1C4 as the ethanol-sulfating enzymes among the eleven SULTs assayed. Cached record is abstract-only, so the full-text panel data could not be checked. |
| human | SYPL2 | PMID:22926577 | Quantitative proteomic analysis of human substantia nigra in Alzheimer's disease, Huntington's disease and Multiple scl… | LOW | The PMID and title are correct and the full dataset may contain SYPL2, but adult disease-tissue proteomics does not support the asserted role in substantia nigra development. Abstract-only in the cache, so SYPL2's presence in the protein table could not be independently checked. |
| human | TBCK | PMID:12471243 | The protein kinase complement of the human genome. | MEDIUM | PubMed-verified (Science 2002;298:1912-34). The paper is sound and is correctly cited as the source of TBCK's placement in the human kinome, but it is a sequence-based census, not a functional study, and cannot support an assertion that TBCK enables protein kinase activity. It does support the accompanying NOT protein phosphorylation row insofar as kinome membership was never a catalytic claim. |
| human | TH | PMID:17520326 | Oxygen dependence of tyrosine hydroxylase. | MEDIUM | Solid in vitro enzymology of TH oxygen (O2) Km, but it does NOT demonstrate a cellular "response to hypoxia" biological process; the GO:0001666 annotation over-interprets the enzyme oxygen-kinetics data. |
| human | TMC1 | PMID:23871232 | TMC1 and TMC2 are components of the mechanotransduction channel in hair cells of the mammalian inner ear. | HIGH | PubMed-verified (Neuron 2013) and scientifically sound, but it is the source of the mouse enables GO:0005245 voltage-gated calcium channel activity annotation that propagates to human TMC1. The paper measures calcium permeability and single-channel conductance of a mechanically gated channel and never claims voltage gating, so it does not support the term it was used for. |
| human | TMEM120A | PMID:42184262 | Biolayer Interferometry for Investigating Membrane Protein-Inhibitor Binding: TACAN Mutant and GsMTx4 As a Model System. | LOW | PubMed-verified (J Vis Exp 2026); cached record is abstract-only. A methods paper demonstrating a biolayer-interferometry protocol, using detergent-solubilised TMEM120A M207A and the gating-modifier peptide GsMTx4 as a convenient model pair. It measures binding kinetics, not channel gating, and the authors describe TMEM120A as "multifunctional" rather than as a channel. Flagged MISCITED in the se… |
| human | TNFRSF21 | PMID:25898930 | Death Receptor 6 and Caspase-6 Regulate Prion Peptide-Induced Axonal Degeneration in Rat Spinal Neurons. | LOW | PubMed-verified as Wang et al., J Mol Neurosci 2015, on DR6 and caspase-6 in prion-peptide-induced axonal degeneration in rat spinal neurons. It is the sole source of the rat IMP that the human GO:0007413 axonal fasciculation IEA is propagated from, and its subject (axonal degeneration) does not match that term (collection of axons into a fascicle). Not cited directly by the human GOA row, listed… |
| human | TTC8 | PMID:22302990 | Direct role of Bardet-Biedl syndrome proteins in transcriptional regulation. | LOW | Study is BBS7-centric; the transcription/RNF2 role is proposed to extend to other BBS proteins. The GO:0061629 label also mischaracterizes RNF2 (a PRC1 E3 ligase) as an RNA Pol II transcription factor. Treat as over-annotation for TTC8. |
| human | UFD1 | PMID:9063746 | UFD1L, a developmentally expressed ubiquitination gene, is deleted in CATCH 22 syndrome. | MEDIUM | Title from UniProt reference list; not cached. Correctly supports the ubiquitin-dependent catabolic-process and developmental/22q11 context, but it is mis-cited for the cysteine-type deubiquitinase activity annotation (UFD1 is not a DUB). |
| human | UPP2 | PMID:11278417 | Uridine phosphorylase association with vimentin. Intracellular distribution and localization. | LOW | Studies UPP1 (the first-identified human uridine phosphorylase) and its vimentin association; published in 2001, two years before UPP2 was cloned (Johansson 2003). GOA annotations attributing type III intermediate filament localization (and the NAS cytosol / uridine metabolic process terms) to UPP2 from this paper are very likely a paralog mis-attribution. The general cytosolic-pool statement is… |
| human | USH1C | PMID:11398101 | Mutations of the protocadherin gene PCDH15 cause Usher syndrome type 1F. | LOW | Full text available and concerns PCDH15/USH1F; it contains no experimental data on USH1C, so the three USH1C IMP annotations citing it appear to carry a misassigned reference. The corresponding GO terms are nonetheless correct for USH1C on the basis of the USH1C patient phenotype (PMID:10973247), so the annotations were accepted with the citation issue flagged here. |
| human | USH1G | PMID:11398101 | Mutations of the protocadherin gene PCDH15 cause Usher syndrome type 1F. | NONE | The PMID resolves to the paper whose title is recorded, but that paper is about PCDH15/USH1F and does not concern USH1G/SANS. The cached record has full_text_available true and a full-text search returns zero occurrences of 'SANS' or 'USH1G'; the paper was published in 2001, two years before SANS was identified as the USH1G gene product (PMID:12588794). It therefore cannot be the source of the th… |
| mouse | Mir100 | PMID:23646144 | Identifying microRNAs involved in degeneration of the organ of corti during age-related hearing loss. | LOW | PMC full text checked and contains no mention of miR-100; the annotation must derive from a supplementary microarray table. The paper is about degeneration and hearing loss, not the normal perception of sound. |
| mouse | Mir100 | PMID:25858512 | miR-26a and miR-384-5p are required for LTP maintenance and spine enlargement. | NONE | PMC full text checked; miR-100 does not appear. The annotation reflects inclusion in the detected-miRNA set, not any finding about mir-100. |
| mouse | Mir127 | PMID:20439489 | miRNA 34a, 100, and 137 modulate differentiation of mouse embryonic stem cells. | LOW | Cached record is abstract-only, and mir-127 is not among the three miRNAs the abstract reports as tested. It is one of the untested members of a 291-row IEP batch drawn from this paper. |
| mouse | Mir127 | PMID:25858512 | miR-26a and miR-384-5p are required for LTP maintenance and spine enlargement. | NONE | PMC full text checked; miR-127 does not appear. The annotation reflects membership of the detected-miRNA set only. |
| mouse | Mir26a-1 | PMID:24205035 | Profiling circulating microRNA expression in experimental sepsis using cecal ligation and puncture. | MEDIUM | The paper genuinely measures miR-26a, so the exosome and RISC rows it supports are sound. It does not support the two process rows drawn from it (response to bacterium, response to wounding); for the bacterium term it argues the opposite. |
| mouse | Mir30e | PMID:24205035 | Profiling circulating microRNA expression in experimental sepsis using cecal ligation and puncture. | LOW | PMC full text checked. mir-30e appears only as a row in a microarray table, not among the 10 qPCR-validated hits, and the paper argues against bacterium-sensing as the driver. |
| mouse | Mir30e | PMID:25858512 | miR-26a and miR-384-5p are required for LTP maintenance and spine enlargement. | MEDIUM | PMC full text checked. mir-30e is genuinely differentially expressed, so the IEP observation is real, but the paper makes no functional claim about it. |
| mouse | Mir384 | PMID:20439489 | miRNA 34a, 100, and 137 modulate differentiation of mouse embryonic stem cells. | LOW | Correctly identified paper, but it does not support a mir-384 role in LIF response. It is the source of a 291-gene IEP batch in which mir-384 is one of the untested members. |
| rat | Arsb | PMID:1682910 | Inability of thiol compounds to restore CNS arylsulfatases inhibited by methyl mercury. | LOW | PubMed-verified. Shows methylmercury inhibits rat CNS arylsulfatase activity; used to support GO:0051597 response to methylmercury but the evidence is a toxicology observation, not evidence of Arsb participating in a biological response process. |
| rat | Arsb | PMID:2290353 | Characterization of liver lysosomal enzyme activity in hepatocytes, Kupffer and endothelial cells during aging: effect… | LOW | PubMed-verified. Shows age- and diet-dependent changes in hepatic lysosomal enzyme activities (including Arsb) but does not demonstrate that Arsb is a component of a nutrient-response pathway. |
| rat | Arsb | PMID:6137211 | The effect of pH on the kinetics of arylsulphatases A and B. | LOW | PubMed-verified. Enzyme-kinetics study of pH dependence of rat liver arylsulphatases; used incorrectly to support GO:0009268 response to pH. |
| rat | Arsb | PMID:8037 | Autophagy-related changes of arylsulphatases A and B in rat liver lysosomes. | MEDIUM | PubMed-verified. Sanghavi & Koenig 1976 describes arylsulphatase A/B activity ratio changes during hepatic autophagocytosis, with A converted to B by other lysosomal hydrolases active in autophagic vacuoles. The paper shows Arsb is a target/substrate of autophagic processing, not evidence that Arsb has a role in autophagy. Flagged for community adjudication under geneontology/go-annotation issue… |
| yeast | CAF120 | GO_REF:0000033 | Annotation inferences using phylogenetic trees | MEDIUM | The IBA pipeline (PANTHER family PTN001969686) propagated protein kinase activity and signal transduction onto CAF120 from Arabidopsis MAP3K reference members that carry a kinase domain CAF120 lacks; the inference is not applicable to this gene. |
| yeast | NIT1 | file:yeast/NIT1/NIT1-deep-research-falcon.md | Falcon deep research report for NIT1 (YIL164C) | LOW | Genuine late falcon artifact, retained for provenance. Its central functional claim for YIL164C is a paralog mis-attribution (it copies the NIT2/YJL126W dGSH-amidase characterization onto YIL164C); NOT used to support any functional annotation here. Flagged as a cautionary example of nitrilase-family over-annotation via name confusion. |
| yeast | SAS3 | PMID:11731479 | The yeast SAS (something about silencing) protein complex contains a MYST-type putative acetyltransferase and functions… | LOW | Correct paper (Osada et al. 2001, PMC312835), but it does not support the two SGD IMP annotations that cite it for SAS3 (GO:0030466, GO:0031509). Full text read at PMC312835 (the local cache is abstract-only): Table 1 lists sas2, sas4, sas5, sir1 and asf1 alleles and the SAS2 M1-M3 acetyl-CoA-motif point mutants, but no sas3 allele, and no SAS3 mutant phenotype is reported. The IMP most likely be… |
| yeast | YDJ1 | PMID:10811660 | Crystal structure and activity of human p23, a heat shock protein 90 co-chaperone. | NONE | The cached abstract is exclusively about human p23 and its Hsp90/progesterone-receptor chaperoning assays; it contains no YDJ1 experiment and does not support direct IDA provenance for the YDJ1 annotation. The reference is retained because the existing GOA annotation cites it. |
DISPUTED (282)
| Organism | Gene | Reference | Title | Rel. | Notes |
|---|---|---|---|---|---|
| 9POAL | NCGR_LOCUS10166 | file:9POAL/NCGR_LOCUS10166/NCGR_LOCUS10166-deep-research-manual.md | Manual deep research on NCGR_LOCUS10166 identifying gene prediction artifact. | HIGH | Preserved historical analysis. Categorical artifact/exclusive-localization conclusions exceed its evidence. Sequential locus numbers did not prove adjacency; deposited coordinates now establish adjacency on chromosome 1 but not an error at the chromosome-2 candidate. ARV warning was misapplied to HDH. |
| 9POAL | NCGR_LOCUS1270 | file:9POAL/NCGR_LOCUS1270/NCGR_LOCUS1270-hypotheses/function-hypothesis-go-0006000/openscientist.md | OpenScientist fructose-process adjudication for NCGR_LOCUS1270 | HIGH | Read in full. Useful FBPase/chloroplast context and separate ontology branches, but absent is_a ancestry is not disjointness. Existing human FBP1/FBP2 TAS and Dictyostelium fbp IDA annotations contradict the blanket free-fructose-substrate rule. Actual PTN004269459 maps to a chloroplast-associated sorghum protein, so cytosolic-paralog misgrafting was not demonstrated. Export contained no executed… |
| ANOGA | TOLL9 | file:ANOGA/TOLL9/TOLL9-deep-research-falcon.md | Deep research report on TOLL9 function | MEDIUM | Useful family background, but pathway-wide and other-receptor results do not establish this target ligand/cascade. Processed Spaetzle recognition was conflated with direct PAMP recognition. Replaced unsupported quoted paraphrases with cached primary evidence. |
| ARATH | AT5G52640 | file:ARATH/AT5G52640/AT5G52640-deep-research-perplexity.md | Perplexity deep research report on Arabidopsis HSP90.1 | MEDIUM | Read in full. Useful broad chaperone context, but conflates paralogs and species for client-specific claims. Primary evidence contradicts the exclusive near-native-client framing, direct PIF4 maturation and unqualified single-gene embryonic essentiality. |
| ARATH | CIPK24 | PMID:14730064 | Calcium sensors and their interacting protein kinases: genomics of the Arabidopsis and rice CBL-CIPK signaling networks. | MEDIUM | CBL-CIPK network genomics; source of CBL1/CBL9 IntAct pairs for CIPK24. Cached record notes an Expression of Concern (2024) and an Erratum (2025); used only for family/interaction context. |
| ARATH | CLV3 | PMID:21499263 | Stem-cell-triggered immunity through CLV3p-FLS2 signalling. | MEDIUM | Full text available; reports CLV3p triggers FLS2/BAK1 immune signaling and binds FLS2 directly. The central claim was contradicted by Segonzac et al. 2012 (PMID:22923673), who could not reproduce CLV3p recognition by FLS2. Citation is correct but the immunity function is contested. |
| ARATH | FLS2 | PMID:21499263 | Stem-cell-triggered immunity through CLV3p-FLS2 signalling. | HIGH | Original CLV3p-FLS2 claim is challenged by the independent primary experiments in PMID:22923673. |
| BACSU | spo0J | file:BACSU/spo0J/spo0J-deep-research-falcon.md | Deep research on spo0J (ParB) in Bacillus subtilis | HIGH | Read full body. Correct clamp, Smc recruitment and Soj regulation synthesis. Direct target PMID:31649139 supersedes homolog-only CTP examples; its CTPgammaS loading result requires separation of binding-driven loading from the proposed hydrolysis-dependent recycling mechanism. |
| CANAL | LPL1 | file:CANAL/LPL1/LPL1-hypotheses/function-hypothesis-go-0047372/openscientist.md | OpenScientist hypothesis investigation - LPL1 monoacylglycerol lipase activity | HIGH | OpenScientist autonomous-compute report directly evaluates the contested IBA monoacylglycerol lipase assignment, and the selected snippets support the wrong-subfamily rationale. However, its claim that QuickGO already lacks GO:0047372 for Q5AMS2 and that PAINT restricts the term to PTN000773838 was not reproducible against the fetched GOA in this repo, which shows the live annotation via PANTHER:… |
| DANRE | opa1 | file:worm/eat-3/eat-3-hypotheses/microtubule-and-peroxisome-capacities/openscientist.md | OpenScientist EAT-3 microtubule/peroxisome capacity adjudication (comparative OPA1-family evidence) | MEDIUM | Actual PANTHER topology independently confirmed: both nodes are ancestors of the OPA1 clade, not exclusive DRP1/classical-dynamin lineages. Report has no delivered computation code; domain/topology differences do not establish loss of every conditional function. Full primary PMID:32228866 explicitly identifies an OPA1 middle domain and C-terminal GED; the report's inferred GED absence is therefor… |
| DANRE | tomt | PMID:10526320 | Defective calmodulin-dependent rapid apical endocytosis in zebrafish sensory hair cell mutants. | HIGH | Correct target mutant/reference; only abstract available. The assay-to-endocytosis inference is unresolved in light of primary channel-permeation studies, not dismissed by title or gene-name inference. |
| DAPPU | DpuGr29 | file:DAPPU/DpuGr29/DpuGr29-hypotheses/daphnia-courtship-and-ligand-gated-channel-function/openscientist.md | OpenScientist focused assessment of DpuGr29 courtship and ligand-gated channel function | HIGH | Complete report, both CSVs and HTML/PDF variants read. Retains useful BmGr9 mechanism and motif lead but rejects courtship from missing target assays and donor species without inspecting topology. Actual target descent supports retention. Mislabels FBgn0032416, calls the donor support one study, double-counts preprint PMID:38187590 and final PMID:38573859, and conflates ORCO absence with GR subun… |
| DESVH | aprA | file:DESVH/Q72DT2/Q72DT2-hypotheses/function-hypothesis-go-0005886/openscientist.md | OpenScientist AprA plasma-membrane localization hypothesis report | HIGH | Incorporated its primary lead PMID:23842468 and the peripheral Qmo-associated pool. The recommendation that peripheral association cannot support GO:0005886 is contradicted by the term definition and ontology ancestry. Absence of a TM helix excludes integral insertion, not this broad compartment. The claimed sequence-analysis CSV is not among the two delivered artifacts and its numbers are not ad… |
| DICDI | acgA | file:DICDI/acgA/acgA-hypotheses/function-hypothesis-go-0001653/openscientist.md | OpenScientist focused acgA GO:0001653 hypothesis report | HIGH | The report omits the primary SDF-1/ACG connection in PMID:21602484 and treats IBA-only provenance and osmosensing as refutation. Its own caveat acknowledges no negative peptide-binding assay. Its CHASE specificity inference is family-level rather than an ACG negative binding experiment. |
| DICDI | carD | file:DICDI/carD/carD-hypotheses/function-hypothesis-go-0007189/openscientist.md | OpenScientist assessment of CAR4 adenylate-cyclase-activating signaling | HIGH | Primary assay scope and the explicit unpublished 1997 observation were verified. Report identifies no target-specific loss, yet treats absence of a CAR4 assay and GSK3 antagonism as paralog carry-over. Its PAINT narrative ignores the actual deep ancestral IBD at PTN000560943. Retain the IBA as non-core without upgrading evidence type. |
| DICDI | dhkA | PMID:12796307 | Genetic interactions of the E3 ubiquitin ligase component FbxA with cyclic AMP metabolism and a histidine kinase signal… | HIGH | Solid genetic evidence linking dhkA to RegA/cAMP and sporulation. The authors' proposed directionality (that DhkA activates RegA phosphodiesterase) explicitly contradicts Wang et al. (PMID:10373524 / PMID:15897458), who show ligand binding to DhkA inhibits phosphorelay to RegA; the direction of DhkA-RegA regulation is therefore contested. Hand-curated expert assessment - the Wang/Anjard-Loomis in… |
| DICDI | pdsA | file:DICDI/pdsA/pdsA-hypotheses/extracellular-pde-and-camp-pathway-regulation/openscientist.md | PdsA/DdPDE1 (DICDI, UniProt P12019) — Adjudication of two negative-regulation GO claims | HIGH | Pool distinction and feedback biology are supported, but direct-only restriction on a regulation BP is unjustified; report itself admits no PdsA-specific PKA test or verified absence of a glucose-GPCR pathway. Fungal source papers not full-text read. Claimed GO:1902660 annotation is absent from this review and not used. PTN002001416 is a node in PTHR28283, not a family identifier. Directionality… |
| DICDI | pten | file:DICDI/pten/pten-hypotheses/nuclear-localization-and-cell-cycle/openscientist.md | Focused OpenScientist assessment of Dictyostelium PTEN nuclear localization and cell-cycle regulation | HIGH | Full substantive report read. It incorrectly calls GO:0051726 absent, despite the source IBA and PTN000959472 IBD; equates nonnuclear-positive imaging with nuclear exclusion; and treats cytokinesis as outside cell-cycle scope. Its low-confidence K289-to-E pairwise alignment has no delivered code/alignment output and does not establish a loss phenotype in Dictyostelium. The cited human K289E defec… |
| DICDI | rasC | file:DICDI/rasC/rasC-hypotheses/function-hypothesis-go-0044351/openscientist.md | OpenScientist focused rasC GO:0044351 hypothesis report | HIGH | The assertion of no RasC deletion/uptake evidence contradicts PMID:15878331. Failure to rescue RasG-dependent suspension growth is not identical to a direct negative macropinocytic mechanism assay. Reported pairwise sequence identities do not locate a functional gain/loss on the PAINT tree, and no executed alignment artifact was delivered. |
| DICDI | regA | file:DICDI/regA/regA-hypotheses/function-hypothesis-go-0047555/openscientist.md | OpenScientist focused regA GO:0047555 hypothesis report | HIGH | The report omits PMID:34063491 and does not reconcile the greater-than selectivity bound in PMID:12429832 with the later ~200 claim. Another enzyme performing cGMP turnover does not establish its absence in RegA. Artifact evidence-tier plots are hardcoded summaries, not a new enzymatic or phylogenetic analysis. |
| DICDI | statA | file:DICDI/statA/statA-hypotheses/proliferation-and-defense-response/openscientist.md | OpenScientist assessment of STATa proliferation regulation and defense response | HIGH | Read complete substantive report. Correctly distinguishes proliferation from differentiation and defense from osmotic stress; correctly identifies STATb among ancestral evidence. Rejection nevertheless rests on donor taxonomic composition, alternative pathway genes and missing target assays. It explicitly acknowledges untested target growth/defense and establishes no STATa-specific loss. |
| DICDI | statC | file:DICDI/statC/statC-hypotheses/proliferation-and-defense-response/openscientist.md | AIGR Gene Hypothesis Deep Research — Dictyostelium discoideum statC (Q54BD4) | HIGH | Primary source leads and noncanonical stress activation are supported. Refuted verdict is not: report admits statC proliferation and bacterial-clearance roles are untested, treats paralog descendants as automatic propagation error, and adds JAK/cytokine requirements absent from the broad GO definitions. TirA/NADPH oxidase necessity does not exclude a STAT transcriptional regulator. No target-spec… |
| DROME | Dic4 | file:DROME/Dic4/Dic4-hypotheses/fly41-thiamine-pyrophosphate-transport/openscientist.md | OpenScientist assessment of Dic4 thiamine pyrophosphate transport | HIGH | Read substantive findings and limitations. The primary negative reconstitution result is relevant, but the report also acknowledges untested ThPP and unresolved folding/assay conditions. Its refuted/paralog-carry-over verdict overstates absent experiments and a coarse identity-based tree; a characterized Tpc1 does not exclude another carrier with overlapping capacity. |
| DROME | Hsp23 | file:DROME/Hsp23/Hsp23-deep-research-falcon.md | Falcon deep research report on Hsp23 (Drosophila melanogaster) | HIGH | Full narrative, table and references read. Its suggestion that direct Hsp23 biochemistry remains a gap overlooks purified-client PMID:16572729. Cytoplasmic localization is established, but the report omits positive 1981/1986 conditional nuclear imaging. Its 2016 muscle-protection lead is verified in full PMID:27483356: perinuclear ubiquitin puncta are clients/aggregates, not proof of Hsp23 nuclea… |
| HETGA | Has2 | PMID:33846452 | Abundance and size of hyaluronan in naked mole-rat tissues and plasma. | HIGH | Full text cached and read. Correctly cited and methodologically careful (two independent sizing methods, quantitative cortex/medulla separation), but marked DISPUTED because its central size result directly contradicts PMID:23783513 and the discrepancy is unresolved - PMID:34476892 attributes it to isolation procedure. That dispute concerns polymer size only; the renal cortex/medulla quantificati… |
| NICAT | NaUGT1_candidate_UGT85A2_0 | file:NICAT/NaUGT1_candidate_UGT85A2_0/NaUGT1_candidate_UGT85A2_0-deep-research-falcon.md | Deep research report on NaUGT1/UGT85A2_0 (Falcon/Edison Scientific Literature) | MEDIUM | Earlier uncertainty about accession mapping is superseded by primary supplemental sequence evidence from the two 2026 papers. The Cell NaUGT1 assay primers identify the g26396 locus, with an extended N-terminal construct; exact activity of the isolated 485-aa database form remains untested. The tobacco paper is Nature Communications, distinct from the independent Cell paper. |
| NICAT | NaUGT1_candidate_UGT85A2_0 | file:NICAT/NaUGT1_candidate_UGT85A2_0/NaUGT1_candidate_UGT85A2_0-hypotheses/function-hypothesis-go-0080043/falcon.md | Existing falcon function-hypothesis report | HIGH | The proposed discriminating substrate panel acknowledges unresolved acceptor specificity. Its absence-of-characterized-UGT85-quercetin argument supports caution, not categorical exclusion. Neither inferred phylogenetic distance nor high predicted structure confidence proves substrate specificity. The family-only and absent-primary-data framing is superseded by the 2026 NaUGT1 gene-level assays an… |
| NICAT | NaUGT1_candidate_UGT85A2_0 | file:NICAT/NaUGT1_candidate_UGT85A2_0/NaUGT1_candidate_UGT85A2_0-hypotheses/function-hypothesis-go-0080043/openscientist.md | Blinded OpenScientist function-assignment report (TreeGrafter audit) | HIGH | The report supports UGT85-family identity but explicitly admits unknown target substrate specificity and broad UGT85 diversity. Its headline refutation and claim that sequence identity gaps are incompatible with shared substrate specificity are too strong. No duplicate quercetin run is needed; the nicotine orthology hypothesis is distinct. The family-only and absent-primary-data framing is supers… |
| NICAT | NaUGT1_candidate_UGT85A2_0 | file:NICAT/NaUGT1_candidate_UGT85A2_0/NaUGT1_candidate_UGT85A2_0-hypotheses/naugt1-candidate-identity-and-nicotinate-glucosylation/openscientist.md | OpenScientist adjudication of NaUGT1 identity and nicotinate glucosylation | HIGH | Read the entire report and computational artifact. Its decisive O-versus-N clade exclusion and non-pathway/storage-branch argument are contradicted by PMID:42151135 and PMID:41928514. The very tobacco accession the report calls an O-glucosyltransferase is explicitly assayed as UGT1/NaGT in the primary supplement. Automated names and missing annotations were treated as experimental proof. Its no-k… |
| POPTR | ndhB1 | file:POPTR/ndhB1/ndhB1-uniprot.txt | UniProtKB reviewed entry for ndhB1 | HIGH | Localization and family identity are useful; the NAD(P)H electron-donor description is not accepted as current mechanistic evidence. See PMID:21505067, PMID:28559282 and PMID:39856350. |
| POPTR | ndhB2 | file:POPTR/ndhB2/ndhB2-uniprot.txt | UniProtKB reviewed entry for ndhB2 | HIGH | Localization and family identity are useful; the NAD(P)H electron-donor description is not accepted as current mechanistic evidence. See PMID:21505067, PMID:28559282 and PMID:39856350. |
| POPTR | ndhD | file:POPTR/ndhD/ndhD-uniprot.txt | UniProtKB reviewed entry for ndhD | HIGH | Localization and family identity are useful; the NAD(P)H electron-donor description is not accepted as current mechanistic evidence. See PMID:21505067, PMID:28559282 and PMID:39856350. |
| POPTR | ndhK | file:POPTR/ndhK/ndhK-uniprot.txt | UniProtKB reviewed entry for ndhK | HIGH | Localization and family identity are useful; the NAD(P)H electron-donor description is not accepted as current mechanistic evidence. See PMID:21505067, PMID:28559282 and PMID:39856350. |
| PSEPK | PP_1451 | file:projects/P_PUTIDA/deep-research/PSEPK__gram_negative_lipoprotein_peptidoglycan_tether_remodeling__ppu00550-deep-research-openscientist.md | OpenScientist PSEPK tether-remodeling pathway report | MEDIUM | Its candidate-uncertain conclusion relied on low-sensitivity k-mer and Smith-Waterman comparisons and is superseded here by exact PTHR36699:SF1 membership plus the PAINT evidence anchored by experimental DpaA/P0AA99. |
| PSEPK | alg8 | file:projects/P_PUTIDA/deep-research/PSEPK__alginate-polymerization-export__ppu00543-deep-research-openscientist.md | OpenScientist PSEPK alginate polymerization and export synthesis | HIGH | Useful operon/mechanistic synthesis. Its evidence-tier description overstates PMID:17601783 as direct KT2440 evidence: the primary abstract specifies P. putida mt-2. PMID:25968647 assays P. aeruginosa. These support related-strain/species transfer, not direct Q88NC5 biochemistry. |
| PSEPK | davB | file:PSEPK/davB/davB-uniprot.txt | UniProtKB entry Q88QV1 for Pseudomonas putida KT2440 davB | HIGH | The accession and gene identity are verified, but the automated tryptophan/auxin function conflicts with KT2440 Dav-pathway evidence. |
| PSEPK | ilvA-I | file:PSEPK/ilvA-I/ilvA-I-deep-research-openscientist.md | OpenScientist functional report for PSEPK IlvA-I | HIGH | Core long-form IlvA assignment and comparative analysis are useful for both paralogs. Do not transfer the precise 23-angstrom E. coli allosteric relay as measured in KT2440 or date duplication from 75% identity alone. The report incorrectly adds water to the net threonine deamination reaction and conflates its 2-aminocrotonate intermediate with serine-derived 2-aminoacrylate; target RHEA:22108 an… |
| PSEPK | ilvA-II | file:PSEPK/ilvA-I/ilvA-I-deep-research-openscientist.md | OpenScientist functional report for PSEPK IlvA-I | HIGH | Core long-form IlvA assignment and comparative analysis are useful for both paralogs. Do not transfer the precise 23-angstrom E. coli allosteric relay as measured in KT2440 or date duplication from 75% identity alone. The report incorrectly adds water to the net threonine deamination reaction and conflates its 2-aminocrotonate intermediate with serine-derived 2-aminoacrylate; target RHEA:22108 an… |
| PYROR | PoMZ_10221 | file:PYROR/PoMZ_10221/PoMZ_10221-hypotheses/rhamnose-enzyme-and-mat-regulatory-capacity/openscientist.md | Focused Curation Report — Pyricularia oryzae PoMZ_10221 / F8U970 | HIGH | Read full delivered report. Correct target and sugar-enzyme evidence; reject the jump from no assay/another catalytic gene/low identity to refuted MAT regulation. The report admits it did not recover exact graft logic. Only HTML/PDF were delivered, so its claimed computed identity percentages are not reproduced primary evidence. No duplicate provider run requested. |
| SACEN | eryCI | file:SACEN/eryCI/eryCI-uniprot.txt | UniProtKB P14290 (eryCI) record | HIGH | The record's FUNCTION ("sensor protein... signal of environmental..."), name ("sensory transduction protein"), and keywords (DNA-binding, Two-component regulatory system, Cell membrane) are a legacy misannotation contradicted by the protein's own DegT/DnrJ/EryC1 family membership (sugar aminotransferases) and PLP keyword, and by the GOA transaminase annotation. Used here only for the (correct) fa… |
| SCHPO | Epe1 | file:SCHPO/Epe1/Epe1-hypotheses/function-hypothesis-go-0032452/openscientist.md | OpenScientist hypothesis investigation - Epe1 histone demethylase activity | HIGH | The report directly evaluates the contested GO:0032452 IBA and its domain/sequence analysis supports removal. However, the report also claims a QuickGO/PomBase NOT annotation on a child demethylase term that is not present in the local fetched GOA file, which instead contains positive PomBase GO:0032454 lines already reviewed separately in this YAML. The cited YAML snippets therefore use only ver… |
| SCHPO | SPAC24C9.08 | file:SCHPO/cps1/cps1-deep-research-falcon.md | Deep research on SPAC24C9.08 M20A metallopeptidase (Falcon) | HIGH | Read report critically. Its suggestion that SPAC24C9.08 localization is unestablished/likely cytosolic conflicts with source PMID:16823372 and explicit target-routing experiments PMID:17660439. M20-family substrate diversity does not establish loss on the fungal CPS branch. |
| SCHPO | pmp20 | file:SCHPO/pmp20/pmp20-deep-research-falcon.md | Falcon (Edison) deep research report on S. pombe pmp20 (O14313) function, localization, and peroxiredoxin-family context | HIGH | Read completely. Comparative Candida/Hansenula/mammalian redox functions are useful background but do not establish target peroxidase activity or peroxisomal location. The target primary negative enzymology takes precedence. OpenAI also wrongly discounts the target nuclear observation without experimental justification; Falcon recognizes holdase but overstates conserved peroxisomal detoxification. |
| SCHPO | pmp20 | file:SCHPO/pmp20/pmp20-deep-research-openai.md | openai research on S. pombe Pmp20 | HIGH | Read completely. Comparative Candida/Hansenula/mammalian redox functions are useful background but do not establish target peroxidase activity or peroxisomal location. The target primary negative enzymology takes precedence. OpenAI also wrongly discounts the target nuclear observation without experimental justification; Falcon recognizes holdase but overstates conserved peroxisomal detoxification. |
| SCHPO | pmp20 | file:SCHPO/pmp20/pmp20-hypotheses/function-hypothesis-go-0008379/openscientist.md | OpenScientist focused review of pmp20 thioredoxin peroxidase activity | HIGH | Read the complete report. Both reports incorrectly state that GSH-dependent activity was untested; primary Results p171 directly contradicts this. Single-Cys architecture does not by itself imply inactivity, but a modeled active-site tetrad also cannot establish catalytic turnover. The first report supports the specific Trx rejection and actual holdase assay; the second raises a valid partner-dep… |
| SCHPO | pmp20 | file:SCHPO/pmp20/pmp20-hypotheses/prediction-peroxiredoxin-activity/openscientist.md | OpenScientist focused review of pmp20 broader peroxiredoxin prediction | HIGH | Read the complete report. Both reports incorrectly state that GSH-dependent activity was untested; primary Results p171 directly contradicts this. Single-Cys architecture does not by itself imply inactivity, but a modeled active-site tetrad also cannot establish catalytic turnover. The first report supports the specific Trx rejection and actual holdase assay; the second raises a valid partner-dep… |
| SCHPO | rqh1 | PMID:7623848 | An alternative eukaryotic DNA excision repair pathway. | MEDIUM | The paper correctly reports the original rad12-502/SPDE phenotype, but the later identification of rad12 as rqh1 explicitly failed to reproduce a UV-dimer-endonuclease defect (PMID:9372918). |
| SCHPO | trm4b | PMID:23074192 | Pmt1, a Dnmt2 homolog in Schizosaccharomyces pombe, mediates tRNA methylation in response to nutrient signaling. | HIGH | The 2019 full-text paper (PMID:30646830) reports that the earlier purported trm4b deletion strain lacked the deletion. The older Trm4b-dependent tRNA-Asp site assignment cannot establish physiological specificity. |
| SCHPO | vms1 | PMID:29632312 | Vms1 and ANKZF1 peptidyl-tRNA hydrolases release nascent chains from stalled ribosomes. | HIGH | Full text cached. Correctly cited, and the ribosome-release biology and catalytic-glutamine requirement stand, but the paper's peptidyl-tRNA hydrolase mechanism was superseded by PMID:31011209, which showed the enzyme cleaves the tRNA 3'-CCA rather than the peptidyl-tRNA ester bond. Cite it for the clade and the catalytic residue, not for the reaction chemistry. |
| WHEAT | RHT1 | file:WHEAT/RHT1/RHT1-deep-research-falcon.md | Deep-research report (falcon / Edison Scientific Literature) - functional annotation of wheat RHT-1 / RHT1 (Q9ST59). | MEDIUM | Useful wheat DELLA background, checked against primary PMID:27327160, PMID:33422695 and PMID:41615756. TaGID2 localization is not a direct RHT1 localization assay. Growth repression does not establish that RHT1 only acts as a transcription corepressor; the 2026 primary study leaves partner-specific transcriptional effects unresolved. |
| human | A1BG | file:human/A1BG/A1BG-hypotheses/membrane-receptor-and-growth-hormone-capacities/openscientist.md | OpenScientist focused assessment of A1BG membrane, receptor and growth-hormone capacities | HIGH | Full report, HTML/PDF, computational provenance and plot read. Canonical hydropathy result reproduced; useful GH expression-versus-participation distinction retained. Categorical REMOVE/NOT leads are not justified: mouse surface localization is missed, NAMPT compartment/directness overstated, topology explicitly not traced, and claimed complete GH-paper access lacks independently checkable proven… |
| human | A2M | PMID:12538697 | Natural substrates and inhibitors of mannan-binding lectin-associated serine protease-1 and -2: a study on recombinant… | MEDIUM | The cited experiment supports soluble MASP-1 inhibition, but its pathway-level physiological extrapolation is contradicted by PMID:23399388, which found no abolition of surface lectin-pathway activation in human serum. |
| human | A3GALT2 | PMID:18630988 | Humans lack iGb3 due to the absence of functional iGb3-synthase: implications for NKT cell development and transplantat… | HIGH | Direct full-text human expression and functional study, but its categorical inactivity conclusion is contested by PMID:23378701. Functional testing used rat/human chimeras rather than native full-length human protein. |
| human | A3GALT2 | file:human/A3GALT2/A3GALT2-hypotheses/human-enzyme-constructs-and-glycan-substrate-specificity/openscientist.md | Human A3GALT2 enzyme constructs and glycan substrate specificity: focused OpenScientist report | HIGH | Full report and four artifacts read. Its pivotal claim that PMID:23378701 contains no human positive experiment was inferred from the pig-focused abstract and is contradicted by full Methods, Figure 6 and Table 1. It also overlooks Figure 5 reverse-mutant experiments in PMID:18630988. Q3V1N9 is mouse, not rat: human U3KPV4 N253 aligns to rat A0A4Z3 Y252 and mouse Q3V1N9 Y283 in sequence version 1… |
| human | AADAC | Reactome:R-HSA-5689000 | AADAC deacetylates PHEN | MEDIUM | The event supports phenacetin deacetylation. Its model places the small-molecule participants in cytosol while the AADAC catalytic bulk is lumen-facing at the ER, so the reaction geometry is internally inconsistent even though the activity is correct. |
| human | AAK1 | PMID:18657069 | AAK1 regulates Numb function at an early step in clathrin-mediated endocytosis. | HIGH | The cached abstract supports the reported Numb localization and cargo-selective phenotypes, but the central Numb-T102 substrate claim is contradicted by PMID:42082516: purified AAK1/BMP2K did not phosphorylate the site, RPE phosphoproteomics did not detect it, and its sequence lacks the refined recognition motif. Preserve the localization phenotype; do not treat T102 phosphorylation as establishe… |
| human | AARS1 | PMID:40156251 | Yeast models for Charcot-Marie-Tooth disease-causing aminoacyl-tRNA synthetase alleles reveal the cellular basis of dis… | MEDIUM | The PubMed identity and recovered full PMC11953622 review were checked. Its AARS1 section usefully distinguishes yeast ALA1 complementation, human cytosolic function and allele-dependent assay limitations. It is a review, not a new direct human activity experiment. The generalization that human AlaRS is predominantly monomeric under normal cellular conditions is not adopted: the original purified… |
| human | AARS1 | PMID:41639505 | The role of Alanyl-tRNA synthetase 1 lactylation in tumors and other diseases. | MEDIUM | Primary PubMed identity and full PMC12961078 review inspected. Bibliographic identity is correct, but the disease-mechanism narrative is not used as independent evidence: Figure 1 and the body expand ALSP differently, and its detailed proofreading/ER-stress disease claims require their original experimental sources. The newly cached biallelic series PMID:34446925 is independently scoped and is no… |
| human | ABCD4 | PMID:9302272 | Identification of a fourth half ABC transporter in the human peroxisomal membrane. | HIGH | The original peroxisome claim conflicts with later full-length human ABCD4 localization and targeting evidence (PMID:19010322, PMID:27456980). The latter paper explicitly discusses the early antibody-based rat hepatic result; this is not a wrong-identifier or wrong-gene allegation. |
| human | ABHD5 | PMID:18606822 | CGI-58, the causative gene for Chanarin-Dorfman syndrome, mediates acylation of lysophosphatidic acid. | HIGH | Identifier/title verified against primary PubMed and indexed full PMC3259832 Methods/Results. Human His-tagged protein in BL21(DE3), nickel purification, PA assays and yeast overexpression were read. Its intrinsic LPAAT interpretation is contradicted by PMID:24879803; human negative lipase/phospholipase observations and mouse WAT localization evidence remain distinct. Local cache is abstract-only. |
| human | ABHD8 | file:human/ABHD8/ABHD8-bioinformatics/RESULTS.md | What are ABHD8's IBA source proteins, and does any of them justify a lipid activity? | HIGH | Source mapping is useful, but donor counts and absent target experiments do not negate an IBD. Its assertion of established ABHD5 LPAAT activity is contradicted by PMID:24879803; plant CGI-58 activities are also challenged by PMID:26745266. The intact target catalytic triad and adaptor role do not establish hydrolase loss. |
| human | ABI3BP | PMID:18559958 | Re-expression of ABI3-binding protein suppresses thyroid tumor growth by promoting senescence and inhibiting invasion. | HIGH | Cached title matches and the paper is correctly cited; the growth-suppression result is used here and is consistent with PMID:40092729 and PMID:31174563 in other human tissues. Marked DISPUTED for the senescence claim specifically: this paper reports that gaining ABI3BP causes senescence while PMID:19338757 reports that losing it causes senescence, and three further papers report that losing it r… |
| human | ABI3BP | PMID:19338757 | Implication of p53-dependent cellular senescence related gene, TARSH in tumor suppression. | MEDIUM | Cached title matches; abstract-only; mouse embryonic fibroblasts. Correctly cited, but marked DISPUTED because its proliferation result contradicts the mouse mesenchymal stem cell knockout and knockdown data in PMID:23666637 and PMID:25296984, and its senescence direction contradicts three later knockdown studies. Used only to document the conflict; no annotation rests on it. |
| human | ABI3BP | PMID:38812032 | ABI3BP promotes renal aging through Klotho-mediated ferroptosis. | LOW | Cached title matches and full text is available; mouse knockout plus human HK2 cells. Correctly cited, but marked DISPUTED because its senescence direction is opposite to PMID:19338757 and PMID:18559958. Cited to document the conflict; no annotation rests on it, and neither ferroptosis nor ageing terms are proposed. |
| human | ABI3BP | PMID:40889718 | ABI3BP deficiency alleviates Ang II-induced VSMC senescence through the Nrf2 signalling pathway. | LOW | Cached title matches; abstract-only; mouse knockout. Correctly cited, but marked DISPUTED because the senescence direction opposes PMID:19338757 and PMID:18559958. Documents the conflict only. |
| human | ABRAXAS1 | file:human/ABRAXAS1/ABRAXAS1-hypotheses/function-hypothesis-go-0008017/openscientist.md | OpenScientist hypothesis report for ABRAXAS1 GO:0008017 | HIGH | Provenance tracing to PTN001272083 and ABRAXAS2 is useful and agrees with the PAINT cache. The conclusion of refutation relies on primary nuclear localization, missing ABRAXAS1 assays and a single paralog source, none of which establishes loss of an ancestral function. The 2015 BRISC paper confirms the source mechanism but does not supply a direct negative ABRAXAS1 comparison. |
| human | ACTL8 | file:human/ACTL8/ACTL8-bioinformatics/RESULTS.md | ACTL8: does the actin fold come with actin's residues, and do its IBA sources transfer? | HIGH | The reproducible sequence/contact calculations and fetched annotation-recipient sets are useful. Inferring a misplaced tree node from recipient membership, alignment identity or donor count is not a phylogenetic reconstruction. The ACTR3 control demonstrates that low canonical-contact conservation does not exclude all structural filament roles. |
| human | ACTR5 | PMID:36563143 | ACTR5 controls CDKN2A and tumor progression in an INO80-independent manner. | HIGH | The only study to assay ACTR5 itself at a promoter, and the basis for the proposed GO:0000122 row. Two caveats recorded rather than hidden. It carries a 2025 erratum (PMID:41071901) correcting Fig. 2D - the CDKN2A western - because the ACTR5 and U87 panels duplicated those of Fig. 3A; the publisher states the conclusions are unaffected, and the RNA-seq, ChIP-qPCR and H3K9me2 data are independent… |
| human | ADAMTSL1 | PMID:30714143 | ADAMTSL1 and mandibular prognathism. | LOW | The genetic association is reported without functional validation, and the proposed mechanism attributes proteolysis to a protein that has no metalloprotease domain, no catalytic activity comment and no EC number. Cited here as an example of the family-name-to-activity conflation, not as evidence for any annotation. |
| human | ADCK1 | PMID:33824271 | ADCK1 activates the β-catenin/TCF signaling pathway to promote the growth and migration of colon cancer cells. | LOW | PubMed-verified, no retraction or erratum, and the co-immunoprecipitation itself is reported at endogenous levels. Marked DISPUTED rather than MISCITED because the citation is accurate and the experiment real, but the conclusion is in tension with every localisation measurement made on this protein and has not been reproduced. Raised as a question rather than acted on. |
| human | ADCK2 | PMID:38425362 | In vitro construction of the COQ metabolon unveils the molecular determinants of coenzyme Q biosynthesis. | MEDIUM | Correctly cited and methodologically substantial, but its central claim about protein kinase activity stands in direct contradiction to PMID:27499294 and UniProt records it only as "may act as a protein kinase". Included here specifically so that this review does not assert an unqualified family-level negative that the literature no longer supports. |
| human | ADIPOQ | PMID:24531262 | Adiponectin reduces thermogenesis by inhibiting brown adipose tissue activation in mice. | HIGH | Correctly cited for GO:0120163 (adiponectin suppresses thermogenesis) and incorrectly cited for the opposite term GO:0120162. The scientific claim is also genuinely disputed: PMID:26166748 reports the opposite direction in the same knockout model. Both papers are attached to both terms in GOA and in MGI. |
| human | ADIPOQ | PMID:26166748 | Adiponectin Enhances Cold-Induced Browning of Subcutaneous Adipose Tissue via Promoting M2 Macrophage Proliferation. | HIGH | Mirror image of PMID:24531262: correctly cited for GO:0120162 (adiponectin promotes cold-induced browning) and incorrectly cited for GO:0120163. Also the second independent report of the adiponectin-T-cadherin interaction, which GO records nowhere. |
| human | ADM5 | file:genes/human/ADM5/ADM5-deep-research-affinage.md | Affinage mechanistic annotation for ADM5 (human) | HIGH | Automated trust gates passed (accession matches C9JUS6, no non-human organism token, pairwise self-evaluation 'win'), and its nine citations are all real and correctly summarised - precision is not the problem. The problem is framing. Every functional experiment it cites is in pufferfish, medaka, Xenopus, pig, rat or sheep, yet the narrative is written as a description of the human gene and never… |
| human | ADPRS | PMID:17015823 | The structure of human ADP-ribosylhydrolase 3 (ARH3) provides insights into the reversibility of protein ADP-ribosylati… | HIGH | The 1.6 A ARH3 structure, and the source of the micromolar ADP-ribose binding measurement. Its statement that ARH3 'efficiently de-ADP-ribosylates poly- but not monoADP-ribosylated proteins' was superseded in 2017: ARH3 is the serine mono-ADP-ribosylhydrolase. The 2006 assay used arginine-linked mono-ADP-ribose, on which ARH3 is genuinely inactive, so the two results are compatible once the subst… |
| human | AEBP2 | PMID:15225548 | SUZ12 is required for both the histone methyltransferase activity and the silencing function of the EED-EZH2 complex. | MEDIUM | Source of the long-standing view that AEBP2 stimulates PRC2, reported here as a minimal complex of EZH2, EED and SUZ12 with AEBP2 required for optimal activity. PMID:41168462 reports that prior in vitro stimulation was obtained with the short isoform or with N-terminally truncated constructs, and that the broadly expressed long isoform inhibits instead. Marked DISPUTED for the direction of the ef… |
| human | AEBP2 | PMID:29681498 | Distinct Stimulatory Mechanisms Regulate the Catalytic Activity of Polycomb Repressive Complex 2. | MEDIUM | Reports that AEBP2 stimulates both PRC2-EZH1 and PRC2-EZH2 by a mechanism independent of and additive to allosteric activation. Like PMID:15225548 this is disputed in direction by PMID:41168462, which attributes prior stimulation results to use of the short isoform or of N-terminally truncated protein. Cited here for the mechanism it describes, not as support for a stimulatory annotation. |
| human | AFF1 | PMID:17135274 | The mixed-lineage leukemia fusion partner AF4 stimulates RNA polymerase II transcriptional elongation and mediates coor… | MEDIUM | PubMed-verified, and cited here ONLY for what it actually shows: a mouse Af4 result. It carries an unflagged 2023 Erratum (PMID:37777189), and its DOT1L-recruitment claim rests on an ENL co-immunoprecipitate that PMID:20159561 reports was misinterpreted. Nothing in this review asserts DOT1L recruitment by AFF1. |
| human | AFF1 | PMID:28955517 | AFF1 and AFF4 differentially regulate the osteogenic differentiation of human MSCs. | HIGH | PubMed-verified. Carries an unflagged 2020 Correction (PMID:32257529) whose content is not stated in its abstract or its PMC record, and one sentence in the overexpression section says "AFF1-depleted" where the figure describes overexpressing cells; both are recorded in the notes and the claims used avoid that sentence. LOW_QUALITY is not the right flag - the experiments are sound and bidirection… |
| human | AGFG1 | PMID:18819912 | Role of HRB in clathrin-dependent endocytosis. | HIGH | Supports the clathrin-dependent-endocytosis role by knockdown, and is one of only two papers the affinage record returned. Marked DISPUTED because its generality is contested rather than because the paper is unsound: it reports transferrin uptake strongly reduced on HRB knockdown, while PMID:18775314 reports no effect of HRB depletion on EGF internalisation or degradation and reads AGFG1 as a SNA… |
| human | AGFG2 | PMID:21284487 | Organization patterns of the AGFG genes: an evolutionary study. | LOW | The mammals-only claim conflicts with the broader AGFG phylogeny in PMID:23433073. This is a claim-specific disagreement, not a judgment that the entire study is low quality; the sequence-symbol census is corroboration rather than an orthology proof. |
| human | AGFG2 | file:human/AGFG2/AGFG2-deep-research-affinage.md | Affinage mechanistic annotation for AGFG2 (human) | MEDIUM | Full body and table read. Useful dual Nef/Vpu narrative agrees with primary evidence. The table contradicts that narrative, detailed binding/localization is partly extrapolated from Hrb, the mammals-only statement conflicts with broader phylogeny, and the report omits the endothelial secretion study. |
| human | AGGF1 | PMID:14961121 | Identification of an angiogenic factor that when mutated causes susceptibility to Klippel-Trenaunay syndrome. | HIGH | The source of seven of AGGF1's twenty-four GOA rows, and the biochemistry it reports - secretion, endothelial binding, CAM angiogenesis, proliferation, TWEAK interaction - has been independently reproduced and extended. Marked DISPUTED rather than VERIFIED for one specific reason: its human-genetic claim, that E133K is a gain-of-function KTS mutation, was contradicted by PMID:16443853 and PMID:17… |
| human | AGGF1 | PMID:36696895 | Protein therapy of skeletal muscle atrophy and mechanism by angiogenic factor AGGF1. | MEDIUM | Not disputed as a citation - the paper says what it says - but its central mechanistic claim is contradicted in direction by PMID:39905000. Recording both, and annotating neither, is the honest treatment. No quote is taken from it. |
| human | AGK | file:human/AGK/AGK-hypotheses/function-hypothesis-go-0001729/openscientist.md | OpenScientist hypothesis report for AGK ceramide kinase activity | HIGH | Read and reused the existing report. Its human ceramide-kinase conclusion is supported by direct negative substrate assays. However, the report incorrectly presents the PAINT assertion as CERK-to-AGK propagation: PTN008994514 is supported by mouse Agk and fly Mulk. It also does not independently resolve ceramide/sphingosine biosynthetic-process annotations. Retain the biochemical observations and… |
| human | AGK | file:human/AGK/AGK-uniprot.txt | UniProt record for human AGK | HIGH | Identity, reactions, domains and localization checked. The C6-ceramide reaction RHEA:43312 is attributed to PMID:15939762, which instead reports a negative assay for C6-ceramide. The separate general ceramide reaction is an ortholog inference. Positive MAG/DAG and TIM22 information is supported independently; do not silently edit the machine record. |
| human | AGO1 | PMID:31330067 | Let-7a-regulated translational readthrough of mammalian AGO1 generates a microRNA pathway inhibitor. | MEDIUM | The original readthrough and AGO1x inhibitory-function study is directly challenged by PMID:40500330 and defended in PMID:40500329. The latter response was verified from PubMed and the publisher full text. Reporter-method disagreement does not by itself invalidate orthogonal proteomic/antibody evidence or all proteoform functions. No seeded GOA row cites this paper. |
| human | AGO1 | PMID:38499809 | Transcript-specific induction of stop codon readthrough using a CRISPR-dCas13 system. | MEDIUM | Full cached dCas13 readthrough-induction study uses AGO1 reporters and endogenous-protein assays. Its reporter interpretation is challenged in PMID:40500330 and defended in PMID:40500329. The criticism is scoped to readthrough measurement, not assumed to refute every dCas13 result or AGO1 function. |
| human | AGO1 | PMID:40500329 | Response to Suresh et al. | MEDIUM | PubMed independently verified the identifier, title and DOI 10.1038/s44318-025-00479-0; publisher full text was read. The normal fetch artifact recovered on 2026-09-27 now supplies XML full text from PMC12264181. The authors contest the critique using prior AGO1 reporter and orthogonal evidence and new HBB reporter comparisons with versus without stop-go sequences. The HBB experiment is not itsel… |
| human | AGO1 | PMID:40500330 | The dual luciferase assay does not support claims of stop codon readthrough on the AGO1 mRNA. | MEDIUM | The full cached methodological critique shows low signal in its insulated reporter and attributes earlier values to reporter effects. Its interpretation is explicitly contested by the co-published response PMID:40500329, which discusses reporter design and orthogonal evidence. This is an unresolved experimental disagreement, not a retraction. The earlier claim that no response existed is supersed… |
| human | AGO2 | PMID:17524464 | An mRNA m7G cap binding-like motif within human Ago2 represses translation. | HIGH | Identifier/title checked against the machine-cached primary record and original GOA citation; access scope is stated below. PubMed confirms the paper proposing a MID-domain cap-binding motif and cap-dependent repression. Direct cap recognition is contested by structural and recombinant-protein studies (PMID:19159466 and PMID:21475248); PMID:23409027 supplies later cap-proximity evidence in cellul… |
| human | AGO2 | PMID:25446899 | Cancer exosomes perform cell-independent microRNA biogenesis and promote tumorigenesis. | HIGH | Identifier/title checked against the machine-cached primary record and original GOA citation; access scope is stated below. PubMed links this original to the 2025 Expression of Concern PMID:40930611, DOI 10.1016/j.ccell.2025.07.022. The full notice was read at the MD Anderson ElsevierPure institutional record on 2026-09-27. It reports an institutional investigation of duplicated/relabelled Dicer,… |
| human | AGO4 | file:human/AGO4/AGO4-hypotheses/mrna-destabilization-versus-slicer-loss/openscientist.md | OpenScientist AGO4 mRNA destabilization versus slicer-loss adjudication | HIGH | Read complete report and CSV tables. Useful PMID:19838187 tethering evidence independently verified. Material error: report asserts no GO:0035279/GO:0090625 negative annotations, but live QuickGO retains both BHF-UCL NOT IDA rows. Definitions of both terms explicitly concern cleavage; related deadenylation synonyms for GO:0035279 do not automatically broaden its formal definition. Reject confiden… |
| human | AHNAK2 | PMID:31011849 | Linkage analysis and whole exome sequencing reveals AHNAK2 as a novel genetic cause for autosomal recessive CMT in a Ma… | MEDIUM | The genetic finding is sound and correctly cited. The disputed element is the background claim that AHNAK2 binds directly to periaxin: a Europe PMC search for AHNAK2 and periaxin returns nothing earlier than PMID:24675079, which reports domain-swapped homodimers of each protein separately rather than a heterodimer. No annotation in this review rests on the interaction claim; it is raised in sugge… |
| human | AIRE | PMID:14734522 | AIRE functions as an E3 ubiquitin ligase. | HIGH | Primary PMC Methods/Results and newly recovered full text inspected. Human AIRE cDNA supplies GST-PHD and full-length preparations; E1/E2/ubiquitin assays with GST/PHD2 controls report positive E3 activity. This is real experimental evidence, not only domain projection. PMID:15649886 reports failure to reproduce intrinsic PHD1 E3 activity/E2 interaction; retain the conflict without inventing cont… |
| human | AKAP12 | GO_REF:0000107 | Automatic transfer of experimentally verified manual GO annotation data to orthologs using Ensembl Compara | MEDIUM | The orthology itself is sound - Q5QD51 is reviewed rat Akap12 in the same single-gene PANTHER subfamily as human AKAP12, so these are ortholog and not paralog transfers. The donor annotations are mixed in quality, and the pipeline propagates them without distinction: three of the eleven transferred rows rest on expression correlation or, in the GO:0043116 case, invert the sign of the experiment t… |
| human | AKIRIN1 | file:human/AKIRIN1/AKIRIN1-deep-research-falcon.md | Research Report: Human AKIRIN1 (UniProt Q9H9L7; Akirin-1/Mighty) — functional annotation | HIGH | Full biological report and evidence-table artifact read. Useful promyogenic synthesis and explicit Akirin1 versus Akirin2 immune distinction. Positive proliferation is condition-specific, and the fibroblast perturbation species was checked in the primary paper. Association studies are not additional process evidence. |
| human | AKT1 | PMID:22797923 | ZNRF2 is released from membranes by growth factors and, together with ZNRF1, regulates the Na+/K+ATPase. | HIGH | Source identity was checked against its seeded IDA citation and the cached primary record. Full original PMC3500867 maps PKBalpha phosphorylation to ZNRF2 Ser19 by HPLC/Edman sequencing (supplementary Figure S2A). Publisher note DOI:10.1242/jcs.259936 reports duplicated bottom GST control blots in Figure 2B and unavailable original LICOR controls; alternate same-experiment controls were oversatur… |
| human | AKT1 | Reactome:R-HSA-199863 | AKT can phosphorylate NR4A1 (NUR77) | HIGH | The live Reactome event summary, participants and catalyst assignment were read at https://reactome.org/content/detail/R-HSA-199863. The event models nuclear AKT phosphorylation of NR4A1 but explicitly notes subsequent work favoring RSK/MSK as the physiological NR4A1 kinase. The physiological NR4A1-kinase assignment is disputed in the wild-type event; the general AKT kinase and nuclear-location a… |
| human | AKT1 | Reactome:R-HSA-2399988 | AKT1 E17K mutant phosphorylates NR4A1 (NUR77) | HIGH | The live Reactome event summary, participants and catalyst assignment were read at https://reactome.org/content/detail/R-HSA-2399988. The event predicts nuclear AKT1 E17K phosphorylation of NR4A1; its wild-type counterpart also records uncertainty about the physiological NR4A1 kinase. The physiological NR4A1-kinase assignment is disputed in the wild-type event; the general AKT kinase and nuclear-… |
| human | AKTIP | file:human/AKTIP/AKTIP-hypotheses/noncatalytic-k63-ubiquitination-and-lesion-bypass/openscientist.md | AKTIP noncatalytic K63 ubiquitination and lesion bypass: focused OpenScientist report | HIGH | Read the full report, citation index, and HTML/PDF report artifacts. Correctly distinguishes missing E2 catalytic cysteine and telomere-replication assays from lesion-bypass assays. Its proposed process refutation is not established: actual PAINT descent includes PTN000630262, experimentally seeded by UBE2N and orthologs, and its negative assertions are on other branches. Pairwise sequence identi… |
| human | ALB | PMID:36979511 | Heme Scavenging and Delivery: The Role of Human Serum Albumin. | MEDIUM | Identifier and full narrative-review text verified. The disputed point is its introductory claim that heme is part of cyanocobalamin; ChEBI identifies cyanocobalamin as a cobalt-centered corrin. This bounded error is not used to reject albumin heme binding, which is independently supported by the original human-albumin transfer experiment. The seeded EXP code remains unchanged, but the review is… |
| human | ALDH7A1 | PMID:40217438 | Precision diagnosis and treatment of vitamin metabolism-related epilepsy. | LOW | Primary PubMed DOI/identifier mapping is verified; full cached clinical review and its ALDH7A1 section were read. The ALDH7A1 subsection switches to ALDH4A1 in its treatment paragraph and uses inconsistent metabolite names, including piperonylic acid in the lysine-pathway caption. These secondary-source wording problems are not adopted as target mechanism. The section is useful clinical context,… |
| human | ALKBH1 | PMID:27027282 | DNA methylation on N(6)-adenine in mammalian embryonic stem cells. | MEDIUM | PubMed-verified (Wu et al., Nature 2016). The mouse ESC precedent for Alkbh1 as a 6mA eraser. Included because it is the origin of the eraser hypothesis; its central claim about mammalian genomic 6mA abundance is directly contested by PMID:32206710, PMID:32203414 and PMID:35113693. |
| human | ALKBH1 | PMID:30017583 | N(6)-Methyladenine DNA Modification in the Human Genome. | HIGH | PubMed-verified (Xiao et al., Mol Cell 2018). Only the abstract is cached. This is one of the two EXP sources for GO:0141131. Its central premise - that 6mA is extensively present in the human genome - is contested by PMID:32206710, PMID:32203414, PMID:35113693 and PMID:41254163. Retained and cited, with the dispute recorded rather than the annotation removed. |
| human | ALKBH1 | PMID:30392959 | N(6)-methyladenine DNA Modification in Glioblastoma. | HIGH | PubMed-verified (Xie et al., Cell 2018). The second EXP source for GO:0141131. Same dispute as PMID:30017583; the glioblastoma 6mA mapping relies on antibody-based detection, which is the specific technique challenged by PMID:32206710. |
| human | ALKBH1 | PMID:40715766 | YTHDF3 recognizes DNA N6-methyladenine and recruits ALKBH1 for 6mA removal from genomic DNA. | HIGH | PubMed-verified (Chen et al., EMBO J 2025). Full text cached. Proposes YTHDF3 as the reader that enables ALKBH1 to act on duplex genomic 6mA. Cited here chiefly because its own abstract concedes that ALKBH1 does not demethylate 6mA in dsDNA and that the genomic activity remains largely debated. Senior authorship overlaps with PMID:30017583, so it is a continuation of the same programme rather tha… |
| human | AP3D1 | PMID:16760431 | BLOC-1 complex deficiency alters the targeting of adaptor protein complex-3 cargoes. | MEDIUM | Sound cell biology - AP-3 and BLOC-1 on shared small vesicles, and mistargeting of LAMP1, PI4KIIalpha and VAMP7-TI. What is disputed is the GO term derived from it: the paper's 'microvesicles' are intracellular transport intermediates, whereas GO:1990742 denotes an extracellular vesicle shed from the plasma membrane. |
| human | AP3D1 | PMID:9545220 | Association of the AP-3 adaptor complex with clathrin. | MEDIUM | Correctly cited for AP-3 binding clathrin, and the binding is real, but its functional inference ('AP-3 function in protein sorting may depend on clathrin') has been superseded: PMID:23761069 shows the association is dispensable and PMID:42139345 shows AP3:ARF1 builds carriers with no clathrin lattice. Also note the interaction it maps is beta3's appendage, not delta's. |
| human | AP3M1 | PMID:9545220 | Association of the AP-3 adaptor complex with clathrin. | MEDIUM | The in vitro clathrin-binding result behind the NAS GO:0035654 row. Correctly cited and a real experiment, but its central claim is contested: the interaction is with beta3 rather than mu3A, the authors' own conclusion is hedged ('may depend on clathrin'), the 2024 review of the field calls the question open, and the 2026 reconstitution shows carrier formation without a clathrin lattice. DISPUTED… |
| human | AP3M2 | file:human/AP3M2/AP3M2-deep-research-affinage.md | Affinage mechanistic annotation for AP3M2 (human) | HIGH | Read all six cited findings. Useful neuronal/pigment-cell context, but calling AP3M2 mu-2 is inaccurate (mu3B); zebrafish and mouse findings are not direct human assays. It does not settle clathrin or TGN questions. |
| human | AP3S2 | PMID:9545220 | Association of the AP-3 adaptor complex with clathrin. | MEDIUM | Correctly cited, and its in vitro finding that the beta-3 appendage binds the clathrin heavy chain is not in question. The conclusion built on it - that AP-3 function involves clathrin coats - has been contested by direct reconstitution and coat tomography [PMID:42139345], which is why the annotation it supports is modified to a coat-agnostic term rather than accepted as written. Flagged DISPUTED… |
| human | AP5M1 | PMID:27136675 | Characterization of MUDENG, a novel anti-apoptotic protein. | MEDIUM | Full text cached. Reports the opposite sign to PMID:18395520 and PMID:31427081, with MuD depletion sensitising astroglioma cells to TRAIL. Its endoplasmic reticulum and mitochondrial localisation rests on a crude microsomal preparation made with a commercial ER isolation kit, a fraction that co-purifies endosomes and lysosomes, so it is not resolution enough to contradict the endolysosomal locali… |
| human | APLP1 | PMID:27708076 | Pathological alpha-synuclein transmission initiated by binding lymphocyte-activation gene 3. | MEDIUM | PubMed-verified (Science 2016). Background for the receptor field rather than direct APLP1 evidence. The central LAG3 claim was subsequently contested by PMID:34309222. |
| human | APOO | GO_REF:0000044 | Gene Ontology annotation based on UniProtKB/Swiss-Prot Subcellular Location vocabulary mapping, accompanied by conserva… | MEDIUM | The UniProt subcellular-location mapping itself operates correctly; what it maps is unreliable for this protein. One of the four rows it generates (mitochondrial inner membrane) is accepted; the other three (Secreted, Golgi apparatus membrane, endoplasmic reticulum membrane) are removed because the UniProt statements behind them rest on a protein species PMID:37279200 showed is not MIC26. Flagged… |
| human | APOO | PMID:16956892 | ApoO, a novel apolipoprotein, is an original glycoprotein up-regulated by diabetes in human heart. | HIGH | The paper that named APOO, and the origin of every secretory annotation on this gene. Correctly cited and historically important, but its central protein assignment has been overturned: the 55 kDa immunoreactive species on which the secretion, glycosylation, lipoprotein-association and cholesterol-efflux observations all rest is unchanged by MIC26 knockdown and knockout and contains no MIC26 pept… |
| human | APOO | PMID:25764979 | The non-glycosylated isoform of MIC26 is a constituent of the mammalian MICOS complex and promotes formation of crista… | HIGH | The single most heavily used reference on this gene, supplying nine GOA rows. Its mitochondrial conclusions - inner-membrane MIC26 as a MICOS subunit interacting with MIC60, MIC27 and MIC10, and required for crista junction number - are accepted in full. Its three-form model is not: the glycosylated/secreted 55 kDa species and its ER/Golgi precursor were withdrawn by the same laboratory in PMID:3… |
| human | APOO | file:human/APOO/APOO-uniprot.txt | UniProtKB entry Q9BUR5 (MIC26_HUMAN) | HIGH | The fetched Q9BUR5 record; accession asserted to be the expected protein (MIC26_HUMAN, 198 aa), not a merged accession resolving elsewhere. Its function, subunit and mitochondrial inner-membrane sections are current and correct. Three things in it are stale: the Secreted / Golgi apparatus membrane / endoplasmic reticulum membrane locations and the chondroitin-sulfate PTM line, all traceable to th… |
| human | ARFGEF3 | GO_REF:0000002 | Gene Ontology annotation through association of InterPro records with GO terms | HIGH | The pipeline itself is sound and correctly applied: ARFGEF3 genuinely carries InterPro IPR000904 (Sec7 domain). What is disputed is the inference the mapping encodes for this entry, namely that presence of the Sec7 domain implies exchange factor activity. ARFGEF3's Sec7 domain has lost the catalytic glutamate, so both GO terms this reference supplies (GO:0005085 and GO:0032012) overstate what the… |
| human | ARFGEF3 | GO_REF:0000044 | Gene Ontology annotation based on UniProtKB/Swiss-Prot Subcellular Location vocabulary mapping, accompanied by conserva… | HIGH | The SL-0086 (Cytoplasm) mapping is exact. The SL-0244 and SL-0245 mappings are not: UniProt's own definition of SL-0244 describes regulated exocytosis of hormones and neurotransmitters, but the mapping sends it to GO:0030133 transport vesicle, which GO defines as the constitutive secretory pathway. GO:0030141 secretory granule and GO:0030133 are disjoint sibling branches under GO:0031410, verifie… |
| human | ARHGAP36 | file:human/ARHGAP36/ARHGAP36-deep-research-affinage.md | Affinage mechanistic annotation for ARHGAP36 (human) | MEDIUM | Full report reviewed. Its PKA, GLI and entosis narrative provides useful primary leads; its broad catalytic/transcription-regulator GO grounding is not direct evidence that ARHGAP36 catalyzes a protein reaction or acts at DNA. Coverage omissions and a provider self-score alone do not establish low scientific quality. Primary structural, catalytic-screen and new accessible entosis data determine t… |
| human | ARHGAP6 | PMID:26628301 | Inhibitory effects of Arhgap6 on cervical carcinoma cells. | MEDIUM | Cited by Reactome R-HSA-9018806 as establishing RAC3-directed GAP activity for ARHGAP6. The paper itself reports a co-immunoprecipitation between Arhgap6 and Rac3 and never measures GTPase activity, so the GAP-activity reading is not supported by it. The paper is correctly cited as a Rac3 interaction; what is disputed is the inference drawn from it. |
| human | ARHGAP6 | Reactome:R-HSA-9018806 | RAC3 GAPs stimulate RAC3 GTPase activity | LOW | Places ARHGAP6 in the tier of GAPs shown to stimulate RAC3 GTPase activity, citing only PMID:26628301, which shows binding and not catalysis. The GO term it produces is still true of ARHGAP6 for a different substrate, so the row is kept; the pathway assertion behind it is what is disputed. |
| human | ARHGEF16 | PMID:21139582 | The HPV16 E6 binding protein Tip-1 interacts with ARHGEF16, which activates Cdc42. | HIGH | Correctly cited and the source of four GO rows, including the TAX1BP3 PDZ interaction which is sound. Marked DISPUTED for its Cdc42 arm alone: PMID:20679435 reports no cellular Cdc42 activation and no binding to nucleotide-free Cdc42 in HEK293T cells, with Zizimin1 as positive control. The two reconcile if Cdc42 activation is conditional on Tip-1 and HPV16 E6, which is how this review treats it.… |
| human | ARHGEF16 | Reactome:R-HSA-205039 | p75NTR indirectly activates RAC and Cdc42 via a guanyl-nucleotide exchange factor | LOW | Source of one cytosol TAS row. DISPUTED rather than MISCITED, because the claim GOA actually draws from this reaction is GO:0005829 cytosol and that claim is correctly supported; what is contested is the reaction's own set-level GTPase output, which never reaches GOA. ARHGEF16 is one of 52 protein participants in a set-level GEF catalyst whose asserted outputs are Rac and Cdc42 - both of which PM… |
| human | ARHGEF16 | Reactome:R-HSA-419166 | GEFs activate RhoA,B,C | LOW | Source of the other cytosol TAS row, and DISPUTED for the same reason as R-HSA-205039: the cytosol claim it supports is fine, the substrate claim it makes is contested and never reaches GOA. The catalyst is a 54-participant DefinedSet of Dbl-family GEFs assembled by domain rather than by measured specificity - it contains TIAM1 and TIAM2 (Rac-only), FGD1-FGD4 and ITSN1 (Cdc42-only), SOS1, SOS2 an… |
| human | ARHGEF18 | PMID:14512443 | G Protein betagamma subunits stimulate p114RhoGEF, a guanine nucleotide exchange factor for RhoA and Rac1: regulation o… | HIGH | Correctly cited and it is the source of the gene's only curated molecular-function annotation, but its Rac1 conclusion is contradicted by the purified-protein experiment in PMID:11085924 and by the depletion experiment in PMID:21258369. Marked DISPUTED for that reason, not for any citation defect. No free full text was found, so the assessment rests on the abstract plus the two contradicting pape… |
| human | ARHGEF19 | PMID:15485661 | WGEF is a novel RhoGEF expressed in intestine, liver, heart, and kidney. | HIGH | PubMed-verified as Wang et al., Biochem Biophys Res Commun 324:1053-8 (2004), and independently corroborated as the ARHGEF19 evidence cited by Reactome R-HSA-8980691. Marked DISPUTED rather than VERIFIED because its substrate claim is contradicted by the later side-by-side retest in PMID:18256687, not because the citation is wrong. Cached as abstract only; the full text was not consulted. |
| human | ARHGEF19 | PMID:38714795 | Dishevelled2 activates WGEF via its interaction with a unique internal peptide motif of the GEF. | HIGH | Important and carefully done, but it contradicts PMID:18256687 on where Dishevelled binds and neither mapping has been reproduced independently, so DISPUTED records a live disagreement rather than a fault. Two things a reader of the database rather than the paper would get wrong: the activity measurements are on Xenopus protein, and PDB 8YR7 - which UniProt lists under Q8IW93, implying a human AR… |
| human | ATP13A1 | file:human/ATP13A1/ATP13A1-hypotheses/kgap-atp13a1-dislocase-vs-topogenesis/openscientist.md | ATP13A1 Dislocase Hypothesis: Final Report | HIGH | Substantive productive-topogenesis finding is supported by primary studies. The 2026 Spf1 experiment measures substrate-stimulated ATP hydrolysis, not a complete purified dislocation cycle; calcium-homeostasis review flag does not establish loss of a process role. |
| human | BAIAP2L2 | file:human/BAIAP2L2/BAIAP2L2-hypotheses/function-hypothesis-go-0005654/openscientist.md | OpenScientist focused report on BAIAP2L2/Pinkbar nucleoplasm localization | HIGH | The report adds a current-database absence check and summarizes direct membrane/vesicle/stereocilia localization. Its own limitations admit that missing nuclear annotations, null HPA fields and a heuristic NLS scan do not exclude a nuclear pool; these data do not justify the stronger refuted verdict. The historical GOA source is known (PTN001022094), although absent from the current cache. Its su… |
| human | BBS2 | PMID:22302990 | Direct role of Bardet-Biedl syndrome proteins in transcriptional regulation. | LOW | Primarily BBS7-centric; shows BBS7 interaction with PcG member RNF2 (Q99496) and proposes a nuclear/transcriptional role for BBS proteins generally. The transcriptional role for BBS2 specifically is indirect and contested versus the well-established cytoplasmic/ciliary trafficking function. |
| human | BBS4 | PMID:22302990 | Direct role of Bardet-Biedl syndrome proteins in transcriptional regulation. | LOW | Primarily a BBS7 study (nuclear role, RNF2/PcG interaction); BBS4-specific transcription-factor-binding is a weak generalization and is treated as an over-annotation here. |
| human | BCL2 | file:human/BCL2/BCL2-hypotheses/conditional-proapoptotic-and-channel-capacities/openscientist.md | BCL2 conditional proapoptotic and channel capacities: focused OpenScientist report | HIGH | The complete report supports retaining conditional proapoptotic and intrinsic channel capacities, consistent with the independently checked primary texts. Its acidic-pH-only permeabilization summary must be restricted to the carboxyfluorescein assay in PMID:9219694: PMID:9144199 directly records neutral-pH ion channels. The report admits that it did not inspect actual PAINT nodes and reviewed no… |
| human | CACNA2D1 | PMID:19818485 | Gabapentin receptor alpha2delta-1 is a neuronal thrombospondin receptor responsible for excitatory CNS synaptogenesis. | HIGH | PubMed-verified (Cell 2009). The origin of the thrombospondin-receptor/synaptogenesis model. Correctly cited and highly influential, but the claim is contested - a 2025 presynaptic triple conditional knockout found synapse development intact. Note that the original claim was specifically postsynaptic, so the 2025 presynaptic result narrows it rather than overturning it outright. Not used to suppo… |
| human | CASP4 | PMID:41702406 | Caspase-4 binds to LPS membranes with positive curvature for non-canonical inflammasome activation. | HIGH | PubMed-verified (Immunity 2026, Broz lab); abstract-only in cache, full text not retrieved. Correctly cited, but its central claim - that caspase-4 engages LPS membranes of defined geometry rather than individual LPS molecules - is in tension with the direct LPS/lipid A binding measurements underlying the GO:0001530 IDAs and with the 2026 HDX-MS pocket mapping. Flagged as DISPUTED for that reason… |
| human | CASP4 | PMID:42044191 | Human non-canonical inflammasomes activate CASP3 to limit intracellular Salmonella replication in macrophages. | HIGH | PubMed-verified (PLoS Pathog 2026); full text cached from PMC and quotes checked verbatim. The CASP3/CASP7 substrate claim is direct biochemistry but single-source, and the accompanying claim that most GSDMD cleavage in non-canonical signalling is CASP1-mediated is a cell-context flux argument that cuts against the widely reproduced CASP4-GSDMD axis. Contested rather than miscited. |
| human | CDH23 | file:human/CDH23/CDH23-hypotheses/catenin-neuronal-and-calcium-process-scope/openscientist.md | CDH23: catenin, neuronal and calcium process scope | HIGH | Full report read. Correctly distinguishes colocalization from binding and identifies the PMCA2 assay context. However, its repeated claim that PMID:30747484 contains no catenin mentions is contradicted by the full-text Results and Discussion. Absence of human neuronal experiments does not refute the inherited process, and a structural contributor need not itself pump calcium. |
| human | CFAP410 | file:human/CFAP410/CFAP410-deep-research-openscientist.md | Focused Primary-Evidence Audit: Human CFAP410 (O43822) Subcellular Localization for GO Curation | HIGH | Useful secondary audit corroborating the independently checked Bai2011 source mismatch. Several recommendations exceed the evidence: it equates unsupported PM source attribution with a biologically contradicted location; proposes outer-segment remapping from abstract wording; treats locally available Fang full text as inaccessible; summarizes HPA without its per-location/antibody limitations; and… |
| human | CRMP1 | file:human/CRMP1/CRMP1-hypotheses/function-hypothesis-go-0016812/openscientist.md | OpenScientist hypothesis investigation - CRMP1 cyclic-amide hydrolase activity | HIGH | The residue-loss analysis and primary CRMP2 structural evidence substantiate loss of the ancestral cyclic-amide reaction. The report overstates that no CRMP has any catalytic activity (the CRMP3/DPYSL4 HDAC report PMID:23443259 is a counterexample needing assessment), and cannot establish that abstract-only PMID:8973361 contains no assay anywhere in its full text. Its author attribution of PMID:2… |
| human | DNER | PMID:15965470 | DNER acts as a neuron-specific Notch ligand during Bergmann glial development. | HIGH | The cached abstract directly supports the Bergmann-glial phenotype and proposed Deltex-dependent/RBP-J-independent signaling, but its claim that DNER binds and activates NOTCH1 was not reproduced in PMID:27622512. |
| human | DPYSL4 | file:human/DPYSL4/DPYSL4-hypotheses/crmp3-deacetylase-and-broad-hydrolase-activity/openscientist.md | OpenScientist assessment of CRMP3 deacetylase and broad hydrolase activity | HIGH | Read entire report and compared Methods/Results to the full primary article. Enzyme-attribution and replication uncertainties are useful, but full-length constructs were tested; HEK293 and Sf9 tags/purification differ; TSA sensitivity does not demonstrate contamination; NOT dihydropyrimidinase does not negate parent hydrolase terms. Exact reported sequence identities/residue maps lack delivered c… |
| human | EIF2AK3 | file:human/EIF2AK3/EIF2AK3-hypotheses/function-hypothesis-go-0005634/openscientist.md | OpenScientist focused report on EIF2AK3/PERK nuclear localization | HIGH | The report usefully identifies HPA nucleoplasm immunofluorescence and the full-length ER-membrane topology. Its assertion that GO:0005634 excludes the nuclear envelope is false: GO:0005635 is part_of GO:0005634 (AmiGO verified 2026-09-20). Absence of direct annotations and primary ER residence do not refute an ancestral IBA. HPA CAB009204 has standard Approved ICC, not enhanced genetic or orthogo… |
| human | EIF4E2 | file:human/EIF4E2/EIF4E2-hypotheses/eif4f-complex-under-hypoxia/openscientist.md | OpenScientist assessment of EIF4E2 and the eIF4F complex under hypoxia | HIGH | Full narrative and delivered plot provenance code read. Supported initiation/repression findings retained. The complex-composition rejection overlooks primary endogenous eIF4E2/eIF4A/eIF4G3 pull-downs and depletion evidence in PMID:26854219 and eIF4A recruitment in PMID:22678294. The report also conflates cap-pocket stacking residues with the opposite dorsal partner-binding surface and labels IBA… |
| human | FTO | PMID:22002720 | N6-methyladenosine in nuclear RNA is a major substrate of the obesity-associated FTO. | HIGH | The in vitro m6A demethylation is not in question and the paper is correctly cited. Its central inference - that nuclear mRNA m6A is the major physiological substrate - is what the direct-sequencing reassessment (PMID:41279954) and the m6Am literature contest. |
| human | FTO | PMID:26458103 | Dynamic m(6)A mRNA methylation directs translational control of heat shock response. | MEDIUM | Correctly cited and a genuine mutant-phenotype experiment, but the m6A readout is antibody-based, which is the methodology under challenge. It is the IMP leg of the contested GO:1990931 annotation. |
| human | FTO | PMID:41190354 | Structure-Based Design of a Highly Potent Dual-Competitive FTO Inhibitor for Targeted m(6)A Demethylase Inhibition in A… | LOW | Representative of the active drug-discovery literature that assumes the m6A-eraser model. Cited here to show that side of the dispute is current, not superseded. Its cellular m6A readouts are antibody-based, and the claim that FTO inhibitors act through FTO is exactly what PMID:41279954 challenges. Cached record is abstract-only. |
| human | GAL3ST1 | PMID:36805701 | Histone tyrosine sulfation by SULT1B1 regulates H4R3me2a and gene transcription. | MEDIUM | PubMed-verified (Nat Chem Biol 2023;19:855-864). Defines the H3Y99sulf mark that PMID:41686426 later attributes to GAL3ST1, and attributes it to a different enzyme (SULT1B1). Formally contested by PMID:40890505; the authors replied (PMID:40890506) and an Author Correction was issued (PMID:40890508). Not retracted. Cited only to document the dispute; no GAL3ST1 annotation derives from it. |
| human | GAL3ST1 | PMID:40890505 | Mass spectrometry and enzyme assays refute histone tyrosine sulfation. | MEDIUM | PubMed-verified (Nat Chem Biol 2025;21:1667-1670), full text available via PMC12978989. A formal Matters Arising arguing that the H3Y99sulf spectra fit a phosphotyrosine peptide, that sulfotyrosine cannot give the reported 100% sulfate-retaining HCD fragments, and that recombinant enzyme gave no histone H3.2 sulfation. Bears on GAL3ST1 because it questions the existence of the mark and the specif… |
| human | GAL3ST1 | PMID:40890506 | Reply to: Mass spectrometry and enzyme assays refute histone tyrosine sulfation. | LOW | PubMed-verified (Nat Chem Biol 2025;21:1671-1674). The original authors' reply. No abstract in PubMed and the cached record is abstract-only, so its arguments could not be assessed. Its existence is why the dispute is treated as unresolved rather than settled. |
| human | GAL3ST1 | PMID:41686426 | GAL3ST1-Mediated Histone Tyrosine Sulfation Induced by Cancer-Associated Fibroblasts Promotes Gastric Cancer Metastasis. | MEDIUM | PubMed-verified (Cancer Res 2026;86:2429-2446), full text available via PMC13176828. Claims GAL3ST1 is a histone sulfotransferase writing H3Y99sulf in the cytosol. Three concerns: the mark itself is contested by PMID:40890505, which this paper does not cite; the readouts rest on the anti-H3Y99sulf antibody class the refutation says needs further validation; and the proposed cytosolic reaction is… |
| human | GET1 | PMID:9544840 | Identification and characterization of a new human cDNA from chromosome 21q22.3 encoding a basic nuclear protein. | LOW | Source of the nucleus annotation; the nuclear localization was not reproduced and is contradicted by later work establishing ER membrane residence. |
| human | GPATCH11 | PMID:20813266 | The protein composition of mitotic chromosomes determined using multiclassifier combinatorial proteomics. | LOW | Source of the disputed kinetochore IDA. Genuine proteomics paper but the per-protein localization assignment for GPATCH11 is based on a low-resolution image, and a dedicated 2024 study (PMID:39572588) contradicts the kinetochore call. Upstream GO curators have flagged the derived annotations in geneontology/go-annotation#6450. |
| human | GPR158 | PMID:28851741 | Gpr158 mediates osteocalcin's regulation of cognition. | MEDIUM | PubMed-verified (J Exp Med 2017, Khrimian et al.). The osteocalcin-receptor assignment rests on mouse genetics, electrophysiology and behaviour; no direct high-affinity binding site has been defined, and UniProt annotates the osteocalcin role only by similarity. Included to document the competing ligand model, not as established fact. |
| human | GPR158 | PMID:40337551 | Osteocalcin and GPR158: linking bone and brain function. | LOW | PubMed-verified (Front Cell Dev Biol 2025). A review that states the osteocalcin-receptor role as settled; cited here to document that framing, not to support it. The primary evidence it rests on is mouse genetic rather than direct binding. |
| human | GPR37 | PMID:16443751 | The neuropeptide head activator is a high-affinity ligand for the orphan G-protein-coupled receptor GPR37. | HIGH | PubMed-verified (J Cell Sci 2006, Rezgaoui et al.). The head-activator deorphanization. Marked DISPUTED because the pairing was not reproduced by Dunham et al. 2009, was not a hit in a genome-wide orphan screen, and the ligand has no human gene; the WITH/FROM entity of the derived GOA row is the Hydra vulgaris peptide UniProtKB:P69251. Its localisation observations remain usable. |
| human | GPR37 | PMID:34678058 | Osteocalcin attenuates oligodendrocyte differentiation and myelination via GPR37 signaling in the mouse brain. | HIGH | PubMed-verified (Sci Adv 2021;7:eabi5811, Qian et al.). Original osteocalcin-GPR37 pairing. Marked DISPUTED because osteocalcin already has two other claimed receptors and because the only follow-up study shares its senior authors. |
| human | GPR37 | PMID:41679312 | A bone-derived hormone permits rapid visual escape via GPR37 receptor in a subpopulation of VTA GABAergic neurons. | HIGH | PubMed-verified (Neuron 2026;114:1594-1608, Liu et al.). Osteocalcin-GPR37 in VTA GABAergic neurons. Marked DISPUTED on independence grounds (shared senior authors with PMID:34678058) and because it carries a four-page published erratum, PMID:42001851. |
| human | GPR50 | PMID:41495223 | Photo-cross-linking-assisted deorphanization deciphers GPR50-L-LEN pairing in metabolism. | HIGH | PubMed-verified (Nat Chem Biol 2026;22:1132-43, Wu et al.). Technically strong single-laboratory deorphanization proposing L-LEN as the endogenous agonist, with Galphai coupling. Marked DISPUTED rather than VERIFIED because a cryo-EM study published five weeks later states that no endogenous agonist has been characterized without engaging this result, and because the reported transducer (Galphai)… |
| human | GPR75 | PMID:17001303 | RANTES stimulates Ca2+ mobilization and inositol trisphosphate (IP3) formation in cells transfected with G protein-coup… | HIGH | PubMed-verified (Br J Pharmacol 2006, Ignatov et al.). This is the single experimental study behind every C-C chemokine receptor annotation on human GPR75, reaching it via the mouse Gpr75 IDA. Correctly cited, but the deorphanization it claims has not been adopted by IUPHAR and is described as controversial in 2025-2026 reviews. |
| human | GPR75 | PMID:29772059 | The orphan G-protein-coupled receptor 75 signaling is activated by the chemokine CCL5. | HIGH | PubMed-verified (J Neurochem 2018, Dedoni et al.). Independent support for CCL5-evoked, GPR75-dependent signalling in cells lacking CCR1/CCR3/CCR5, and the source of the IC plasma membrane annotation. Cited as the strongest evidence for the chemokine model; still short of demonstrating direct chemokine binding, which the 2026 structures make harder to envisage. |
| human | GSDMB | PMID:32299851 | Granzyme A from cytotoxic lymphocytes cleaves GSDMB to trigger pyroptosis in target cells. | HIGH | Identifier verified against PubMed: Science 2020;368(6494):eaaz7548, Zhou et al. Flagged DISPUTED because Science published a Published Erratum for it in 2026 (PMID:42424476, doi 10.1126/science.aej7549), and the content of that erratum could not be retrieved - PubMed carries no abstract, Europe PMC records it as subscription-only and not in PMC, and science.org returns HTTP 403. I do not know wh… |
| human | HSPA13 | file:human/HSPA13/HSPA13-hypotheses/function-hypothesis-go-0044183/openscientist.md | OpenScientist hypothesis investigation - HSPA13 protein folding chaperone | HIGH | Structural divergence is supported. Its universal impossibility/refutation language goes beyond the GO definition and assays; promoting degradation does not generally exclude a chaperone role, and positive ER localization is nonexclusive. The recent papers were read independently. |
| human | HSPA14 | file:human/HSPA14/HSPA14-uniprot.txt | UniProt entry Q0VDF9 (HSP7E_HUMAN), Heat shock 70 kDa protein 14 / HSP70L1 | HIGH | The record correctly identifies mRAC and cytosolic location, but its statement that DNAJC2 stimulates HSPA14 ATPase is not established by full PMID:21245388. Direct nascent-chain contact by HSPA14 is also proposed rather than resolved. |
| human | HSPA14 | file:yeast/SSZ1/SSZ1-hypotheses/function-hypothesis-go-0016887/openscientist.md | OpenScientist SSZ1 ATPase adjudication (comparative scope) | MEDIUM | Actual report read for homolog coverage. It explicitly calls HSPA14 activity untested and omits PMID:21245388, which performed the inconclusive near-background assay; it cannot adjudicate the human target. No direct transfer of fungal ATPase loss is made. |
| human | IL23R | file:human/IL23R/IL23R-hypotheses/function-hypothesis-go-0004925/openscientist.md | OpenScientist hypothesis report for IL23R prolactin receptor activity | HIGH | The ligand-specific prolactin-receptor conclusion is supported by primary IL-23 receptor structural biology and the PRLR source trace. However, PMID:12023369 tests IL-23 versus IL-12, not prolactin; k-mer dissimilarity cannot prove ligand incompatibility. The report separately accepts peptide hormone binding, which this review leaves UNDECIDED pending a precise hormonal-scope judgment. |
| human | KIAA1614 | file:human/KIAA1614/KIAA1614-deep-research-falcon.md | Deep research report on KIAA1614 | MEDIUM | Useful unknown-function summary, but treating IPR051741 as a physical PAR6 domain and asserting no detected partners/localization omits independently curated high-throughput evidence. Lack of target studies does not refute IBA. |
| human | KIAA1614 | file:human/KIAA1614/KIAA1614-hypotheses/kgap-kiaa1614-par6-polarity-scaffold/openscientist.md | OpenScientist KIAA1614 Par6-like polarity-scaffold hypothesis report | HIGH | Substantive report read and reused. InterPro confirms divergent annotated architecture, but also a PTHR14102:SF12 match (reported score 4.6e-32); a four-mer Jaccard comparison does not refute homology. IBA is not circular just because the target lacks experimental GO rows. Missing domains or interaction-database edges do not establish inability to bind. AlphaFold numeric claims lack delivered com… |
| human | LOXL2 | PMID:22483618 | Lysyl oxidase-like 2 deaminates lysine 4 in histone H3. | HIGH | PubMed classifies this record as a Retracted Publication and explicitly links retraction notice PMID:27392148. Exclude its H3K4me3 oxidation and chromosome-catalysis claims from accepted evidence. |
| human | LOXL2 | PMID:24239292 | Regulation of heterochromatin transcription by Snail1/LOXL2 during epithelial-to-mesenchymal transition. | MEDIUM | Abstract-only mouse mechanistic study whose H3K4-deaminase premise overlaps the subsequently retracted PMID:22483618 claim; do not use as secure evidence of chromosome-associated LOXL2 catalysis. |
| human | LOXL2 | PMID:27735137 | Lysyl oxidase-like 2 (LOXL2) oxidizes trimethylated lysine 4 in histone H3. | HIGH | Later abstract-only paper from substantially the same author group reasserting the H3K4me3 chemistry after PMID:22483618 was retracted. It does not erase the retraction or provide independently verified support suitable for confident GO curation. |
| human | LOXL2 | file:human/LOXL2/LOXL2-uniprot.txt | UniProtKB record for human LOXL2 (Q9Y4K0) | HIGH | The sequence, precursor status, extracellular localization, catalytic reaction, and cofactors are authoritative. However, the chromosome/chromatin statements depend on disputed PMID:27735137, and the record still mentions retracted PMID:22483618 in its intracellular-location note; those nuclear catalytic claims must be separated from the reliable extracellular enzyme record. |
| human | LPA | PMID:2531657 | Lipoprotein(a) binds to fibronectin and has serine proteinase activity capable of cleaving it. | HIGH | Correct source for the curator-read fibronectin-binding and experimental protease annotations, but only the abstract is cached. Its catalytic interpretation conflicts with PMID:3472206 and the recombinant apo(a) tests in PMID:7495809. Binding to a heparin-binding region of fibronectin is not direct evidence that apo(a) binds heparin. |
| human | LPA | file:human/LPA/LPA-uniprot.txt | UniProtKB entry P08519 (APOA_HUMAN), cached in the gene folder | HIGH | The sequence, domain, polymorphism, and particle-subunit statements are authoritative for the cached reference allele. Its catalytic FUNCTION wording cites only PMID:2531657 and conflicts with PMID:3472206 and PMID:7495809, so that specific claim is disputed rather than treated as established core activity. |
| human | LPCAT4 | Reactome:R-HSA-1482635 | 2-acyl LPG is acylated to PG by LPGAT | MEDIUM | This 2-acyl LPG reaction requires acylation at free sn-1, but PMID:41740885 directly found no LPCAT4 activity toward sn-2-dominant LPG. The older non-position-resolved LPG result does not establish this positional reaction. |
| human | LPCAT4 | Reactome:R-HSA-1482925 | Acyl chain remodelling of PG | MEDIUM | The pathway combines a plausible 1-acyl LPG-to-PG reaction with a contradicted 2-acyl LPG reaction. LPCAT4's PG remodeling is therefore position dependent rather than uniformly supported. |
| human | LPGAT1 | PMID:15485873 | Identification and characterization of a gene encoding human LPGAT1, an endoplasmic reticulum-associated lysophosphatid… | HIGH | Direct human enzyme and ER-localization study, verified from the PubMed abstract. Its LPG substrate interpretation is now disputed by position-resolved studies, especially PMID:42173283, which proposes LPG acyl migration as the source of the historical assignment and finds no significant PG remodeling in vivo. The ER localization remains strong. |
| human | LPGAT1 | PMID:30831319 | Defective Phosphatidylglycerol Remodeling Causes Hepatopathy, Linking Mitochondrial Dysfunction to Hepatosteatosis. | MEDIUM | The knockout phenotypes are direct mouse evidence, but the attribution to defective PG remodeling is challenged by PMID:42173283, which reports unchanged PG composition and proposes a broader phospholipid-regeneration defect instead. |
| human | LPGAT1 | PMID:37917582 | LPGAT1 controls MEGDEL syndrome by coupling phosphatidylglycerol remodeling with mitochondrial transport. | HIGH | Full-text mouse and cell-based mechanistic study. Its PG-specific transport model is directly challenged by PMID:42173283, which found no significant PG remodeling role and instead linked mitochondrial effects to general PC/PE/PS regeneration; the interaction and phenotype observations remain relevant but context-bounded. |
| human | LPGAT1 | Reactome:R-HSA-1482539 | 1-acyl LPG is acylated to PG by LPGAT | MEDIUM | The human ER reaction is traceable to the original LPGAT1 report, but the PG-remodeling assignment is challenged by position-resolved and in-vivo lipidomic evidence in PMID:35131264 and PMID:42173283. |
| human | LPGAT1 | Reactome:R-HSA-1482635 | 2-acyl LPG is acylated to PG by LPGAT | MEDIUM | The human ER reaction is traceable to the original LPGAT1 report, but the acceptor-position interpretation is especially vulnerable to LPG acyl migration and is not supported by the newer in-vivo lipidomics in PMID:42173283. |
| human | LPGAT1 | Reactome:R-HSA-1482925 | Acyl chain remodelling of PG | MEDIUM | Reactome provides pathway provenance for PG Lands-cycle remodeling. LPGAT1 participation is contested by newer position-resolved biochemistry and Huh-7/mouse lipidomics, so the pathway record should not outweigh the direct evidence. |
| human | LPGAT1 | file:human/LPGAT1/LPGAT1-hypotheses/secondary-acylation-capacities-and-lipid-pathway-scope/openscientist.md | LPGAT1 secondary acylation capacities and lipid pathway scope: OpenScientist focused report | HIGH | Read full report. Correctly retains secondary capacity and distinguishes mouse MGAT provenance, but PI-paralog attribution is speculative and omits the actual ancestral node; absence of a UniProt LPI reaction is not loss evidence. It conflates lower serum TAG in 20018982 with liver TAG, which slightly increased, and overreads 2026 accumulation as excluding a biosynthetic role. RHEA labels alone d… |
| human | LRP5L | PMID:32789677 | The novel mutation P36R in LRP5L contributes to congenital membranous cataract via inhibition of laminin γ1 and c-MAF. | HIGH | PubMed identifier/title and the complete cached abstract were manually checked. This is the only located primary experimental study directly about human LRP5L, but its causal protein interpretation is now materially contested by current NCBI Gene/RefSeq annotation, and the unavailable full text prevents independent assessment of construct identity, segregation, and assay controls. The reported ex… |
| human | LRP5L | file:human/LRP5L/LRP5L-uniprot.txt | UniProtKB/Swiss-Prot A4QPB2 (LRP5L_HUMAN) record | HIGH | The local flat file and accession were manually checked. It accurately records UniProtKB's current reviewed PE2 assertion, sequence, repeats, splice variants, and external proteomics database links, but the protein-coding interpretation is disputed by current NCBI and HGNC pseudogene classifications. The PE2 assignment shows that UniProtKB has not accepted the linked resources as protein-level ex… |
| human | MAGT1 | PMID:18455129 | Oligosaccharyltransferase-subunit mutations in nonsyndromic mental retardation. | MEDIUM | PubMed-verified. Reports the MAGT1/IAP V311G variant in X-linked NSMR and links OTase activity to cognition. UniProt CAUTION notes the pathological role is questionable (PubMed:23871722); underpins the contested cognition (GO:0050890) annotation. |
| human | MAGT1 | PMID:19717468 | Mammalian MagT1 and TUSC3 are required for cellular magnesium uptake and vertebrate embryonic development. | MEDIUM | PubMed-verified. Key study for the Mg2+-uptake phenotype and the plasma-membrane transporter model. The direct-transporter interpretation is now contested; the Mg2+ effect is widely regarded as indirect/downstream of the OST-B glycosylation role. |
| human | MAGT1 | Reactome:R-HSA-5339538 | MAGT1 transports Mg2+ from extracellular region to cytosol | LOW | Reactome reaction encoding the historical direct Mg2+-transporter model; the direct transporter activity is contested. Title left exactly as fetched. |
| human | NAT10 | PMID:30449621 | Acetylation of Cytidine in mRNA Promotes Translation Efficiency. | HIGH | PubMed-verified (Arango et al., Cell 2018). Correctly cited for the annotations that rest on it, but its central mRNA-ac4C claim is contested by PMID:32555463 and PMID:38640895 and was defended in PMID:38640896. The acRIP-seq method is antibody-based and not base-resolution. |
| human | NAT10 | PMID:35679869 | Direct epitranscriptomic regulation of mammalian translation initiation through N4-acetylcytidine. | HIGH | PubMed-verified (Arango et al., Mol Cell 2022; note a published erratum, Mol Cell 82:2912). The base-resolution positive result for human mRNA ac4C. Directly challenged by PMID:38640895 on reproducibility and duplicate-read grounds and defended in PMID:38640896. The cryo-EM component, showing how ac4C at position -1 perturbs the initiator tRNA t6A37 interaction, is independent of the mapping disp… |
| human | NAT10 | PMID:38640896 | Detection of ac4C in human mRNA is preserved upon data reassessment. | HIGH | PubMed-verified (Beiki et al., Mol Cell 2024). The back-to-back rebuttal defending RedaC:T-seq and raising depth and reduction-efficiency problems in the ac4C-seq data. Marked DISPUTED because it is one half of an unresolved exchange, not because the citation is faulty. |
| human | NAT10 | PMID:41956987 | NAT10 promotes cisplatin resistance and immune escape by increasing the expression of DUSP1 and PD-L1 in gastric cancer. | LOW | PubMed-verified (Cell Death Discov 2026). Representative of the 2025-2026 disease literature that asserts NAT10-dependent mRNA ac4C on named transcripts using antibody-based ac4C-RIP plus knockdown, without citing or addressing PMID:38640895. Does not provide independent methodological support for the contested substrate claim. |
| human | NAT10 | PMID:42315153 | NAT10 deficiency in aging drives impaired liver regeneration by disrupting PARP10 in an ac4C-dependent manner. | LOW | PubMed-verified (Clin Mol Hepatol 2026); abstract-only in the local cache. Another ac4C-RIP-based single-transcript mRNA acetylation claim that does not engage the mapping dispute. |
| human | NAT10 | PMID:42592486 | Targeting NAT10 with Remodelin reshapes the pro-fibrotic microenvironment in renal fibrogenesis via inhibiting exosome… | LOW | PubMed-verified (Acta Pharm Sin B 2026). Invokes both contested activities at once - nuclear mRNA ac4C and cytoplasmic lysine acetylation - and relies on remodelin, which impairs ribosome biogenesis, so the phenotype does not discriminate between the proposed mechanisms. |
| human | NLRP3 | PMID:26098997 | The receptor NLRP3 is a transcriptional regulator of TH2 differentiation. | MEDIUM | Bruchard et al., Nat Immunol 2015, doi 10.1038/ni.3202 - PubMed-verified. The sole primary source behind every DNA-binding, IRF4-binding, nuclear and T helper 2 annotation on human NLRP3, all of which reach the human gene by IBA or ISS from mouse Nlrp3. Single laboratory, mouse only, carries a published corrigendum (PMID:26580508), and no NLRP3 DNA-binding domain or NLRP3-DNA structure has been r… |
| human | NOTCH1 | file:human/NOTCH1/NOTCH1-hypotheses/function-hypothesis-go-0007411/openscientist.md | OpenScientist focused assessment of human NOTCH1 axon guidance | HIGH | The fly Notch/Disabled/Trio/Rac mechanism and mammalian neurite-versus-pathfinding distinction are useful, supported by independently fetched primary papers. However, the report misidentifies FBgn0264089 and MGI:1315202/1315203/1315205 as Notch orthologs: FlyBase/MGI identify them as sli, Slit3, Slit1 and Slit2. Fresh QuickGO confirms the identifiers, not the report’s names. STRING edge absence a… |
| human | NQO2 | file:human/NQO2/NQO2-hypotheses/function-hypothesis-go-0003955/openscientist.md | OpenScientist blinded function-assignment report for NQO2 (NAD(P)H dehydrogenase (quinone) activity, GO:0003955) | HIGH | Read and reused the existing focused report without a duplicate run. Its NRH-preference finding is sound, but the categorical absence/removal inference omits the measurable human NADH turnover summarized in PMID:35517822. The C-terminal chimera in PMID:9367528 also shows that a missing C-terminal segment alone does not explain donor preference. Treat the report as a useful, partially disputed sec… |
| human | NSUN2 | PMID:22395603 | The tRNA methyltransferase NSun2 stabilizes p16INK⁴ mRNA by methylating the 3'-untranslated region of p16. | MEDIUM | Full text available and read. Reports in vitro methylation of the p16 3'UTR and an mRNA half-life effect, but the site is given as "A988", and the paper carries an erratum (Nat Commun 2013;4:2703). Supports the mRNA activity in principle while illustrating why site-level claims in this literature need orthogonal confirmation. |
| human | NSUN2 | PMID:31358969 | 5-methylcytosine promotes pathogenesis of bladder cancer through stabilizing mRNAs. | MEDIUM | Abstract only in our cache. Influential single-nucleotide mRNA m5C map in bladder cancer and the source of the YBX1 reader model. The site-level conclusions depend on bisulfite mapping whose specificity is questioned in PMID:42587746. |
| human | NSUN2 | PMID:41707999 | The RNA methyltransferase NSUN2 catalyzes 5-methylcytosine (m(5)C) on IL1B mRNA to promote transcript stability. | LOW | Full text available. Representative of the recent single-transcript m5C literature (IL1B, read by YBX1). Cited here as an example of the pattern, not as support for any annotation; no orthogonal, site-resolved validation is presented. |
| human | NSUN7 | PMID:41062800 | NSUN7-mediated m5C methylation of NLRP3 promotes pyroptosis in ovarian granulosa cells in polycystic ovary syndrome. | LOW | Knockdown plus MeRIP and actinomycin D chase in a granulosa cell line. Catalysis is assumed rather than tested. |
| human | NSUN7 | PMID:41275291 | NSUN7-mediated m(5)C modification of circNTRK2 regulates stemness properties of glioblastoma cells by activating STK31. | LOW | Uses the word "catalyzes" of a knockdown-plus-m5C-mapping result in glioblastoma lines. No direct test of catalysis, and NSUN7 is not appreciably expressed outside testis. |
| human | NSUN7 | PMID:41871257 | The m5C orchestrator NSUN7 drives SPARC/HMGB1 axis-mediated inflammation to exacerbate kidney injury. | MEDIUM | Abstract only in our cache (PMC full text is publisher-restricted). Asserts that NSUN7 is the main driver of kidney m5C, which conflicts with the direct biochemistry and with the testis-restricted expression of NSUN7. A fall in bulk m5C after deleting a protein does not establish that the protein is the methyltransferase; the abstract reports no SAM binding, no in vitro methylation and no catalyt… |
| human | NT5C2 | PMID:36159777 | TRIM22 orchestrates the proliferation of GBMs and the benefits of TMZ by coordinating the modification and degradation… | MEDIUM | Full text available. Sole source of the ubiquitin-ligase (GO:0061630), K48-linked ubiquitination (GO:0070936) and antiviral (GO:0050689) annotations. This is a moonlighting claim well outside the enzyme's HAD-fold metabolic role (no recognizable E3 domain; interaction with RIG-I is indirect via TRIM22) and is not independently corroborated; treated as a likely over-annotation. |
| human | P2RX7 | PMID:17785580 | Evidence for functional P2X4/P2X7 heteromeric receptors. | MEDIUM | PubMed-verified (Mol Pharmacol 2007). The P2X4/P2X7 heteromer claim. Correctly cited but the claim is contested - the paper was published with an accompanying commentary and heteromeric assembly has not become consensus. |
| human | PHGDH | PMID:40273909 | Transcriptional regulation by PHGDH drives amyloid pathology in Alzheimer's disease. | MEDIUM | PubMed-verified (Cell 2025, PMID:40273909, PMC12204802; full text cached). Not disputed by a published rebuttal; marked DISPUTED because the claim is a major, enzyme-independent reassignment of function that rests on a single laboratory, on an AlphaFold-predicted DNA-binding subdomain embedded in the NAD(H)-binding Rossmann fold, and on overexpression phenotypes. No GOA annotation currently depen… |
| human | PHGDH | PMID:41701839 | RNA-binding activity of PHGDH drives amyloid-beta production in a human brain organoid model of sporadic Alzheimer's di… | MEDIUM | PubMed-verified (PNAS 2026, PMID:41701839, PMC12933074; full text cached). Carries a correction (PMID:41838922). Marked DISPUTED on the same grounds as PMID:40273909, plus two specific to it: it is the second mutually-independent moonlighting mechanism proposed for PHGDH by the same senior author within a year, and it is a PNAS Contributed-track submission (contributed by co-author Shu Chien), so… |
| human | PIK3C3 | file:human/PIK3C3/PIK3C3-hypotheses/pexophagy-and-peroxisome-localization/openscientist.md | OpenScientist focused adjudication of PIK3C3 pexophagy and peroxisome localization (recovered cancelled-job artifact) | HIGH | Exact provider final_report.md recovered through supported artifact download from cancelled job dc13b854-c862-4926-82a1-8ae985f1ebe0; report endpoint refused cancelled status. Useful PAINT/yeast-complex evidence, but its cargo-confusion explanation conflicts with the actual biogenesis-localization experiments and its absence-of-human-assay argument does not refute conservation. PI3P synthesis is… |
| human | PIK3R4 | file:human/PIK3C3/PIK3C3-hypotheses/pexophagy-and-peroxisome-localization/openscientist.md | OpenScientist focused adjudication of PIK3C3 pexophagy and peroxisome localization (recovered cancelled-job artifact) | HIGH | Exact provider final_report.md recovered through supported artifact download from cancelled job dc13b854-c862-4926-82a1-8ae985f1ebe0; report endpoint refused cancelled status. Useful PAINT/yeast-complex evidence, but its cargo-confusion explanation conflicts with the actual biogenesis-localization experiments and its absence-of-human-assay argument does not refute conservation. PI3P synthesis is… |
| human | PIK3R4 | file:human/PIK3R4/PIK3R4-hypotheses/nucleus-vacuole-junction-and-nuclear-lysosomal-contacts/openscientist.md | OpenScientist focused adjudication of PIK3R4 nucleus-vacuole-junction localization | HIGH | Actual report and both CSV artifacts read. Useful primary/source leads, but its refuted/REMOVE-or-NOT recommendation is not supported: lysosome is_a lytic vacuole in current GO; the S. cerevisiae clause describes an example; late-endosomal residence, missing partner annotations and absent human assays do not establish exclusion or ancestral loss. No executed PAINT topology analysis was delivered.… |
| human | PIK3R4 | file:human/PIK3R4/PIK3R4-hypotheses/protein-phosphorylation-complex-contribution/openscientist.md | OpenScientist focused adjudication of PIK3R4 protein-phosphorylation complex contribution | HIGH | Full generated report, citations, HTML and extracted PDF read. Its removal lead for GO:0006468 conflates intrinsic MF with noncatalytic BP participation, despite explicitly admitting that complex contribution remains unresolved. It did not retrieve full PMID:8999962 or PMID:39913640, relying on abstract/seed. Its claimed executed motif code and CSV were not delivered: artifacts contain only final… |
| human | PMVK | PMID:17180682 | Localization of the pre-squalene segment of the isoprenoid biosynthetic pathway in mammalian peroxisomes. | MEDIUM | PubMed-verified citation, but its conclusion of peroxisomal localization of presqualene enzymes (including PMVK) is contradicted by PMID:14729858 and PMID:27052676 and by the UniProt CAUTION note; the field consensus is cytosolic. |
| human | PNPLA3 | PMID:22560221 | Adiponutrin functions as a nutritionally regulated lysophosphatidic acid acyltransferase. | HIGH | PubMed-verified (Cell Metab 2012;15:691-702). Correctly cited, but the central claim - that PNPLA3 is a CoA-dependent LPA acyltransferase, and that I148M increases that activity - is directly contradicted by PMID:21878620, which assayed the same reaction with CGI-58 as a positive control and detected nothing. The disagreement also runs the opposite way on the direction of the variant's effect. Re… |
| human | PNPLA3 | PMID:38844467 | PNPLA3 is a triglyceride lipase that mobilizes polyunsaturated fatty acids to facilitate hepatic secretion of large-siz… | HIGH | PubMed-verified (Nat Commun 2024;15:4847). Not cited by any current GOA row; added here because it is the loss-of-function pole of the dispute. Its in vivo findings - hepatic Pnpla3 knockout aggravating steatosis and lowering VLDL-TG - are explicitly contradicted by PMID:39550037, which reports no hepatic phenotype in Pnpla3-null mice and increased rather than decreased phospholipid PUFA in 148M… |
| human | PRSS23 | PMID:39920289 | PRSS23-eIF4E-c-Myc axis promotes gastric tumorigenesis and progression. | MEDIUM | PubMed-verified (Oncogene 2025). Independently supports non-proteolytic PRSS23 function, since eIF4E binding and progrowth phenotypes were retained in catalytic triad mutants. Marked DISPUTED rather than VERIFIED because the central claim - a secreted, signal-peptide-bearing, glycosylated protein binding the cytosolic translation initiation factor eIF4E - is topologically hard to reconcile with t… |
| human | RELN | PMID:11689558 | Reelin is a serine protease of the extracellular matrix. | MEDIUM | Correctly cited primary report, but its protease conclusion was directly challenged by PMID:20522975. |
| human | RFT1 | PMID:11807558 | Translocation of lipid-linked oligosaccharides across the ER membrane requires Rft1 protein. | HIGH | PubMed-verified. The founding genetic paper for the flippase model. Correctly cited, but its conclusion is the contested one - the authors show Rft1 is "required for" translocation in vivo, which later biochemistry (PMID:18668045, PMID:19494107) and genetics (PMID:23720757) showed does not entail that Rft1 catalyses the step. A formal Brief Communication Arising was published against it. |
| human | RFT1 | Reactome:R-HSA-4570573 | Defective RFT1 does not flip the N-glycan precursor | MEDIUM | The reaction name asserts the canonical flippase model as settled. That model is contested in the primary literature (PMID:19494107, PMID:23720757, PMID:42417535), so the two GOA rows sourced from it should not be read as independent confirmation of the mechanism. |
| human | RIMBP2 | file:human/RIMBP2/RIMBP2-hypotheses/function-hypothesis-go-0007274/openscientist.md | AIGR Gene Hypothesis Deep Research — RIMBP2 and Neuromuscular Synaptic Transmission (GO:0007274) | HIGH | Read the report and evidence/decision CSV artifacts, checked its key primary papers and independently reproduced the O15034 QuickGO query. Useful synthesis of studied synapse types and current annotation provenance. Its recommendation against inherited annotation solely from missing target experiments is not adopted. Lack of skeletal-muscle expression does not exclude a presynaptic motor-neuron r… |
| human | SAMD8 | PMID:35503176 | Sphingomyelin Synthase Family and Phospholipase Cs. | MEDIUM | Secondary review supports PE-PLC physiology, but its universal PE-only and absent-in-vivo-ceramide-regulation framing requires qualification by conflicting human PMID:33621517 and conditional mouse PMID:40998032 results. |
| human | SAMD8 | file:human/SAMD8/SAMD8-deep-research.md | Deep research on SAMD8 function | HIGH | Useful synthesis, but universal PE-only/predominant-in-vivo activity, exclusive ER residency and universal apoptosis dependence are too strong. Primary PMID:33621517 shows broader human hydrolysis; PMID:28120887 shows a minor wild-type Golgi pool and viable stable knockout cells. |
| human | SAMD8 | file:human/SAMD8/SAMD8-hypotheses/lipid-reaction-spectrum-pathways-and-compartments/openscientist.md | OpenScientist focused SAMD8 reaction-spectrum, pathway and compartment report | HIGH | Report body and all three artifacts read. Supports PE-dependent CPE, ER residence and broad hydrolase capacity, but GO:0002950 and GO:0047493 recommendations conflate donor chemistry; wild-type Golgi is misread as mutant-only; pairwise identity is not PAINT topology; physiological negatives omit the conditional 2025 SPT evidence. Explicitly could not retrieve 34332077/38388831. No alignment scrip… |
| human | SEPHS1 | PMID:20471958 | Human selenophosphate synthetase 1 has five splice variants with unique interactions, subcellular localizations and exp… | HIGH | PubMed-verified; full text available, and correctly cited. Key source for the in-vitro catalytic deficiency, isoform localizations, homodimerization, and cell-cycle/growth links. Marked DISPUTED for one specific statement rather than for the paper as a whole: its "SPS1 has an arginine at the position corresponding to Sec in SPS2" is asserted generically for SPS1 and does not hold for the human pr… |
| human | SLC25A19 | PMID:11226231 | The human mitochondrial deoxynucleotide carrier and its role in the toxicity of nucleoside antivirals. | MEDIUM | PubMed-verified. Foundational paper but its central claim (SLC25A19 is the mitochondrial deoxynucleotide carrier) was overturned by later knockout and transport studies; the physiological substrate is thiamine pyrophosphate, not deoxynucleotides. |
| human | SLC45A4 | PMID:41266324 | SLC45A4 encodes a peroxisomal putrescine transporter that promotes GABA de novo synthesis. | HIGH | PubMed-verified: Colson et al., Nat Commun 16(1):10198, 2025 (Nov 20), PMC12634670. Correctly cited and methodologically careful - endogenous protein, validated antibody, two independent peroxisome isolation methods. Marked DISPUTED because its central localisation claim (peroxisome membrane) is incompatible, as stated, with the plasma-membrane localisation reported three months earlier in Nature… |
| human | SLC52A2 | PMID:17197387 | Cloning and functional characterization of a gamma-hydroxybutyrate receptor identified in the human brain. | MEDIUM | Cloned GPR172A/GHBh1 (= SLC52A2) as a putative GHB receptor. The GHB-receptor function is now superseded by the protein's identity as riboflavin transporter RFVT2; UniProt retains it only as "Probable". Basis of the GO:0062124 IDA (marked over-annotated) and one plasma-membrane IDA (accepted; localization is sound). |
| human | SLC8B1 | PMID:20018762 | NCLX is an essential component of mitochondrial Na+/Ca2+ exchange. | HIGH | PNAS 2010; verified. The paper that identified NCLX as the mitochondrial Na+/Ca2+ exchanger and the source of most of this gene's experimental annotations. Correctly cited, but its central identification is now directly challenged by PMID:40931067, which reports intact mito-NCX in four independent NCLX-knockout lines. |
| human | SLC8B1 | PMID:40691517 | TMEM65 functions as the mitochondrial Na(+)/Ca(2+) exchanger. | MEDIUM | Nat Cell Biol 2025; verified. Cited here for context: it assigns mitochondrial Na+/Ca2+ exchange to TMEM65 rather than NCLX. Its conclusion is contradicted by PMID:40200126, and its liposome assay measures 45Ca2+ accumulation rather than transmembrane flux (PMID:41061666). |
| human | SLC8B1 | PMID:40931067 | Structure and mechanism of the mitochondrial calcium transporter NCLX. | HIGH | Nature 2025; verified (title, journal and year match the cached record). Cryo-EM structures plus functional reassignment of NCLX to H+/Ca2+ exchange, and the report that mito-NCX survives NCLX knockout. Correctly cited, but contradicted on the counterion by PMID:42431881 and on the transport-site residues by PMID:28130126. Note that it shares senior authors with PMID:40691517 (TMEM65 as mito-NCX)… |
| human | SLC8B1 | PMID:42431881 | Molecular mechanisms of mitochondrial Ca(2+) exchanger NCLX. | HIGH | Nat Commun 2026; verified. Independent cryo-EM study that reproduces the structural observation (missing Na+-coordinating residues, trimeric assembly) but concludes broader selectivity rather than H+ specificity, and explicitly declines to adjudicate the TMEM65 question. |
| human | SPG21 | PMID:26978163 | Loss of Maspardin Attenuates the Growth and Maturation of Mouse Cortical Neurons. | LOW | PubMed-verified (Neurodegener Dis 2016;16:260-72). Reports reduced axonal branching in SPG21-null mouse cortical neurons - the opposite direction to PMID:20661613, which reported increased branching in the same knockout. The 2026 J Cell Biol paper notes the discrepancy explicitly. Cited here only to record that the neuronal morphology phenotype is unsettled; no annotation is derived from it. Cach… |
| human | SULT1B1 | PMID:36805701 | Histone tyrosine sulfation by SULT1B1 regulates H4R3me2a and gene transcription. | HIGH | PubMed-verified (Nat Chem Biol 2023;19:855-864). The claim that SULT1B1 is a histone sulfotransferase acting on H3 Tyr99 is formally contested by PMID:40890505, which reannotated the original raw mass spectra as a phosphotyrosine peptide and found no sulfation of recombinant histone H3.2 by recombinant SULT1B1. The authors replied (PMID:40890506) and an Author Correction was published (PMID:40890… |
| human | SULT1B1 | PMID:40890505 | Mass spectrometry and enzyme assays refute histone tyrosine sulfation. | HIGH | PubMed-verified (Nat Chem Biol 2025;21:1667-1670), full text available via PMC12978989. A formal Matters Arising refuting PMID:36805701 on three independent grounds: sulfotyrosine is too labile to give the 100% sulfate-retaining HCD fragments reported, the reported fragment masses fit phosphotyrosine to <3 ppm versus 12.5-31.5 ppm for sulfotyrosine, and a recombinant SULT1B1 assay that efficientl… |
| human | SULT1B1 | PMID:40890506 | Reply to: Mass spectrometry and enzyme assays refute histone tyrosine sulfation. | HIGH | PubMed-verified (Nat Chem Biol 2025;21:1671-1674). The original authors' reply to PMID:40890505. Cached record is abstract-only and PubMed carries no abstract for it, so the substance of their rebuttal could not be assessed here. Its existence is why this review treats the question as open rather than settled. |
| human | SULT1B1 | PMID:41686426 | GAL3ST1-Mediated Histone Tyrosine Sulfation Induced by Cancer-Associated Fibroblasts Promotes Gastric Cancer Metastasis. | MEDIUM | PubMed-verified (Cancer Res 2026;86:2429-2446), full text available via PMC13176828. Relevant to SULT1B1 because it attributes the same H3Y99sulf mark to a different enzyme, GAL3ST1, and reports SULT1B1 levels unchanged under the inducing conditions. It cites PMID:36805701 as settled background and does not cite the refutation PMID:40890505, and its H3Y99sulf readouts rest on the antibody class t… |
| human | TBCK | PMID:23977024 | TBCK influences cell proliferation, cell size and mTOR signaling pathway. | MEDIUM | PubMed-verified (PLoS One 2013). Correctly cited, but its central mTOR claim has not reproduced in human neural progenitors or consistently across patient-derived cells, and the proliferation, cell-size and actin phenotypes are siRNA knockdown readouts several steps from any molecular activity of TBCK. |
| human | TMEM120A | PMID:32084332 | TACAN Is an Ion Channel Involved in Sensing Mechanical Pain. | HIGH | PubMed-verified (Beaulieu-Laroche et al., Cell 2020). This is the origin of the TACAN name and of every channel-related annotation on this gene, including the mouse donor annotation behind the human GO_REF:0000024 ISS rows. The channel claim was not reproduced by four independent groups in 2021, and GO now carries three human NOT|enables annotations against the same term. Correctly cited but the… |
| human | TMEM120A | PMID:33232830 | Involvement of TACAN, a Mechanotransducing Ion Channel, in Inflammatory But Not Neuropathic Hyperalgesia in the Rat. | MEDIUM | PubMed-verified (J Pain 2020). Antisense knockdown in rat supports a contribution to inflammatory mechanical hyperalgesia, which is relevant to the pain-process annotations. Marked DISPUTED only because the paper adopts the mechanotransducing-ion-channel framing that has since been refuted; the behavioural observation stands independently of that framing. |
| human | TMEM120A | PMID:41967766 | From nociception to therapy: The expanding role of TMEM proteins in pain. | LOW | PubMed-verified (Life Sci 2026); cached record is abstract-only. A narrative review that still lists TMEM120A/TACAN among pain-relevant TMEM proteins. Its own wording is "modulating ion channels", which is compatible with the PIEZO2 model rather than with TMEM120A being a channel. Recorded to document that the channel framing persists in secondary literature; it presents no new data and is not a… |
| human | TMEM65 | PMID:26403541 | Evolutionarily conserved intercalated disc protein Tmem65 regulates cardiac conduction and connexin 43 function. | MEDIUM | Nat Commun 2015; verified. Correctly cited for the intercalated-disc/connexin-43 role, but the surface localisation it reports has not been reproduced by any subsequent mitochondrial study of TMEM65 and is hard to reconcile with a cleaved mitochondrial targeting sequence. Cached record is abstract-only. |
| human | TMEM65 | PMID:40200126 | TMEM65 regulates and is required for NCLX-dependent mitochondrial calcium efflux. | HIGH | Nat Metab 2025; verified (title, journal and year match the cached record). Source of the NOT|enables GO:0005432 annotation. Correctly cited and methodologically sound, but its central conclusion - TMEM65 is a regulator of NCLX, not the exchanger - is directly contradicted by PMID:40691517. Recorded as DISPUTED rather than MISCITED: the paper is cited for exactly what it claims. |
| human | TMEM65 | PMID:40691517 | TMEM65 functions as the mitochondrial Na(+)/Ca(2+) exchanger. | HIGH | Nat Cell Biol 2025; verified (title, journal and year match the cached record). Source of the positive enables GO:0005432 annotation. Correctly cited, but the key liposome reconstitution measures 45Ca2+ accumulation rather than transmembrane flux, a limitation raised in PMID:41061666, and the conclusion is contradicted by PMID:40200126. Note also that this paper shares senior authors with PMID:40… |
| human | TMEM65 | PMID:40931067 | Structure and mechanism of the mitochondrial calcium transporter NCLX. | MEDIUM | Nature 2025; verified. Cited here for context only - it does not assay TMEM65 (the word does not appear in the paper). Relevant because it reassigns NCLX to H+/Ca2+ exchange, which would remove the alternative candidate for mito-NCX and so indirectly supports Position A. |
| human | TNFRSF21 | PMID:19225519 | APP binds DR6 to trigger axon pruning and neuron death via distinct caspases. | HIGH | RETRACTED. PubMed-verified as Nikolaev et al., Nature 2009, "APP binds DR6 to trigger axon pruning and neuron death via distinct caspases"; retracted January 2024 (see PMID:38110576) for image duplication, statistical errors and superseded mechanistic conclusions. Listed here only because it is the origin of the APP-receptor model for DR6. Critically, it is cited by no GO annotation on this gene… |
| human | TNFRSF21 | PMID:28285993 | Death Receptor 6 Promotes Wallerian Degeneration in Peripheral Axons. | HIGH | PubMed-verified (Curr Biol 2017, Gamage et al.). The claim that DR6 is required for Wallerian degeneration and Schwann cell remodelling after axotomy. Directly non-replicated by PMID:41891813 using two knockout lines including the original. |
| human | TRMT6 | PMID:29072297 | The m1A landscape on cytosolic and mitochondrial mRNA at single-base resolution. | HIGH | PubMed-verified (Nature 2017, 551(7679), 251-255). Correctly cited, and itself a conservative study, reporting cytosolic mRNA m1A as rare, sub-stoichiometric and confined to tRNA T-loop-like structures while attributing those sites to the TRMT6/TRMT61A complex. Marked DISPUTED because the underlying claim of internal mRNA m1A is contested by later biochemical validation (PMID:42337368), not becau… |
| human | TRMT6 | PMID:29107537 | Base-Resolution Mapping Reveals Distinct m(1)A Methylome in Nuclear- and Mitochondrial-Encoded Transcripts. | HIGH | PubMed-verified (Mol Cell 2017, 68(5), 993-1005). Independent base-resolution study reaching the same narrow conclusion about a GUUCRA tRNA-like motif class of TRMT6/61A-dependent sites. Same DISPUTED rationale as PMID:29072297. The cached entry is abstract-only. |
| human | TRMT6 | PMID:41103012 | RNA m(1)A methyltransferase TRMT61A promotes colorectal tumorigenesis by enhancing ONECUT2 mRNA stability and is a pote… | MEDIUM | PubMed-verified (Cancer Commun 2025, 45(12), 1616-1644). Same status as PMID:42003777, and reviewed in more detail on the TRMT61A review. Not used to support any annotation here. |
| human | TRMT6 | PMID:42003777 | TRMT6-Mediated m(1)A Modification of CDK9 mRNA is a Dual-Pronged Pathogenic Driver for HBV-Related Hepatocellular Carci… | MEDIUM | PubMed-verified (Adv Sci 2026, 13(39), e14172). Reports TRMT6-dependent mRNA m1A stabilising CDK9 in hepatocellular carcinoma. Cited to document the opposing stream, not to support an annotation, since the mechanism rests on the class of mapping method challenged by PMID:42337368 and the reported direction of effect is opposite to that in PMID:40897821. |
| human | TRMT61A | PMID:29072297 | The m1A landscape on cytosolic and mitochondrial mRNA at single-base resolution. | HIGH | PubMed-verified (Nature 2017, 551(7679), 251-255). Correctly cited, and itself a conservative, sceptical study, since it reports cytosolic mRNA m1A as rare, sub-stoichiometric and confined to tRNA T-loop-like structures while explicitly attributing those sites to the TRMT6/TRMT61A complex. Marked DISPUTED rather than MISCITED because the underlying claim of internal mRNA m1A is contested by later… |
| human | TRMT61A | PMID:29107537 | Base-Resolution Mapping Reveals Distinct m(1)A Methylome in Nuclear- and Mitochondrial-Encoded Transcripts. | HIGH | PubMed-verified (Mol Cell 2017, 68(5), 993-1005). Independent base-resolution study reaching the same narrow conclusion about a GUUCRA tRNA-like motif class of TRMT6/61A-dependent sites. Same DISPUTED rationale as PMID:29072297. The cached entry is abstract-only. |
| human | TRMT61A | PMID:41103012 | RNA m(1)A methyltransferase TRMT61A promotes colorectal tumorigenesis by enhancing ONECUT2 mRNA stability and is a pote… | MEDIUM | PubMed-verified (Cancer Commun 2025, 45(12), 1616-1644). Reports TRMT61A-dependent mRNA m1A stabilising ONECUT2 in colorectal cancer. Cited here to document the opposing stream, not to support an annotation. The mechanism rests on m1A-sequencing, the class of method challenged by PMID:42337368, and the reported direction of effect on target mRNA stability is opposite to that in PMID:40897821. |
| human | TRMT61A | PMID:42003777 | TRMT6-Mediated m(1)A Modification of CDK9 mRNA is a Dual-Pronged Pathogenic Driver for HBV-Related Hepatocellular Carci… | MEDIUM | PubMed-verified (Adv Sci 2026, 13(39), e14172). Same status as PMID:41103012. It documents the ongoing mRNA-m1A cancer literature but depends on the contested mapping methods, and is not used to support any annotation. |
| human | TTC19 | PMID:20208530 | PtdIns(3)P controls cytokinesis through KIF13A-mediated recruitment of FYVE-CENT to the midbody. | LOW | Reports TTC19 as a ZFYVE26/CHMP4B partner with centrosome/midbody localization and a cytokinesis role. UniProt records a CAUTION that the midbody localization could not be confirmed by others (PMID:21278747); not the core mitochondrial function of TTC19. |
| human | TUSC3 | PMID:19717468 | Mammalian MagT1 and TUSC3 are required for cellular magnesium uptake and vertebrate embryonic development. | MEDIUM | PubMed-verified citation. Experimental basis for the magnesium-transport annotations, but the interpretation as a direct plasma-membrane Mg2+ transporter is disputed by later work (PMID:25135935, PMID:26864433) that localizes TUSC3/MAGT1 to the ER as OST oxidoreductase subunits; any Mg2+ role is likely indirect. |
| human | UBAC2 | file:human/UBAC2/UBAC2-hypotheses/function-hypothesis-go-0004252/openscientist.md | OpenScientist hypothesis investigation - UBAC2 serine-type endopeptidase activity | HIGH | The catalytic-loss conclusion is supported by primary evidence and the cached sequence/topology. Do not adopt the report's unverified assertion that PANTHER topology itself is wrong, its treatment of PMID:23297223 as the original IBA reference (GOA cites GO_REF:0000033), or its proposed GO:0140318 ER-phagy label: QuickGO identifies that ID as protein transporter activity. |
| human | YTHDF3 | PMID:28106072 | YTHDF3 facilitates translation and decay of N(6)-methyladenosine-modified RNA. | HIGH | Full text cached. Correctly cited and a primary source for YTHDF3 function, but its translation-promoting conclusion is directly contradicted by PMID:32492408; the mRNA-decay conclusion is not. |
| human | YTHDF3 | PMID:28106076 | Cytoplasmic m(6)A reader YTHDF3 promotes mRNA translation. | HIGH | Full text cached. Correctly cited; the m6A-binding and ribosome-association data stand, but the translation-promotion conclusion is the one that PMID:32492408 could not reproduce and that GOA now also carries as a NOT annotation. |
| human | YTHDF3 | PMID:32194978 | Cytoplasmic m(1)A reader YTHDF3 inhibits trophoblast invasion by downregulation of m(1)A-methylated IGF1R. | MEDIUM | Full text cached. Correctly cited for YTHDF3, and the IGF1R turnover and trophoblast phenotypes are real; but the m1A-reader framing depends on m1A-seq, whose specificity was challenged the previous year (PMID:31719534). |
| human | YTHDF3 | PMID:40715766 | YTHDF3 recognizes DNA N6-methyladenine and recruits ALKBH1 for 6mA removal from genomic DNA. | MEDIUM | PubMed-verified (Chen et al., EMBO J 2025). Full text cached. Proposes an entirely new substrate class for YTHDF3 - N6-methyladenine in genomic DNA. Recorded here because it is the most recent functional claim about this gene, but not annotated: the mark being read is itself contested (PMID:32206710, PMID:32203414, PMID:35113693, PMID:41254163), the paper's own abstract concedes that the ALKBH1 d… |
| mouse | Bcl2 | file:human/BCL2/BCL2-hypotheses/conditional-proapoptotic-and-channel-capacities/openscientist.md | OpenScientist BCL2 conditional proapoptotic and channel-capacity adjudication | HIGH | Read complete generated Output, distinct from prompt. Supports conditional proapoptotic and channel capacities, but it did not inspect actual PAINT topology, does not settle pore stoichiometry, and its sequence-only mouse discussion misses direct mouse experiments PMID:18835031 and PMID:18443228. Acidic-pH carboxyfluorescein efflux is distinct from neutral-pH small-ion channels. No provider seque… |
| rat | Aldh1l1 | file:rat/Aldh1l1/Aldh1l1-hypotheses/cofactor-specificity-and-mitochondrial-localization/openscientist.md | OpenScientist assessment of rat Aldh1l1 cofactor specificity and mitochondrial localization | HIGH | Read the complete output and inspected its N-terminal CSV. The mitochondrial targeting distinction was verified against the rat P28037 N terminus and primary PMID:20498374/21238436. The report headline overstates NAD+ refutation: its own limitations acknowledge no NAD+ kinetics, and the structural study says NADP preference rather than zero NAD usage. Its repeated weaker-evidence-code framing and… |
| rat | Hmgcs2 | file:rat/Hmgcs2/Hmgcs2-hypotheses/mevalonate-and-isoprenoid-synthesis/openscientist.md | OpenScientist assessment of Hmgcs2 mevalonate and isoprenoid synthesis | HIGH | Read the full report and both CSV artifacts. Ketogenic physiology is consistent with the existing review, but the absolute compartment exclusion and proposed NOT are not established. The report missed PMID:7961793 despite the hypothesis naming complementation, mistakes the live isoprenoid IEA as absent, and treats IBA evidence as a weak automatic transfer. Live QuickGO and primary-abstract checks… |
| worm | clk-1 | PMID:25961505 | A nuclear role for the respiratory enzyme CLK-1 in regulating mitochondrial stress responses and longevity. | HIGH | Full-text primary paper reporting a distinct nuclear CLK-1/COQ7 pool that regulates ROS metabolism, the UPRmt, transcription, and longevity independently of ubiquinone. The data are genuine but the nuclear/non-mitochondrial model is debated (worm N-terminus lacks the COQ7 NTS; limited independent confirmation). Underlies the nucleus, ROS-regulation, and transcription-regulation annotations, which… |
| worm | csr-1 | file:worm/csr-1/csr-1-hypotheses/corrected-argonaute-mirna-binding-specificity/openscientist.md | OpenScientist focused assessment of CSR-1 miRNA-binding specificity | HIGH | Full report body, evidence/decision CSVs and corresponding HTML read. Useful identity verification and 22G/isoform specificity synthesis are retained. The recommendation to REMOVE or add NOT is not supported by its admitted absence of a miRNA-binding exclusion test; IBA-only status is not evidence against the ancestral assertion. The report used PubMed abstracts and did not independently derefere… |
| worm | eat-3 | file:worm/eat-3/eat-3-hypotheses/microtubule-and-peroxisome-capacities/openscientist.md | EAT-3: microtubule and peroxisome capacities | HIGH | Full report read. Correctly emphasizes mitochondrial fusion/topology and recovers the relevant ancestral nodes. However, both challenged nodes are ancestors of the OPA1 node, not transfers from an exclusively separate DRP1 clade. Missing conventional dynamin modules do not establish loss of all binding or process capacities. No executable tree/domain artifact was supplied. Full PMID:32228866 expl… |
| worm | pgl-1 | file:worm/pgl-1/pgl-1-deep-research.md | OpenAI deep research report on pgl-1 (C. elegans) | HIGH | The synthesis correctly connects PGL self-association, IFE-1 recruitment, RNase and germline phenotypes. Its suggested in vivo RNA-catabolic function and direct SIR-2.1 sequestration exceed demonstrated mechanism; several GO identifiers in the prose are incorrect. Autophagic removal treats PGL as cargo and is not evidence that PGL executes autophagy. |
| worm | pgl-1 | file:worm/pgl-1/pgl-1-hypotheses/pgl-fold-and-inherited-helicase-rna-processing/openscientist.md | OpenScientist focused adjudication of PGL-1 fold and inherited helicase, splicing and export capacities | HIGH | Full report, HTML/PDF artifacts and citation list assessed. Useful primary RNase/FBF-2 context is incorporated. Stronger refutation and GLH-partner-confusion claims were not reproduced. Independent sequence controls, live InterPro coordinates, actual tree and full 2016/2021 primary scope are documented in pgl-1-report-assessment.md. Existing source IEA RNase-T1 evidence code is preserved despite… |
| worm | prg-1 | file:worm/prg-1/prg-1-hypotheses/nuclear-localization-and-pirna-processing/openscientist.md | OpenScientist focused adjudication of PRG-1 nuclear localization and piRNA processing | HIGH | Actual report incorporated with primary checks. A positive cytoplasm comment and perinuclear imaging do not refute every nuclear pool or justify NOT. PMID:26919432 explicitly models trimming before/independently of PRG-1 loading, so mutant precursor association alone does not demonstrate normal precursor scaffolding during trimming. The report's noncatalytic wording must also be restricted to mat… |
| yeast | ACL4 | file:yeast/ACL4/ACL4-hypotheses/function-hypothesis-go-0008320/openscientist.md | OpenScientist hypothesis report for ACL4 GO:0008320 | HIGH | The seed/report omitted the source contributes_to qualifier. Lack of a transmembrane helix does not refute contribution by a soluble complex subunit; known Rpl4 chaperoning is not evidence that all other roles are impossible. PAINT topology/branch-specific loss was not demonstrated. |
| yeast | DCV1 | file:yeast/DCV1/DCV1-deep-research-falcon.md | Falcon research report on yeast DCV1 | MEDIUM | Actual complete narrative and artifact table read. Correctly labels most proposals as family inference; it misses the PMID:33002606 curated nuclear-envelope source and PMID:22042866 genetic-screen source. Therefore its claim that no direct cell-biological characterization exists is incomplete, and its specific Rim101/scaffold assignments cannot be promoted to target molecular functions. Existing… |
| yeast | ESL1 | file:yeast/ESL1/ESL1-deep-research-falcon.md | Falcon literature report for yeast ESL1 | MEDIUM | Read full scientific body. Useful EST/SMG/PIN background, but it missed PMID:23893744 and Esl2 and therefore cannot adjudicate the target negatives or establish target telomere/NMD/nuclease functions. The dedicated full-text study takes precedence. No exact OpenScientist report existed at the initial cache check; the newly generated focused report is separately assessed here. |
| yeast | ESL1 | file:yeast/ESL1/ESL1-hypotheses/telomerase-association-and-nucleic-acid-binding/openscientist.md | OpenScientist focused assessment of ESL1 telomerase association and nucleic-acid binding | HIGH | Entire report, HTML version and all three CSV artifacts read. The actual PAINT lineage already verifies positive IBD PTN000403280 above target PTN007651903; a family/subfamily label is not the origin or validity test of that assertion. Zero STRING experimental-channel scores mean no recorded positive link, not tested noninteraction; uniform database scores do not identify their underlying source.… |
| yeast | HST1 | file:yeast/HST1/HST1-deep-research-openscientist.md | OpenScientist GO-focused literature synthesis for yeast HST1 | HIGH | Actual report reviewed with primary sources. Its opening H3K9/K14/H4K16 list derives partly from Sir2/Hst2; it is not a direct site-specific Hst1 assay. Its native-role language must not erase conditional SUM1-1 silencing or the H4K16-positive loci in PMID:18990212. |
| yeast | ILT1 | file:yeast/ILT1/ILT1-deep-research-falcon.md | Falcon literature report for yeast ILT1 | MEDIUM | Full body reviewed. Its identity/domain and comparative transporter synthesis is informative, but its assertion of no target localization and its naming history miss PMID:30045857. Primary PM microscopy supplements rather than excludes the inherited vacuolar role. No exact cached OpenScientist report was found. |
| yeast | MIM1 | PMID:19345216 | The three domains of the mitochondrial outer membrane protein Mim1 have discrete functions in assembly of the TOM compl… | HIGH | Not in the GOA. Corroborates the topology, but its central claim is in direct tension with PMID:18177669 from the same period: this paper assigns "discrete functions" to the three domains and attributes regulation of the SAM early reaction to the N-terminal domain, whereas Popov-Celeketic et al. found that deleting either soluble domain caused no growth or TOM-assembly phenotype and concluded the… |
| yeast | SSA1 | file:yeast/SSA1/SSA1-hypotheses/existing-go-0005886-keep-as-non-core/openscientist.md | OpenScientist hypothesis review of the SSA1 plasma-membrane annotations | HIGH | The report's evidence-weight conclusion supports non-core treatment of the stripped-plasma-membrane HDA, but its live QuickGO claim that P10591 has no plasma-membrane IBA conflicts with the pinned SSA1 GOA snapshot, which contains GO:0005886 IBA from GO_Central dated 2025-09-03. The live-query absence claim and recommendation to correct the evidence code are not used. |
| yeast | SSA2 | file:yeast/SSA2/SSA2-hypotheses/existing-go-0005886-keep-as-non-core/openscientist.md | OpenScientist assessment of SSA2 plasma-membrane localization as non-core | MEDIUM | The report supports KEEP_AS_NON_CORE and correctly emphasizes the weak, high-throughput fractionation evidence. However, its live QuickGO result claims that no IBA plasma-membrane annotation exists, whereas the repository's pinned SSA2 GOA contains both GO_REF:0000033 IBA and PMID:16622836 HDA rows for GO:0005886. Its use of the histatin paper is valid because that study directly assays S. cerevi… |
| yeast | SSA3 | file:yeast/SSA3/SSA3-hypotheses/existing-go-0005886-keep-as-non-core/openscientist.md | OpenScientist review of the SSA3 plasma-membrane IBA annotation | HIGH | The report's conclusion that SSA3 lacks target-specific plasma-membrane evidence is useful, but its live QuickGO claims that the P09435 IBA was retired and that Ssa4 had no such annotation conflict with the pinned SSA3/SSA4 GOA snapshots. Those live-query claims are not used as evidence for removal. |
| yeast | SSA4 | file:yeast/SSA4/SSA4-hypotheses/existing-go-0005886-keep-as-non-core/openscientist.md | OpenScientist hypothesis review of the SSA4 plasma-membrane IBA annotation | HIGH | The report correctly found no target-specific plasma-membrane evidence and its live absence result is corroborated by the current local PAINT snapshot. It nevertheless mischaracterizes the pinned SSA4 GOA snapshot, which does contain the 2025 IBA, and speculates about donor origins; those claims are not used. |
| yeast | SSQ1 | file:yeast/SSQ1/SSQ1-deep-research-falcon.md | Falcon (Edison) deep research report: SSQ1 (Q05931), mitochondrial Hsp70 chaperone for Fe-S cluster biogenesis | HIGH | Report and evidence table read in full and checked against primary sources. Core ISC-transfer synthesis is supported. Quantitative claim of full rescue requiring 1000-2000-fold Ssc1 overexpression is misleading: PMID:10779357 measured twofold Ssc1 increase, approximately 2000-fold relative to baseline Ssq1, with only partial rescue. Narrow native specificity does not negate the observed in-vitro… |
| yeast | YAR1 | file:yeast/YAR1/YAR1-hypotheses/function-hypothesis-go-0001228/openscientist.md | OpenScientist hypothesis review of YAR1 GO:0001228 | HIGH | The domain distinction informs GO:0001228, but blanket removal of regulation/MBF/SBF does not follow: non-DNA-binding subunits need not carry APSES. Bulk acidity is not a sufficient exclusion test, and actual PAINT topology was not recovered. |
LOW_QUALITY (172)
| Organism | Gene | Reference | Title | Rel. | Notes |
|---|---|---|---|---|---|
| BRUMA | cpi-2 | GO_REF:0000118 | TreeGrafter-generated GO annotations | LOW | Source of generic, phylogenetically propagated cytoplasm and vesicle localizations that are not specifically supported for this secreted protein. |
| HETGA | Cd44 | PMID:11944887 | Hyaluronidases and CD44 undergo differential modulation during chondrogenesis. | LOW | The human source behind the cartilage development row, annotated IEP. The citation is correct, but an expression-pattern correlation is the weakest experimental basis for a process annotation and does not survive projection into another species. Title verified against PubMed; no quotation used. |
| HETGA | Cd44 | PMID:19577615 | bFGF induces changes in hyaluronan synthase and hyaluronidase isoform expression and modulates the migration capacity o… | LOW | The human source behind the cellular response to fibroblast growth factor stimulus row. The study measures bFGF-driven changes in hyaluronan synthase and hyaluronidase expression in a fibrosarcoma line, which is a weak basis for asserting that CD44 itself mediates an FGF response. Title verified against PubMed; no quotation used. |
| HETGA | Cgas | file:HETGA/Cgas/Cgas-deep-research-falcon.md | Functional Annotation Report: Heterocephalus glaber Mb21d1 / cGAS | MEDIUM | The report supplies mammalian DNA sensing, cGAMP-STING signaling and nucleosome restraint, but missed the direct naked-mole-rat HR study PMID:41066557. Its human/mouse sequestration model does not demonstrate loss of naked-mole-rat immune signaling. Use conserved mechanism with this recall limitation; do not use lack of retrieved species assays to refute inherited function. |
| HETGA | Scn9a | file:HETGA/Scn9a/Scn9a-deep-research-affinage-human-ortholog.md | Affinage mechanistic annotation for human SCN9A used as a conserved-mechanism baseline | MEDIUM | Human-ortholog record used only as a conserved-mechanism baseline, never as evidence about the naked mole-rat protein, and no sentence from it is quoted as supporting_text anywhere in this review. Marked LOW_QUALITY rather than VERIFIED because the record tripped a soft trust gate: its own frontmatter reports self_evaluation_pairwise as tie rather than win against the curated UniProt reference, s… |
| HETGA | Tac1 | file:HETGA/Tac1/Tac1-deep-research-affinage-human-ortholog.md | Affinage mechanistic annotation for TAC1 (human) | LOW | Machine-generated record for the HUMAN ortholog (P20366), supplied as a conserved-mechanism baseline because affinage does not run on non-human species. Two limitations, recorded because provider recall is being measured. First, it contains no naked mole-rat content: none of its 36 citations is a naked mole-rat paper, and none of the six papers that actually decide this review appears in it, so i… |
| PSEPK | PP_0075 | file:PSEPK/PP_0075/PP_0075-deep-research-openscientist.md | OpenScientist deep-research report for PSEPK PP_0075 | MEDIUM | Architecture and absence-of-direct-assay observations were retained. The report's proposed sulfate cargo and assignment of intact COS uptake to an unspecified ABC route are speculative and are not adopted. |
| PSEPK | PP_0076 | file:PSEPK/PP_0076/PP_0076-deep-research-openscientist.md | OpenScientist deep-research report for PSEPK PP_0076 | MEDIUM | The explicit knowledge gaps were retained. The report's definitive ABC complex model is rejected because it incorrectly identifies PP_0074/Q88RQ5 as an ABC ATPase; PP_0074 is shikimate dehydrogenase AroE, and PP_0075 has SLC26A/SulP-STAS rather than ABC-permease architecture. |
| PSEPK | PP_2928 | file:PSEPK/PP_2928/PP_2928-deep-research-openscientist.md | OpenScientist functional review of PP_2928 | MEDIUM | Genuine provider artifact used for literature retrieval. Its unsupported quantitative identity and route-absence claims were not used in the final annotation rationale. |
| PSEPK | PP_2928 | file:projects/P_PUTIDA/deep-research/PSEPK__bacterial-carboxyspermidine-biosynthesis__ppu00330-deep-research-openscientist.md | OpenScientist PSEPK carboxyspermidine pathway review | MEDIUM | Useful retrieval and pathway synthesis, but internally inconsistent route claims and unreproducible identity percentages were deliberately excluded; exact family, synteny, and primary-literature evidence carry the assignment. |
| PSEPK | betC | file:PSEPK/betC/betC-deep-research-openscientist.md | OpenScientist deep-research report for PSEPK betC | HIGH | Exact target identity, motif, direct KT2440 phenotype, and absence of target-protein kinetics were retained. The report's assertion that the adjacent proteins form the importer is rejected: current PP_0075 architecture is SLC26A/SulP-STAS, and no cognate local ABC ATPase or target-specific transport assay is established. |
| PSEPK | guaD | file:PSEPK/guaD/guaD-deep-research-openscientist.md | OpenScientist deep-research report for Pseudomonas putida KT2440 guaD | MEDIUM | Useful retrieval synthesis for target identity and the family-level reaction, but it promotes inferred residues, localization, regulation, and promiscuity to target-specific claims without direct Q88F18 experiments. |
| PSEPK | moaA | file:PSEPK/moaA/moaA-deep-research-openscientist.md | OpenScientist gene research for moaA | HIGH | The conserved GTP cyclase reaction and division of labor with MoaC agree with reviewed target UniProt. The report explicitly acknowledges that its mechanistic and structural evidence comes from orthologs, but presents unsaved target residue/motif analysis, target homodimer and structural transfer, heterologous xanthine-oxidase production, and rate-defining or nonredundant physiological claims as… |
| PSEPK | moaB-I | file:PSEPK/moaB-I/moaB-I-deep-research-openscientist.md | OpenScientist gene research for moaB-I | HIGH | The report correctly identifies the MoaB family and acknowledges the absence of a Q88L15 assay, but then overstates synteny and family homology as a robust MogA-equivalent catalytic assignment. That conclusion conflicts with the target UniProt no-adenylyltransferase statement and is not used here. |
| PSEPK | moaB-II | file:PSEPK/moaB-II/moaB-II-deep-research-openscientist.md | OpenScientist gene research for moaB-II | HIGH | The report appropriately preserves the absence of a Q88E67 assay and the unresolved physiological role, consistent with the target UniProt no-MPT-adenylyltransferase statement. Its exact proteome census, pairwise identity values, genomic-context interpretation, and structural predictions are unsaved provider analyses and are not propagated as target evidence. |
| PSEPK | moaD | file:PSEPK/moaD/moaD-deep-research-openscientist.md | OpenScientist gene research for moaD | HIGH | Conserved carrier chemistry and complex membership were cross-checked against the UniProt/PANTHER family assignment. The report overstates target-specific operon, uniqueness, non-redundancy, and downstream-client conclusions; no direct KT2440 biochemical assay was located, so those claims are not used as established evidence. |
| PSEPK | moaE | file:PSEPK/moaE/moaE-deep-research-openscientist.md | OpenScientist gene research for moaE | HIGH | The conserved MoaD-MoaE reaction was cross-checked against the target UniProt entry and ortholog evidence. The report's target-specific alignment, AlphaFold, operon, and downstream-client conclusions are not independently reproducible local evidence, and no direct KT2440 enzyme assay was found. |
| PSEPK | moeA | file:PSEPK/moeA/moeA-deep-research-openscientist.md | OpenScientist gene research for moeA | HIGH | The target MoeA reaction agrees with UniProt, but the report misassigns several KT2440 MoaD/MoaE/MoeB/MobA locus tags and overstates moaB-I as the missing active MPT adenylyltransferase. Those organism-specific claims were not used; no direct KT2440 assay was found. |
| PSEPK | nspC | file:projects/P_PUTIDA/deep-research/PSEPK__bacterial-carboxyspermidine-biosynthesis__ppu00330-deep-research-openscientist.md | OpenScientist PSEPK carboxyspermidine pathway review | MEDIUM | Useful retrieval and pathway synthesis, but unreproducible identity percentages were excluded; exact PANTHER subfamily, InterPro family, synteny, and primary literature carry the decision. |
| PSEPK | speA | file:projects/P_PUTIDA/deep-research/PSEPK__bacterial-carboxyspermidine-biosynthesis__ppu00330-deep-research-openscientist.md | OpenScientist PSEPK carboxyspermidine pathway review | MEDIUM | Useful pathway-retrieval context, but its quantitative and route-absence claims were not used as annotation evidence; the exact UniProt reaction and process record ground this review. |
| PSEPK | speB | file:projects/P_PUTIDA/deep-research/PSEPK__bacterial-carboxyspermidine-biosynthesis__ppu00330-deep-research-openscientist.md | OpenScientist PSEPK carboxyspermidine pathway review | MEDIUM | Useful pathway-retrieval context, but its quantitative and route-absence claims were not used as annotation evidence; the exact UniProt reaction and process record ground this review. |
| PSEPK | xdhA | file:PSEPK/xdhA/xdhA-deep-research-openscientist.md | OpenScientist deep-research report for Pseudomonas putida KT2440 xdhA | MEDIUM | Useful synthesis of target architecture and homologous XDH biochemistry, but several exact electron-path, localization, maturation, and oligomeric claims are extrapolated from other strains or species. |
| PSEPK | xdhB | file:PSEPK/xdhB/xdhB-deep-research-openscientist.md | OpenScientist deep-research report for Pseudomonas putida KT2440 xdhB | MEDIUM | The target identity and complementary cofactor architecture are useful corroboration. The report nevertheless promotes homolog-derived stoichiometry, localization, maturation, electron-path, active-site, and substrate-range claims to Q88F20; those details were not imported. |
| SCHPO | cmp7 | PMID:16823372 | ORFeome cloning and global analysis of protein localization in the fission yeast Schizosaccharomyces pombe. | LOW | Genome-wide overexpression GFP screen; gives a coarse cytoplasmic signal for cmp7 that does not reflect its specific functional NE / SPB-attachment site compartment characterized by later work. |
| XENTR | LOC101732730 | file:XENTR/A0A8J0SCI2/A0A8J0SCI2-deep-research-falcon.md | Deep research summary for LOC101732730 in Xenopus tropicalis | MEDIUM | Adjudicated against the UniProt record, GOA WITH/FROM donors, and the 2026-08-29 re-review notes. The report targets the correct gene, is candid that no direct experimental studies exist, and its general C2H2-ZF mechanism claims (the only parts used as supporting_text here) are sound and consistent with UniProt/PANTHER. Marked LOW_QUALITY because it is an LLM-generated inference document with no… |
| XENTR | aldh1a2 | file:XENTR/aldh1a2/aldh1a2-deep-research.md | Deep research synthesis for aldh1a2 | MEDIUM | LLM-generated synthesis whose citations are largely to Wikipedia and secondary web pages rather than primary literature, and which contains at least one factual slip, giving the protein length as approximately 519 aa where the UniProt record is 511 aa. Used only where a statement could be quoted verbatim and is corroborated by a primary source or by the UniProt record. |
| human | A2ML1 | file:human/A2ML1/A2ML1-deep-research-affinage.md | Affinage mechanistic annotation for A2ML1 (human) | MEDIUM | Trust gate TRIPPED - affinage's own head-to-head self-evaluation scored this record pairwise = loss against the curated UniProt reference (recorded in the record's own self_evaluation_pairwise frontmatter field), so it was treated with extra scepticism and every claim used was re-verified against the cited PMIDs and UniProt. The narrative did hold up on the core biology and it usefully reports bo… |
| human | A3GALT2 | PMID:32915357 | APOL1 polymorphism modulates sphingolipid profile of human podocytes. | LOW | A3GALT2 is incidental to an APOL1/podocyte study. The negative expression result is data-not-shown from one immortalized podocyte model and cannot support organism-wide non-expression. |
| human | ABHD14A | file:human/ABHD14A/ABHD14A-deep-research-affinage.md | Affinage mechanistic annotation for ABHD14A (human) | LOW | Machine-fetched Affinage record. Gates passed, but it must not be used as it stands, for a reason worth recording for the whole campaign: its two most substantive claims - that a truncated ABHD14A hydrolyses short-chain esters with CoA enhancement, and that the full-length protein localises to the Golgi - both cite 'PMID:bio_10.1101_2025.11.28.691245', which is not a PubMed identifier at all but… |
| human | ABHD8 | file:human/ABHD8/ABHD8-deep-research-affinage.md | Affinage mechanistic annotation for ABHD8 (human) | MEDIUM | Affinage record whose own self_evaluation_pairwise scored 'tie', not 'win', against curated UniProt, tripping the tool's trust gate, so it is used here only for the one thing a corpus-level synthesis is good evidence for: the negative that no catalytic activity or substrate has been characterised. Every mechanistic claim in this review is cited to PMID:39225180 or to the UniProt record instead, d… |
| human | ABI3BP | PMID:40092729 | ABI3BP can inhibit the proliferation, invasion, and epithelial-mesenchymal transition of non-small-cell lung cancer cel… | MEDIUM | Cached title matches and full text is available; human cells, and the direction of effect agrees with PMID:18559958 in thyroid and PMID:31174563 in gallbladder. Marked LOW_QUALITY rather than VERIFIED because it is a small single-laboratory overexpression study in a low-impact venue with no in vivo arm and no receptor-level mechanism; it is used as corroboration of a pattern, not as primary evide… |
| human | ABR | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | Crude NK-cell membrane-fraction proteomics. Its own abstract reports only ~40% of hits as plausible membrane proteins, which is the basis for marking the GO:0016020 row over-annotated. Cited for the caveat, not for the localisation. |
| human | ABRA | GO_REF:0000054 | Gene Ontology annotation based on curation of intracellular localizations of expressed fusion proteins in living cells | LOW | The LIFEdb GFP-fusion survey, and the sole source of the plasma-membrane call. Its own title states the method: localisation of expressed fusion proteins. For a muscle-restricted protein assayed in a non-muscle line at non-native levels this is a weak basis for a located_in annotation, and it is recorded as IDA, which reads as far stronger than it is. |
| human | ACAA1 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | Abstract-only in cache; high-throughput membrane-proteomics co-detection producing a spurious membrane localization for a soluble matrix enzyme. |
| human | ACAA1 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | Full text available; systematic Y2H screen with no functional context for the ACAA1 interactions, so the derived bare protein-binding annotations are uninformative/over-annotated. |
| human | ACAP2 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | Cached title matches. Correctly cited, but a bulk membrane-fraction proteome of one NK-like cell line is weak support for a localisation claim, and the study's own caveats undercut it for a peripheral membrane protein. Basis for the GO:0016020 HDA row, which is marked over-annotated on those grounds rather than on any doubt that the peptides were detected. |
| human | ACAP2 | file:human/ACAP2/ACAP2-deep-research-affinage.md | Affinage mechanistic annotation for ACAP2 (human) | MEDIUM | Machine-fetched Affinage record; self_evaluation_pairwise win, faith 100%. Its value here was the reference list rather than the prose. It surfaced PMID:22045739, PMID:22344257, PMID:24600047, PMID:25694427, PMID:25853217, PMID:16806385 and PMID:40251363, four of which materially shaped this review, and none of which is cited by any GOA row. The prose is unreliable at residue resolution, which is… |
| human | ACRBP | PMID:21630459 | Proteomic characterization of the human sperm nucleus. | LOW | Correctly cited, and the mass-spectrometric identification is not disputed, but the paper is a catalogue rather than a localisation study and cannot support a nucleoplasmic assignment for a signal-peptide-bearing secretory protein whose organelle is physically apposed to the sperm nucleus. Basis for MARK_AS_OVER_ANNOTATED on GO:0005634, not for REMOVE. |
| human | ACRV1 | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | Correctly cited and a legitimate large-scale resource, but methodologically weak as support for individual gene annotations. IntAct shows all five ACRV1 partners come from this one publication, each recorded three times as three yeast two-hybrid variants (two hybrid array, two hybrid pooling, validated two hybrid) with both partners over-expressed and MI-score 0.56, so the apparent NbExp=3 is one… |
| human | ACSL4 | PMID:36899378 | Specificity protein 1-mediated ACSL4 transcription promoted the osteoarthritis progression through suppressing the ferr… | MEDIUM | Primary PubMed identity/title/DOI checked against https://pubmed.ncbi.nlm.nih.gov/36899378/ in the source15 identity audit; exact normal-fetch bytes inspected. Extracted article body; no complete supplementary-material appraisal is claimed. Full Methods/Results use human CHON-001 cells, IL-1beta, ACSL4 knockdown and SP1/promoter-reporter perturbations, with no in vivo experiment. Results describe… |
| human | ACSL4 | PMID:40932861 | ACSL4 at the helm of the lipid peroxidation ship: a deep-sea exploration towards ferroptosis. | MEDIUM | Primary PubMed identity/title/DOI checked against https://pubmed.ncbi.nlm.nih.gov/40932861/ in the source15 identity audit; exact normal-fetch bytes inspected. Extracted review body; no complete supplementary-material appraisal is claimed. Secondary ACSL-family/lipid-peroxidation overview. Organelle and cancer claims aggregate underlying sources; no newly measured ACSL4 localization or enzymology… |
| human | ACTL7B | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | The citation is correct and the study is a legitimate large-scale resource, but for ACTL7B specifically it yields four unreplicated Y2H candidate interactions from a screen whose biological context (neurodegeneration, brain tissue) does not overlap the gene's expression domain. Flagged LOW_QUALITY as support for ACTL7B molecular function, not as a criticism of the paper's own claims. |
| human | ACTL7B | PMID:38464253 | Novel Nuclear Roles for Testis-Specific ACTL7A and ACTL7B Supported by In Vivo Characterizations and AI Facilitated In… | MEDIUM | Correctly cited and the PMID resolves to the intended record, but PubMed types it as a Preprint (bioRxiv, PMCID PMC10925299) with no peer-reviewed version found by PubMed search as of this review. Its in vivo localisation and HDAC observations are used only as corroboration of the nucleus annotation; its INO80 and SWI/SNF interaction claim is in silico docking and is not used to support any GO te… |
| human | ACTL7B | file:human/ACTL7B/ACTL7B-deep-research-affinage.md | Affinage mechanistic annotation for ACTL7B (human) | MEDIUM | The record carries no self_evaluation_pairwise score; its narrative correctly surfaced PMID:36617158 and PMID:38464253, but the record missed PMID:37800308 entirely - the Development 2023 Actl7b-null paper that supplies the only informative molecular partner - which was found instead by a direct PubMed search of the gene symbol. Its own GO grounding (GO:0008092 cytoskeletal protein binding) is un… |
| human | ACTL8 | PMID:32125225 | Actin‑like protein 8 executes a promoting function in the malignant progression of endometrial cancer: identification o… | NONE | Retraction confirmed from the cached PubMed metadata, which records the publication type and the retraction notice citation. The affinage deep-research record cites this paper as a dated finding without flagging the retraction, and it is the sole source of that record's p21, E-cadherin and EMT-marker claims. Flagged is_invalid so it is not re-imported; the proliferation and migration conclusions… |
| human | ACTL8 | file:human/ACTL8/ACTL8-deep-research-affinage.md | Affinage mechanistic annotation for ACTL8 (human) | MEDIUM | Treated as a preliminary lead source rather than evidence, per its own header. It passed its internal gates, but it cites the retracted PMID:32125225 as a dated finding without flagging the retraction, and that paper is the sole source of its p21, E-cadherin and EMT-marker claims - so its narrative is partly built on withdrawn work. Its remaining claims were independently verified against the cit… |
| human | ACTMAP | file:human/ACTMAP/ACTMAP-deep-research-affinage.md | Affinage mechanistic annotation for ACTMAP (human) | MEDIUM | gates_passed True and faith 100%, and the narrative is substantially correct including the 2026 covalent-probe result, which is a genuine lead. Marked LOW_QUALITY for one citation defect: its single 2026 finding is attributed to "PMID:42159598, PMID:41757055", which are the JACS paper and its own bioRxiv preprint presented as two sources. Its GO grounding (GO:0140096, GO:0016787, GO:0005829) was… |
| human | ACTR10 | PMID:39697424 | ACTR10 Overexpression Facilitates the Progression and Tyrosine Kinase Inhibitor Resistance in Hepatocellular Carcinoma. | LOW | Entirely in-silico reanalysis of public expression, DepMap and co-expression data; no wet-lab experiment on ACTR10 protein function. Cited here only because it documents the downstream cost of the Reactome granule annotations: it restates them as ACTR10 biology ("suggested to be present in the cytosol, extracellular region and secretory granules"). |
| human | ACTR1B | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | Correctly cited, but methodologically weak for this purpose. All three IntAct records for the ACTR1B-HTT pair -- two hybrid array, two hybrid pooling, validated two hybrid -- originate from this single publication, so UniProt's NbExp=3 does not represent three independent studies. Basis for marking that row over-annotated. |
| human | ACTR5 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | The HuRI two-hybrid map. Its single ACTR5 row pairs a nuclear chromatin-remodelling subunit with periplakin, and IntAct logs that one screen under three method names, which is where UniProt's NbExp=3 comes from. Judged an unreplicated screen hit. |
| human | ACTR5 | PMID:40386946 | Trio Whole Exome Sequencing in Chinese Childhood-Onset Lupus Reveals Novel Candidate Genes. | LOW | Cited by the affinage record. Its only ACTR5 datum is that one de novo variant enhanced type I interferon signalling in an IFN-beta luciferase reporter, in a candidate-gene screen of 50 trios. Too preliminary to support any annotation; recorded so that the affinage finding is accounted for rather than silently dropped. |
| human | ACTR5 | file:human/ACTR5/ACTR5-deep-research-affinage.md | Affinage mechanistic annotation for ACTR5 (human) | LOW | The machine-fetched affinage record (self_evaluation_pairwise: win, 2 citations, 3 dated findings). Its narrative is faithful to PMID:36563143 for the HCC/CDKN2A phenotype, which is why that one sentence is cited here, but its corpus coverage is narrow: it found 2 papers where UniProt lists 12 primary references, and it missed the entire structural literature on this gene - both human INO80-nucle… |
| human | ACTR8 | file:human/ACTR8/ACTR8-deep-research-affinage.md | Affinage mechanistic annotation for ACTR8 (human) | NONE | Recorded so the provider gap is machine-readable, and cited by nothing in this review. The record is empty - n_discoveries 0, citation_count 0, no findings table, no citations, and no self_evaluation_pairwise score - so the trust gates were clear only because there was nothing to gate. That is a provider gap, not a literature gap: the UniProt RN list alone carries 13 references, including a 2.6 A… |
| human | ACTRT2 | PMID:40811009 | ACTRT2 deficiency increases spermatogonia vulnerability to ferroptosis. | HIGH | PubMed-verified, no retraction, and the only knockout study of this gene, so its relevance is high. Marked LOW_QUALITY on methodological grounds rather than citation grounds: the in vivo busulfan comparison uses heterozygotes rather than nulls, the mechanism rests on expression changes in four regulators rather than a demonstrated activity, and the cell type is several stages earlier than the spe… |
| human | ACTRT2 | file:human/ACTRT2/ACTRT2-deep-research-affinage.md | Affinage mechanistic annotation for ACTRT2 (human) | MEDIUM | Machine-fetched precomputed research, gates_passed true with 6 citations, used as a lead rather than as evidence. Two concrete defects. One of its six citations, PMID:bio_10.1101_2025.03.27.645694, is a bioRxiv DOI in a PMID-shaped field rather than a PubMed identifier, and nothing in this review rests on it. More importantly it omits PMID:35793634, the calicin paper that is the actual source of… |
| human | ACTRT3 | file:human/ACTRT3/ACTRT3-deep-research-affinage.md | Affinage mechanistic annotation for ACTRT3 (human) | MEDIUM | Machine-generated, gates_passed true, and its narrative is broadly right. Two defects were found. Its four citations include one bioRxiv DOI in a PMID-shaped field, and comparing that entry with PMID:41668650 shows the two are the same study - the same theca localisation, the same cap-phase acrosome defect, the same GM130 and TGN46 result and the same co-immunoprecipitation set - so the record pr… |
| human | ADAMTSL3 | PMID:32266537 | Genome-wide CRISPR knockout screens identify ADAMTSL3 and PTEN genes as suppressors of HCC proliferation and metastasis… | LOW | The ADAMTSL3 result rests on sgRNAs appearing twice in one genome-wide screen in one cell line, with no mechanistic follow-up and no independent replication. Cited so the claim is on record and explicitly not annotated. |
| human | ADAMTSL3 | file:human/ADAMTSL3/ADAMTSL3-deep-research-affinage.md | Affinage mechanistic annotation for ADAMTSL3 (human) | LOW | gates_passed: False. The tripped gate is a suspected symbol collision raised because the narrative opens on C. elegans; on inspection that is a false positive, since the narrative correctly identifies human ADAMTSL3/punctin-2 and is discussing the worm orthologue Ce-Punctin/madd-4. Every claim taken from the record was nonetheless re-verified against the cited PMIDs, and four defects were found.… |
| human | ADAMTSL5 | PMID:23962539 | Elastic fibres in health and disease. | LOW | Not miscited - the review is a legitimate source for elastic fibre composition, and TAS is the honest evidence code for it. But it is weak support for a gene-specific localisation: full text is unavailable, the abstract never names ADAMTSL5, and the same reference underpins 41 other elastic-fiber assignments. Flagged LOW_QUALITY on those grounds rather than MISCITED. Per campaign policy I did not… |
| human | ADAMTSL5 | file:genes/human/ADAMTSL5/ADAMTSL5-deep-research-affinage.md | Affinage mechanistic annotation for ADAMTSL5 (human) | LOW | gates_passed is False. The tripped gate is specifically the self-evaluation pairwise tie, not a faithfulness failure (faith_pct is 100.0), and no citation is a bioRxiv PMID:bio_* identifier. Marked LOW_QUALITY per campaign policy for a failed gate. Every claim used from it was re-verified against the cited PMID directly and is quoted from the PMID rather than from the provider prose; its GO groun… |
| human | ADCK2 | file:human/ADCK2/ADCK2-deep-research-affinage.md | Affinage mechanistic annotation for ADCK2 (human) | MEDIUM | Machine-generated provider record, gates_passed True. Used only to widen the literature search: it surfaced the three cancer papers and the 2024 mouse study, all of which were then read and judged on their own. No provider sentence is quoted as evidence for any mechanistic claim in this review, per the campaign rule, and no annotation verdict rests on it. Marked LOW_QUALITY rather than MISCITED b… |
| human | ADCK5 | PMID:32277958 | aarF domain containing kinase 5 gene promotes invasion and migration of lung cancer cells through ADCK5-SOX9-PTTG1 path… | MEDIUM | Not miscited and not disputed in the literature, but methodologically thin for the weight placed on it elsewhere: the abstract hedges every step ("might regulate", "might be required"), reports no in vitro kinase assay, and full text is unavailable. The affinage record restates it as an established fact that ADCK5 phosphorylates SOX9 at serine 181, which the abstract does not support. No GOA row… |
| human | ADGRA1 | file:human/ADGRA1/ADGRA1-deep-research-affinage.md | Affinage mechanistic annotation for ADGRA1 (human) | LOW | Provider record, gates_passed: True, no bioRxiv-DOI pseudo-PMIDs in its citation list, and it surfaced the two most load-bearing references here (PMID:28935861, PMID:41961591). Two defects mean nothing in this review rests on it. First, it lists PMID:40766348 (2025 bioRxiv) and PMID:41961591 (2026 Cell Reports) as two separate dated findings; PubMed records UpdateIn: 41961591 on the former, so th… |
| human | ADIRF | file:human/ADIRF/ADIRF-deep-research-affinage.md | Affinage mechanistic annotation for ADIRF (human) | LOW | Recorded so the provider's performance on this gene is measurable, and deliberately NOT cited as supporting_text for any annotation. Precision is fine: gates_passed is True, both citations (PMID:19444912, PMID:23467766) are real numeric PubMed ids rather than bioRxiv DOIs in a PMID-shaped field, both resolve to papers genuinely about this gene under its former names, and neither is retracted. Rec… |
| human | ADPRH | file:human/ADPRH/ADPRH-deep-research-affinage.md | Affinage mechanistic annotation for ADPRH (human) | LOW | The record's trust gate failed - the frontmatter reads gates_passed: False, with the explanation that Affinage's own head-to-head self-evaluation scored the record pairwise = tie rather than a win. Re-checking every claim against the cited PMIDs found two genuine defects, both name/paralog collisions - PMID:12464675 is about the LDL receptor adaptor LDLRAP1 rather than ADP-ribosylhydrolase 1, and… |
| human | AEBP2 | file:human/AEBP2/AEBP2-deep-research-affinage.md | Affinage mechanistic annotation for AEBP2 (human) | MEDIUM | gates_passed is True and faith_pct 100, and recall was unusually good on this gene: the record surfaced PMID:41168462, which is the reference that reframes the annotation set. Marked LOW_QUALITY for two specific defects rather than as a blanket judgement. First, two of its 24 citations are bioRxiv DOIs in a PMID-shaped field (PMID:bio_10.1101_2025.11.09.687442, PMID:bio_10.1101_2025.10.14.682307)… |
| human | AFF1 | file:human/AFF1/AFF1-deep-research-affinage.md | Affinage mechanistic annotation for AFF1 (human) | LOW | Adjudicated and deliberately not used as evidence anywhere in this review; recorded here so the judgement lives in the repository. Its gates_passed: True and faith_pct: 100.0 concern PRECISION - the 23 citations are all well-formed numeric PMIDs with no preprint ids in PMID-shaped fields. Recall was measured, not estimated: of the 7 PMIDs that AFF1's GOA rows actually cite, affinage returned 0, i… |
| human | AFF3 | file:human/AFF3/AFF3-deep-research-affinage.md | Affinage mechanistic annotation for AFF3 (human) | LOW | Adjudicated and deliberately not quoted as evidence anywhere in this review; recorded here so the judgement lives in the repository. Its gates_passed True and faith_pct 100.0 concern PRECISION only - all 15 citations are well-formed numeric PMIDs with no preprint ids in PMID-shaped fields, and none of them is retracted or carries an expression of concern. They say nothing about recall, and recall… |
| human | AFF4 | PMID:31466050 | Genipin attenuates mitochondrial-dependent apoptosis, endoplasmic reticulum stress, and inflammation via the PI3K/AKT p… | LOW | Not an AFF4 paper. It is the sole evidence, by IEP, behind rat Aff4's ER-stress annotation, which Ensembl Compara then projected onto human AFF4. A drug-intervention study in a rat lung-injury model; the projection test shows it produces exactly one annotation on one entity in the whole of GOA. Marked LOW_QUALITY as support for the propagated term, not as a judgement on the paper's own subject. |
| human | AGGF1 | PMID:39251607 | Systematic identification of post-transcriptional regulatory modules. | LOW | The study is sound; the derived annotation is not. IntAct records AGGF1 as experimentalRoleA bait in all 635 proximity records, and a BirA fusion biotinylates itself, so the self-proximity datum carries no information about homodimerisation. The paper also carries an unflagged Author Correction (PMID:39468017); reading it shows the correction adds a missing author affiliation and affects no data. |
| human | AGT | PMID:17159080 | Cross-talk between angiotensin II receptor types 1 and 2: potential role in vascular remodeling in humans. | LOW | PubMed types this Comment / Editorial / Review; it is a two-page commentary on another paper, and the cached record contains no abstract text at all. It is nonetheless the sole source of eleven GOA rows on this gene, three of them coded TAS and eight NAS. Several of the terms it supports are correct RAS physiology and are accepted here on the biology rather than on this citation; the generic ones… |
| human | AGT | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | A competent large-scale study whose output is unsuitable for functional annotation of this gene. It yields ten bare protein-binding rows on AGT, a protein unrelated to neurodegeneration; the paper itself calls the results candidate interactions. None of the ten partners shares AGT's secreted compartment (0/10, computed in AGT-bioinformatics), and eight of them carry a curated intracellular locati… |
| human | AGT | file:human/AGT/AGT-deep-research-affinage.md | Affinage mechanistic annotation for AGT (human) | LOW | The script's trust gate tripped (self_evaluation_pairwise: tie, not win), and the record fails independently of that. It confuses AGT with AGXT: the narrative asserts that AGT has peroxisomal alanine:glyoxylate aminotransferase activity, citing a primary hyperoxaluria type 1 study in AgxtQ84-/- rats. 'AGT' is also the common abbreviation for that enzyme, whose HGNC symbol is AGXT (P21549) - a dif… |
| human | AHI1 | file:human/AHI1/AHI1-deep-research-affinage.md | Affinage mechanistic annotation for AHI1 (human) | MEDIUM | Genuine existing affinage output is preserved unchanged and used as a lead, not primary evidence. Its incorrectly PMID-shaped bioRxiv identifier resolves to the real CEP290 preprint PMID:39896654 (DOI 10.1101/2025.01.20.633784), identified through the primary PubMed/PMC pages and now read in the normal cached Methods and Results. The AHI1 localization experiment uses CEP290-mutant mouse retina; s… |
| human | AIG1 | file:human/AIG1/AIG1-deep-research-affinage.md | Affinage mechanistic annotation for AIG1 (human) | MEDIUM | Trust gates passed and all seven citations are numeric PMIDs that resolve to papers genuinely about AIG1, with no bioRxiv identifiers in a PMID-shaped field and no retraction or expression of concern among them. Recall was good here: it surfaced four papers absent from GOA (PMID:27040980, PMID:32152231, PMID:38816388, PMID:40303337), two of which changed this review. Marked LOW_QUALITY solely for… |
| human | AIMP2 | file:human/AIMP2/AIMP2-deep-research-cyberian.md | Cyberian deep research on AIMP2 | MEDIUM | Genuine preserved provider artifact, used as a discovery lead rather than independent experimental evidence. It contains overgeneralized domain/stoichiometry and signaling claims that require primary-source correction. Fresh provider execution failed before contacting a provider; no replacement report was authored. |
| human | AIMP2 | file:human/AIMP2/AIMP2-deep-research-falcon.md | Falcon deep research on AIMP2 | MEDIUM | Preserved generated literature lead, not independent experimental evidence. Its summary presents direct JNK phosphorylation of AIMP2 more categorically than the unresolved kinase assignment in primary PMID:18695251. Structural and signaling claims used in the review are checked against primary sources; the report bytes are unchanged. |
| human | AIMP2 | file:human/AIMP2/AIMP2-deep-research-openai.md | OpenAI deep research on AIMP2 | MEDIUM | Preserved generated literature lead, not independent experimental evidence. Its assertion that JNK phosphorylates AIMP2 does not retain the primary PMID:18695251 uncertainty about the direct kinase. Only independently verified primary findings support annotation decisions; registering the artifact records provenance without endorsing all claims. |
| human | AIMP2 | file:human/AIMP2/AIMP2-deep-research-perplexity.md | Perplexity deep research on AIMP2 | MEDIUM | Preserved generated literature lead with specific source-scope errors. It calls AIMP2 and AIMP3 aminoacyl-tRNA synthetases in a GST-domain summary and overstates JNK phosphorylation as necessary and sufficient. AIMP2 is the nonenzymatic scaffold in the reviewed MSC mechanism; primary evidence and its limits supersede these generated statements. |
| human | AKAP12 | PMID:21423176 | Analysis of the myosin-II-responsive focal adhesion proteome reveals a role for β-Pix in negative regulation of focal a… | LOW | Correctly cited and a sound paper, but as a GO source it supplies GO:0005925 to 285 annotations, all HDA and all UniProt-assigned, including RPLP0, RPLP1 and RPLP2 - cytoplasmic large-ribosomal-subunit proteins that are not focal adhesion components. Flagged LOW_QUALITY as an annotation source, not as science: a proteomic inventory of isolated adhesions is not per-protein validation. |
| human | AKAP12 | PMID:27683220 | AKAP12 mediates PKA-induced phosphorylation of ATR to enhance nucleotide excision repair. | NONE | RETRACTED. The retraction notice (PMID:33091128) states the authors requested retraction after a University of Kentucky investigation determined the paper contains fabricated and/or falsified data, including substituted blank panels in negative controls. The affinage record cites this paper as a high-confidence finding with no flag, and it is the sole source of that record's DNA-repair claims and… |
| human | AP1S1 | PMID:16548883 | Transcriptomic and proteomic analyses of rhabdomyosarcoma cells reveal differential cellular gene expression in respons… | LOW | Correctly identified, but a single-cell-line infection microarray in which 191 transcripts moved more than twofold and AP1S1 is not named in the abstract. Supports 'the transcript responds' and not 'the protein participates'. Projects to 20 distinct gene products in QuickGO. |
| human | AP1S1 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | Bulk NK-cell membrane proteomics. Correctly cited but maximally generic: QuickGO returns 1142 annotations over 1142 distinct gene products for this reference, one per protein identified. The authors themselves note many recovered proteins are only transiently membrane-associated, which is sigma1A's case. |
| human | AP1S1 | PMID:9714600 | The AP-3 complex: a coat of many colours. | LOW | Donor-side only. QuickGO records this as the TAS behind mouse Ap1s1's own GO:0016192 row. Checking the cached title shows it is a review of AP-3, not of AP-1, which makes it thin support for a vesicle-transport statement about an AP-1 subunit. Recorded honestly in the IBA block: the mouse donor's value is its orthology to the target, not this citation. |
| human | AP1S1 | file:human/AP1S1/AP1S1-deep-research-affinage.md | Affinage mechanistic annotation for AP1S1 (human) | MEDIUM | Machine-generated deep-research record whose own head-to-head self-evaluation scored pairwise = tie rather than win, tripping the trust gate; marked LOW_QUALITY on that basis and every claim taken from it was re-verified against its PMID. It does describe the right protein and found the disease and cell-biology literature well. Two errors: it presents PMID:2040623 as though the cDNA were human wh… |
| human | AP1S2 | file:human/AP1S2/AP1S2-deep-research-affinage.md | Affinage mechanistic annotation for AP1S2 (human) | MEDIUM | Machine-generated deep-research record whose own head-to-head self-evaluation scored pairwise tie rather than win against the curated UniProt reference, tripping the trust gate. It describes the correct protein and its narrative matched the primary papers where I checked it, and its ten citations are all numeric PMIDs with no preprints. Marked LOW_QUALITY per the tripped gate: every claim taken f… |
| human | AP1S3 | PMID:34367317 | Adaptor Protein Complex 1 Sigma 3 Is Highly Expressed in Glioma and Could Enhance Its Progression. | LOW | A TCGA-mining plus knockdown study reporting AP1S3 overexpression in glioblastoma and reduced proliferation on silencing. Correctly identified and about this gene, but methodologically thin - the mechanistic part is database enrichment analysis - and far downstream of a housekeeping trafficking adaptor. No annotation proposed. |
| human | AP1S3 | PMID:40555003 | AP1S3 affects lipid metabolism in breast cancer cells by regulating the PI3K/AKT/mTOR pathway. | LOW | A 2025 breast-cancer study linking AP1S3 knockdown to reduced proliferation and lipid droplet accumulation via PI3K/AKT/mTOR. Correctly cited, but a cell-line association study several steps removed from the protein's biochemistry. Abstract-only in the cache. No annotation proposed. |
| human | AP3D1 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | ARBA machine-learned rules. Of the six uses, four fire on a single AP3D1-specific CATH FunFam (3.30.450.50:FF:000001) and give reasonable terms; one (ARBA00092758, GO:0010496 intercellular transport) gives a term with the wrong topology from the same evidence; and one (ARBA00028739, GO:0032502) cannot be reproduced from any of the rule's 893 published condition sets. |
| human | AP3D1 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | NK-cell membrane proteome. Correctly cited as an HDA, but a full QuickGO reference projection returns 1142 distinct gene products, and the paper itself notes that much of its catalogue is only transiently membrane-associated. Not informative about AP3D1. |
| human | AP3M1 | file:human/AP3M1/AP3M1-deep-research-affinage.md | Affinage mechanistic annotation for AP3M1 (human) | LOW | The script's trust gate tripped (self_evaluation_pairwise: loss, faith_pct 50.0), and the record is thin rather than wrong: one citation, PMID:29032074, reporting that AP3B1 depletion lowers AP3M1 abundance. That observation is real and correctly cited, but it is complex stoichiometry rather than function, and the record's own claim that no further mechanistic detail has been characterised is fal… |
| human | AP3M2 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | ARBA00026971 was fetched from rest.uniprot.org and its 2,388 condition sets were filtered against the seven InterPro signatures InterPro reports for P53677 and against PTHR10529. The two sets that match a signature also require taxon Saccharomyces, or the sigma-subunit families IPR016635/IPR027156, so none can fire on this protein. The conferred term happens to be true. |
| human | AP3M2 | PMID:39488930 | AP3M2: A key regulator from the nervous system modulates autophagy in colorectal cancer. | LOW | Abstract-only. Knockdown of AP3M2 in three colorectal-cancer lines reduces viability and correlates with autophagy and ROS genes; no mechanism and no rescue. Cited only as evidence that human AP3M2 is functionally relevant outside neurons. |
| human | AP3S1 | PMID:9792713 | Interaction of insulin receptor substrate-1 with the sigma3A subunit of the adaptor protein complex-3 in cultured adipo… | LOW | The sole source of the insulin-receptor-signalling row. Correctly cited and the interaction is real as reported, but the paper itself frames the functional claim as a hypothesis and nothing has confirmed it in 28 years: 5 PMC-indexed citing papers, and UniProt's SUBUNIT line for Q92572 does not list IRS1. Graded LOW_QUALITY for the annotation it supports, not for the biochemistry it reports. Cach… |
| human | AP4M1 | file:human/AP4M1/AP4M1-deep-research-affinage.md | Affinage mechanistic annotation for AP4M1 (human) | LOW | Trust gate tripped - the record's own head-to-head self-evaluation scored pairwise = tie against the curated UniProt reference, and the narrative bears that out. Its six citations are the cloning paper plus five clinical/genetic/iPSC-line papers, and it states that the cargo-recognition mechanism has not been further characterized in the available corpus, which is the opposite of the truth for th… |
| human | AP4S1 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | The ARBA machine-learning reference behind the GO:0005737 and GO:0012505 rows. Marked LOW_QUALITY on the basis of a specific check rather than a prior about machine learning: fetching both rules from https://rest.uniprot.org/arba/ and intersecting all 2388 and 533 condition sets against AP4S1's three real InterPro matches shows that neither annotation can be reproduced from its own rule's publish… |
| human | AP5M1 | PMID:18395520 | A novel protein, MUDENG, induces cell death in cytotoxic T cells. | MEDIUM | Cited by UniProt for the cell-death function. The evidence is ectopic overexpression in Jurkat and HeLa cells arising from a randomised ribozyme screen, with no mechanism and no link to AP-5. Abstract-only in cache. Correctly cited, methodologically weak for a function claim. |
| human | AP5M1 | PMID:31427081 | BAX is an essential key mediator of AP5M1-induced apoptosis in cervical carcinoma cells. | LOW | Abstract-only. Overexpression of AP5M1 is pro-apoptotic and BAX-dependent in cervical carcinoma lines. The same overexpression limitation as PMID:18395520, and the opposite sign to PMID:27136675 and PMID:41915273. |
| human | AP5M1 | file:human/AP5M1/AP5M1-deep-research-affinage.md | Affinage mechanistic annotation for AP5M1 (human) | LOW | Trust gates clear and self_evaluation_pairwise win, and each of its five cited findings is correctly summarised, but all five are MUDENG cell-death papers and the record returned nothing at all on AP-5. Its closing scope statement, that no adaptor-complex trafficking function has been characterised for AP5M1, is false: PMID:22022230 is the source of nine GOA rows and of the UniProt function line.… |
| human | APLN | file:human/APLN/APLN-deep-research-affinage.md | Affinage mechanistic annotation for APLN (human) | MEDIUM | Trust gates cleared (self_evaluation_pairwise: win, faith_pct 100) and the record describes the right protein - no symbol collision. Marked LOW_QUALITY not because it is wrong but because it is a lead list whose bibliography barely intersects the record under review: of the seven PMIDs cited by GOA rows on this gene it cites exactly one, and it omits the apelin knockouts, both donor-side papers b… |
| human | APLP1 | PMID:41270468 | Binding domains of alpha-synuclein receptors with monomeric/oligomeric alpha-synuclein: Implications for Parkinson's di… | MEDIUM | PubMed-verified (Biomed Pharmacother 2025; review article with AlphaFold3 predictions). Useful for framing the three-receptor dispute. Structure predictions alone cannot support a GO molecular-function annotation, and no annotation is proposed on its basis. |
| human | APOO | PMID:42647630 | The molecular basis of mitochondrial crista formation by the MIC10 complex. | MEDIUM | The current structural model of the MIC10 subcomplex, and the only source making any cardiolipin-recruitment claim about MIC26 itself. Graded LOW_QUALITY not because the work is poor - it is careful and explicit about its limits - but because the evidence class is AlphaFold 3 models plus coarse-grained and all-atom molecular dynamics, with no experimental measurement. Cited here to define a knowl… |
| human | APOO | file:human/APOO/APOO-deep-research-affinage.md | Affinage mechanistic annotation for APOO (human) | MEDIUM | Machine-generated deep research (Affinage, LLM-authored). Trust gates clear and self_evaluation_pairwise 'win'; the narrative is about the right protein, with no symbol collision, and it correctly leads with the 2023 refutation of the secreted isoform, which is the most important fact for this gene's annotations. Graded LOW_QUALITY for two defects: it cites PMID:26217776 twice for MIC26 content w… |
| human | APOOL | file:human/APOOL/APOOL-deep-research-affinage.md | Affinage mechanistic annotation for APOOL (human) | MEDIUM | LLM-generated deep research. Trust gates cleared (self_evaluation_pairwise win, faith 100%) and the record describes the correct protein with no symbol collision, which is why it is listed at all; graded LOW_QUALITY because it is machine-authored and unreviewed, so nothing in it is used as supporting text. Its dated-findings table checked out against the primary papers wherever tested. It missed… |
| human | ARC | PMID:21834987 | Identification and characterization of a set of conserved and new regulators of cytoskeletal organization, cell morphol… | MEDIUM | The source of both human IDAs in this file (GO:0005737, GO:0005886) and, via UniProt, of the cytoskeleton SubCell keyword. Full text verified. Two problems, neither fatal: the localisation data are GFP-ARC overexpressed in PC3 prostate carcinoma cells, which do not express ARC, and the paper carries a 2024 author correction stating that the siARC panels of Fig. 6 duplicated the siZRANB1 panels. T… |
| human | ARGLU1 | PMID:35082911 | miR-335-5p Inhibits Progression of Uterine Leiomyoma by Targeting ARGLU1 | NONE | RETRACTED. PubMed lists this record with publication type Retracted Publication and a RetractionIn pointer to PMID:37503407. The affinage deep-research record for this gene cites it as a dated finding with no retraction flag, which is a provider defect worth recording. Nothing in this review rests on it, and it is listed here solely so that the retraction is on the record for the next reviewer. |
| human | ARHGAP23 | PMID:38022849 | Pan‑cancer analysis identified ARHGAP23 as a potential biomarker for pancreatic adenocarcinoma. | NONE | Listed so that its absence from the evidence base is deliberate rather than accidental: it is one of only two papers with ARHGAP23 in the title, so an automated pass is likely to weight it heavily. It states that ARHGAP23 is known to activate RHO-GTPase, which is the opposite of what a GTPase-activating protein does, and it contains no primary data. |
| human | ARHGAP23 | file:human/ARHGAP23/ARHGAP23-deep-research-affinage.md | Affinage mechanistic annotation for ARHGAP23 (human) | MEDIUM | Cited both for what it got right and for what it missed. Every citation it returned resolves and is correctly used, so nothing here is a citation error; the defect is coverage, and the resulting explicit assertion that no experimental characterization exists is contradicted by three papers found by an independent search of Europe PMC and NCBI gene2pubmed. Flagged LOW_QUALITY rather than MISCITED… |
| human | ARHGAP4 | PMID:26707211 | ARHGAP4 mutated in a Chinese intellectually challenged family. | LOW | A single-family exome study with no functional assay, whose candidate was selected by database filtering and SIFT. T491 lies outside the Rho-GAP domain (507-695). Recorded because it is one of only two ARHGAP4 variant reports in humans, not because it establishes a function. It produced no GO annotation. |
| human | ARHGEF16 | file:human/ARHGEF16/ARHGEF16-deep-research-affinage.md | Affinage mechanistic annotation for ARHGEF16 (human) | MEDIUM | Machine-fetched provider record; trust gates clear, self_evaluation_pairwise win, faith_pct 100.0, and all six of its citations verified here as genuinely about this protein - so nothing in it is miscited, and MISCITED would say something the rest of this note disclaims. LOW_QUALITY instead, for a methodological reason: its retrieval is keyed on the symbol ARHGEF16 and so returned none of the sev… |
| human | ARHGEF18 | file:human/ARHGEF18/ARHGEF18-deep-research-affinage.md | Affinage mechanistic annotation for ARHGEF18 (human) | MEDIUM | Blocking trust gates passed (accession Q6ZSZ5, human), but the soft gate tripped. Affinage's own head-to-head self-evaluation scored this record as a tie rather than a win against the curated UniProt reference. Precision was good, since every PMID it returned resolves to a real paper about this gene, and the three identifiers checked most closely (20810787, 23648482, 26483385) match the papers it… |
| human | ARHGEF19 | PMID:34813497 | ARHGEF19 promotes the growth of breast cancer in vitro and in vivo by the MAPK pathway. | MEDIUM | The third independent tumour type showing the same MAPK effect, which is why it is cited, but it is the weakest of the three: a low-visibility journal, no mechanism beyond immunoblot, and no domain mapping. Marked LOW_QUALITY for that reason rather than because the citation is wrong - PubMed-verified, cached as abstract only. It contributes convergence, not evidence. |
| human | ASS1 | PMID:22658674 | Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. | LOW | Single high-throughput mRNA-interactome-capture screen that recovered many metabolic enzymes, including ASS1, as putative RBPs. No orthogonal validation of an RNA-binding function for ASS1; basis for flagging the RNA binding annotation as over-annotated. |
| human | C5orf46 | PMID:35504177 | Preliminary study on the role of the C5orf46 gene in renal cancer. | MEDIUM | Cached full text read; retraction and correction checks clean. Marked LOW_QUALITY on methodological grounds rather than citation grounds - the identifier and title are correct and the paper is honestly presented, its own title beginning 'Preliminary study'. The in vitro part is siRNA knockdown in two lines with proliferation, migration and apoptosis read-outs, no rescue, no mechanism, and no mole… |
| human | C5orf46 | file:human/C5orf46/C5orf46-deep-research-affinage.md | Affinage mechanistic annotation for C5orf46 (human) | MEDIUM | A machine-generated secondary source, assessed rather than trusted. Its value here was real: it surfaced all three of the papers that GOA does not cite, including the sole functional characterisation of the gene, so both proposed terms trace to a lead it provided. Marked LOW_QUALITY on two grounds, neither of which is a fabricated citation. First, the cytotoxicity finding is framed without the di… |
| human | CACNA2D1 | PMID:41740819 | Chaihu-Baishao confers antidepressant effects in CSDS mice by ameliorating hippocampal synaptic plasticity impairment v… | LOW | PubMed-verified (J Ethnopharmacol 2026). The nearest thing found to an out-of-house test of the alpha-2/delta-1-NMDAR model, using Cacna2d1-knockout cells and gabapentin, but it is a herbal-medicine efficacy study with molecular docking, and one author has a long publication history with the originating laboratory. Not treated as independent replication. |
| human | CBLIF | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | PubMed-verified. HuRI high-throughput binary (Y2H) interactome; the CBLIF hits (FFAR2, SLC13A4, SLC22A23, SLC7A1, SLC7A14, TMEM237) are unvalidated screen interactions with no established biological role for a secreted Cbl-binding protein. |
| human | CH25H | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | Source of the two bare "protein binding" IPI annotations (KRTAP12-4, OTX1) via the HuRI systematic Y2H interactome. These are high-throughput binary interactions with no established functional relevance to CH25H's cholesterol-hydroxylase function; flagged over-annotated rather than removed (experimental IPI). |
| human | CRYAA | file:human/CRYAA/CRYAA-hypotheses/function-hypothesis-go-0042026/openscientist.md | OpenScientist hypothesis report for CRYAA protein refolding (GO:0042026) | HIGH | Mechanistically sound and useful -- the holdase-not-foldase reasoning, the no-NBD architectural argument, and the identification of PMID:8093612 and PMID:1438232 as paralog-derived/mixed-oligomer evidence were all checked and hold up. However, the report's annotation-set enumeration is wrong in two ways that undercut its recommendations, so its verdict was not adopted. (1) It reports the GO:00420… |
| human | FANCD2 | PMID:35384245 | Physical and functional interactome atlas of human receptor tyrosine kinases. | LOW | PubMed-verified high-throughput RTK interactome atlas. The reported EGFR-FANCD2 interaction is proteome-scale with no established FANCD2 functional context; basis for removing the generic protein binding annotation. |
| human | FANCD2 | PMID:35512704 | Systematic discovery of mutation-directed neo-protein-protein interactions in cancer. | LOW | PubMed-verified systematic neo-PPI screen. Reported EGFR interaction lacks characterized FANCD2 functional relevance; basis for removing the generic protein binding annotation. |
| human | FANCG | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | Aggregation-prone interactome screen; the many non-FA partners reported for FANCG are likely non-specific and are treated as over-annotation. |
| human | FANCL | PMID:25416956 | A proteome-scale map of the human interactome network. | LOW | High-throughput yeast two-hybrid interactome (HuRI); source of numerous generic 'protein binding' (GO:0005515) annotations that are not linked to FANCL's characterized function. |
| human | FANCL | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | HuRI reference binary interactome (high-throughput Y2H); source of generic 'protein binding' annotations not informative for FANCL's molecular function. |
| human | FANCL | PMID:35512704 | Systematic discovery of mutation-directed neo-protein-protein interactions in cancer. | LOW | Systematic neo-PPI screen (partner SMAD4); a generic 'protein binding' interaction not representing a constitutive functional interaction of wild-type FANCL. |
| human | FBXO47 | PMID:34445249 | The SCF Complex Is Essential to Maintain Genome and Chromosome Stability. | LOW | PubMed-verified (Int J Mol Sci 2021;22(16):8544, PMC8395177); abstract-only in cache (full_text_available: false). This is a general SCF-family review and does not establish FBXO47-specific function. It is the NAS source used by ComplexPortal for both FBXO47 SCF annotations, but it provides only family-level support, not direct evidence for FBXO47 acting as a canonical SCF substrate receptor. |
| human | FLG | PMID:21630459 | Proteomic characterization of the human sperm nucleus. | LOW | Correctly cited - FLG does appear in this sperm nuclear proteome - but a shotgun catalogue of a tissue where FLG is not the relevant biology is weak support for nuclear localisation, and large abundant skin proteins are a standard contamination confound. The nucleus annotation is retained on the strength of PMID:9800950 and PMID:12230510 instead. |
| human | FLG | PMID:28344315 | Proteomic characterization of human multiple myeloma bone marrow extracellular matrix. | NONE | A myeloma bone-marrow ECM proteome. FLG's appearance there is best explained by skin-derived sample contamination; it is an intracellular epidermal protein with no signal peptide and no reported expression in bone marrow. Not a credible basis for the extracellular matrix annotation. |
| human | GALK1 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | High-throughput NK-cell membrane proteome. The resulting "membrane" localization is a proteomic co-purification and is directly contradicted by the curated NOT|located_in membrane IDA (PMID:8908517). |
| human | GCSH | PMID:16189514 | Towards a proteome-scale map of the human protein-protein interaction network. | LOW | Large-scale yeast two-hybrid interactome (CCSB-HI1). Source of a bare protein binding IPI; the interactor is not a GCS/lipoylation partner and does not inform GCSH function. |
| human | GCSH | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | HuRI binary interactome map. Source of multiple bare protein binding IPIs (NMI, MAGEA6/11, MED11, MIS18A, RHBDD2); none are GCS or lipoylation-pathway partners, so they do not inform GCSH molecular function. |
| human | KASH5 | PMID:26842404 | Depletion of the LINC complex disrupts cytoskeleton dynamics and meiotic resumption in mouse oocytes. | MEDIUM | Mouse oocyte siRNA/morpholino knockdown only; spindle and F-actin phenotypes are indirect; likely source of oocyte-specific IBA/IEA terms. |
| human | LRCH2 | PMID:35351988 | Novel genes bearing mutations in rare cases of early-onset ataxia with cerebellar hypoplasia. | MEDIUM | PubMed metadata and full cached article were checked. This is a single-family candidate-gene report supported by conservation and expression analyses, without segregation across multiple families or direct functional validation of p.Lys258Glu. It informs tissue and disease hypotheses, not normal molecular function or definitive gene-disease causality. |
| human | NAT10 | PMID:19946888 | Defining the membrane proteome of NK cells. | LOW | Membrane-fraction proteomics; the resulting GO:0016020 annotation for a soluble nucleolar protein is best read as co-fractionation. |
| human | NDUFA11 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | Genome-wide binary Y2H interactome (HuRI); source of the bare "protein binding" IPI. NDUFA11 is not discussed individually (one PPI among >50,000); the reported partner MEOX2 is not a plausible functional partner for a Complex I membrane subunit. Uninformative for molecular function. |
| human | NDUFA12 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | HuRI high-throughput binary (Y2H) interactome. Source of the three IntAct IPI 'protein binding' annotations (TMED8, MYO15B, SYT4). Full text available but the NDUFA12 interactions appear only in supplementary data, not the main text; these binary hits are not the physiologically relevant Complex I contacts, hence the uninformative MF is treated as over-annotation. |
| human | NDUFA13 | PMID:17500595 | Huntingtin interacting proteins are genetic modifiers of neurodegeneration. | LOW | Large-scale huntingtin-interactor screen; source of a bare HTT protein-binding IPI of uncertain physiological relevance to NDUFA13. Uninformative as a molecular function. |
| human | NDUFA13 | PMID:31617661 | Global Interactome Mapping of Mitochondrial Intermembrane Space Proteases Identifies a Novel Function for HTRA2. | LOW | High-throughput interactome map of IMS proteases; source of a bare HTRA2 protein-binding IPI. Consistent with the HtrA2 association but uninformative as a molecular function. |
| human | NDUFA13 | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | Large-scale interactome/aggregation map; source of a bare HTT protein-binding IPI of uncertain relevance to NDUFA13. Uninformative as a molecular function. |
| human | NDUFA13 | PMID:40205054 | Multimodal cell maps as a foundation for structural and functional genomics. | LOW | Multimodal cell-map / structural-genomics interactome; source of a bare HTRA2 protein-binding IPI. Uninformative as a molecular function. |
| human | NEK1 | file:human/NEK1/NEK1-deep-research-openscientist.md | Focused Primary-Evidence Audit: Human NEK1 (Q96PY6) Uncertain Claims for GO Curation | MEDIUM | Secondary research report with explicit incomplete full-text inspection. Correctly distinguishes ATR priming from direct Thr1989 phosphorylation and flags the RAD54 dispute, but overstates direct NEK1-CFAP410 binding despite the primary authors explicitly leaving directness unproved. Its all-stages localization and old-artifact framing overlook the observed axonemal pool and human localization ev… |
| human | NINJ1 | PMID:26677008 | Ninjurin1 regulates lipopolysaccharide-induced inflammation through direct binding. | MEDIUM | Correctly cited for the GO:0001530 annotation, but the evidence is a pull-down from transfected HEK293T lysates with biotinylated LPS, from one laboratory, never independently reproduced and without structural corroboration in any NINJ1 cryo-EM structure. Basis for demoting both the LPS binding and the TLR4 regulation annotations to non-core. |
| human | NSUN7 | PMID:24384068 | T26248G-transversion mutation in exon7 of the putative methyltransferase Nsun7 gene causes a change in protein folding… | MEDIUM | Abstract only in our cache. Small candidate-gene association study in one Iranian cohort with in silico folding predictions rather than functional work; supports human relevance but should not be treated as establishing a mechanism. |
| human | PCCB | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | Large-scale yeast two-hybrid interactome; the PCCB-ACTN3 hit is not biologically corroborated for a mitochondrial matrix enzyme and is most likely a high-throughput artifact. Supports only a bare protein-binding IPI. |
| human | PDZD7 | PMID:22664934 | Comparison of tear protein levels in breast cancer patients and healthy controls using a de novo proteomic approach. | LOW | Correctly cited (the paper exists and is a tear-fluid proteomic survey), but it is a pooled-sample de novo MALDI-TOF-TOF screen unrelated to PDZD7, which is never discussed in the text. Not an adequate basis for an extracellular-region annotation on an intracellular scaffold. |
| human | PMM2 | PMID:25416956 | A proteome-scale map of the human interactome network. | LOW | High-throughput binary interactome screen; supports only a generic, uninformative 'protein binding' annotation with no demonstrated functional relevance to PMM2. |
| human | PMM2 | PMID:26871637 | Widespread Expansion of Protein Interaction Capabilities by Alternative Splicing. | LOW | High-throughput interactome screen; supports only a generic 'protein binding' annotation, not a core function. |
| human | PMM2 | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | HuRI reference binary interactome; generic 'protein binding' hits with no functional meaning for a cytosolic metabolic enzyme. |
| human | PNMT | PMID:32296183 | A reference map of the human binary protein interactome. | LOW | PubMed-verified systematic interactome paper, but the PNMT binary interactions are high-throughput screen hits with no functional context; basis for a bare protein-binding annotation only. |
| human | QDPR | PMID:34800366 | Quantitative high-confidence human mitochondrial proteome and its dynamics in cellular context. | LOW | High-throughput mitochondrial-proteome mapping in which QDPR is detected; weak, single-source evidence for organellar residence of an abundant cytosolic SDR enzyme lacking mitochondrial targeting features. Treated as over-annotation, not core localization. |
| human | RFT1 | file:human/RFT1/RFT1-deep-research-cyberian.md | Cyberian deep research on RFT1 function | LOW | Machine-generated. Explicitly frames the controversy as resolved in 2024 ("This controversy persisted for over two decades until 2024"), which the 2026 primary literature contradicts. |
| human | RFT1 | file:human/RFT1/RFT1-deep-research-falcon.md | Deep research on RFT1 function (falcon) | MEDIUM | Machine-generated. Useful on topology and general background, but it presents the flippase question as closed by the 2024 reconstitution and does not surface PMID:19494107, PMID:23720757 or PMID:42417535. Do not rely on it for the molecular function call. |
| human | RPN1 | PMID:22658674 | Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. | LOW | Correctly cited high-throughput RBP atlas, but the RNA-binding assignment is a screen-level call without a defined molecular function for RPN1; treated as an over-annotation. |
| human | SLC25A19 | PMID:21630459 | Proteomic characterization of the human sperm nucleus. | LOW | PubMed-verified. High-throughput sperm-nucleus proteome; source of the HDA nucleus annotation, which is not biologically informative for a mitochondrial carrier and is likely a contaminant. |
| human | SLC25A5 | PMID:22658674 | Insights into RNA biology from an atlas of mammalian mRNA-binding proteins. | LOW | High-throughput mRNA-interactome capture; abundant mitochondrial membrane proteins are frequent nonspecific hits, so the "RNA binding" MF is not a credible molecular function for ANT2. |
| human | SLC45A4 | PMID:25164149 | Proton-associated sucrose transport of mammalian solute carrier family 45: an analysis in Saccharomyces cerevisiae. | HIGH | PubMed-verified: Bartolke et al., Biochem J 464(2):193-201, 2014. Correctly cited and not retracted, but methodologically weak as the sole basis of a molecular-function call: sucrose uptake was measured only in Saccharomyces cerevisiae heterologously expressing mouse SLC45A2/3/4, with Km around 5 mM. It is the seed for every sucrose annotation in this family, including the mouse IDAs that the hum… |
| human | SLCO5A1 | PMID:42231149 | Deorphanisation and functional characterisation of OATP5A1 as transport protein for amino acids and vitamins. | HIGH | PubMed-verified: Kohlmann et al., Cell Mol Biol Lett 31(1):75, 2026 (Jun 2), PMC13231527. Full text available and read. The sole source for every substrate claim about OATP5A1, and correctly cited for all of them. Marked LOW_QUALITY not because the work is poorly done - the controls (4 degree Celsius counter-controls, saturation kinetics with six biological replicates, inhibitor panels, thiamine… |
| human | STING1 | PMID:40861013 | Beyond interferons: Non-canonical roles of MITA/STING. | LOW | PubMed-verified: Yang Y et al. (Cell Insight 2025), 'Beyond interferons: Non-canonical roles of MITA/STING', doi 10.1016/j.cellin.2025.100266. A review, and the sole support for four NAS annotations (GO:0006914, GO:0045820, GO:0055088, GO:0090398). Correctly cited, but review prose is weak evidence for process annotations; two of the four (glycolysis, lipid homeostasis) are marked as over-annotat… |
| human | SUN5 | PMID:31144711 | SPAG4L/SPAG4Lβ interacts with Nesprin2 to participate in the meiosis of spermatogenesis. | LOW | Abstract only; claims SPAG4L/Nesprin2 LINC complexes in meiosis from co-IP/IF; contrasts with normal meiosis in Sun5-null mice. |
| human | TCN1 | PMID:32814053 | Interactome Mapping Provides a Network of Neurodegenerative Disease Proteins and Uncovers Widespread Protein Aggregatio… | LOW | Large-scale yeast two-hybrid interactome screen; the single TCN1-JPH3 interaction underlies only an uninformative "protein binding" IPI with no functional follow-up. |
| human | TNFRSF21 | PMID:22761420 | Death receptor 6 induces apoptosis not through type I or type II pathways, but via a unique mitochondria-dependent path… | MEDIUM | PubMed-verified (J Biol Chem 2012). Correctly cited, but the apoptosis it reports depends on ectopic overexpression of a death-domain receptor, which is weak evidence for a physiological role; hence KEEP_AS_NON_CORE for the apoptotic process annotation. |
| human | TRAF3 | PMID:11279055 | A diverse family of proteins containing tumor necrosis factor receptor-associated factor domains. | LOW | PubMed-verified and correctly cited, but the interaction it supports is non-selective: isolated TRAF domains of MUL/TRIM37 and USP7 bound all six TRAF proteins in vitro. Weak basis for the GO:0031625 and GO:0031996 IPI rows, which are consequently kept only as non-core. Abstract-only in cache. |
| human | TRMT6 | PMID:40897821 | m1A methylase TRMT6 promotes neuroblastoma development by demethylating SST mRNA in an m1A/YTHDF2-dependent manner. | LOW | PubMed-verified (Br J Cancer 2025, 133(9), 1354-1364). The identifier is correct and the title really does read this way. The title is internally incoherent, describing TRMT6 as an m1A methylase acting "by demethylating" its target, which a writer enzyme cannot do, while the abstract describes deposition ("TRMT6 mediates m1A modification of SST"). It also reports m1A destabilising its target, the… |
| human | TRMT61A | PMID:40897821 | m1A methylase TRMT6 promotes neuroblastoma development by demethylating SST mRNA in an m1A/YTHDF2-dependent manner. | LOW | PubMed-verified (Br J Cancer 2025, 133(9), 1354-1364). The identifier is correct and the title really does read this way. The title is internally incoherent, describing TRMT6 as an m1A methylase acting "by demethylating" its target, which a writer enzyme cannot do, while the abstract describes deposition ("TRMT6 mediates m1A modification of SST"). It also reports m1A destabilising its target, the… |
| human | USH1C | PMID:15219944 | Expression of AIE-75 PDZ-domain protein induces G2/M cell cycle arrest in human colorectal adenocarcinoma SW480 cells. | LOW | Heterologous overexpression in a cancer cell line lacking endogenous harmonin; not evidence of physiological cell-cycle function, and never corroborated. Basis for the MARK_AS_OVER_ANNOTATED assessment of GO:0000086. |
| rat | Arsb | PMID:15040 | Enzymic changes in the cervix of the rat and hamster during the oestrous cycle and the effect of steroids. | LOW | PubMed-verified. Shows estrogen-induced upregulation of cervical arylsulphatase B activity in ovariectomised rats. Supports a tissue-specific hormonal regulation observation but not a core Arsb function. |
| worm | clk-1 | PMID:11959146 | CLK-1 protein has DNA binding activity specific to O(L) region of mitochondrial DNA. | LOW | Correctly cited but an isolated 2002 in-vitro study of mtDNA (O_L) binding, not widely replicated; basis for the debated transcription cis-regulatory region binding annotation. Treated as a possible moonlighting activity, non-core. |
| worm | dyf-11 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | ARBA machine-learning annotations. The "animal organ development" and "system development" terms are over-general electronic inferences that are not informative for a nematode sensory-cilium protein; "regulation of microtubule cytoskeleton organization" over-generalizes the structural axoneme-assembly role. |
| worm | dyf-6 | GO_REF:0000117 | Electronic Gene Ontology annotations created by ARBA machine learning models | LOW | ARBA machine-learning annotation to the over-general grouping term "plasma membrane bounded cell projection" (GO:0120025), whose informative descendants (cilium, dendrite) are already annotated. Uninformative for this protein. |
| yeast | GTT3 | file:yeast/GTT3/GTT3-deep-research-falcon.md | Falcon deep research report for GTT3 (YEL017W) | HIGH | Genuine Edison/falcon deep-research report (17 citations). Used for its well-supported "no direct evidence" statements about GTT3 and the paralog contrast. Note one factual error: it states GTT3 is "285 amino acids", whereas UniProt P39996 is 337 aa; that length claim was not propagated into this review. |
| yeast | MOH1 | PMID:28173693 | Pro-Apoptotic Role of the Human YPEL5 Gene Identified by Functional Complementation of a Yeast moh1Δ Mutation. | HIGH | PubMed-verified; the only dedicated functional study of MOH1. Abstract-only in cache. It establishes a robust stress-sensitizing phenotype and cross-species complementation, but the "pro-apoptotic/mitochondria-dependent apoptosis" interpretation rests on a single heterologous-complementation viability study in the contested yeast programmed-cell-death paradigm; it does not establish a molecular a… |
| yeast | NVJ3 | PMID:41542480 | Nvj3 regulates Dga1-mediated triacylglycerol synthesis and lipid droplet formation at ER contact sites. | MEDIUM | bioRxiv preprint (2026), NOT peer-reviewed; full text not cached. Metadata and abstract verified via PubMed (PMID:41542480). Used only to frame knowledge gaps around the DAG/Dga1 axis of lipid-droplet biogenesis; NOT used as the basis for any positive core-function claim. No verbatim full-text quote is available. |