Human Genes Annotation Re-Review
Bottom line: each gene review assigns an action (ACCEPT, REMOVE, MODIFY
and so on) to every existing GO annotation, and those actions are only as
good as the reviewer who set them. We re-read the actions in every human
review file then in the repo, 1,323 genes from AAAS to ZSWIM8, and asked of
each one whether we agreed, using the YAML and cached publications without
new literature searches. Agreement was about 99.7%: only three actions were
changed (ABL1 UNDECIDED → REMOVE, ADRM1 REMOVE → UNDECIDED, BRCA2 NOT-row
REMOVE → ACCEPT), and all three edits are in the current YAMLs. We also
adjudicated all 143 NOT annotations and logged four fence cases (AGO3/AGR2,
ATP23, GAPDH, HSPA1B) and two literature questions (ASCL1, ATF3) for a
curator. The sweep finished in July 2026; since then TRA2B, flagged below as
never reviewed, has been reviewed, and the human set has grown to 2,048
review files, so about 725 newer reviews have not had this second pass.
We did this to measure how far the first-pass actions can be trusted before
building summaries and modules on them, and to find any systematic reviewer
errors worth a guideline change.
Purpose
A second-pass, manual re-review of the per-annotation curation actions
(ACCEPT, KEEP_AS_NON_CORE, MARK_AS_OVER_ANNOTATED, REMOVE, MODIFY,
NEW, UNDECIDED) recorded in each genes/human/<GENE>/<GENE>-ai-review.yaml.
For every gene I read the existing_annotations block and ask, for each action:
do I agree? Judgments are made from the information already in the YAML
(term, evidence code, summary, reason, supporting_text) plus the cached
publications/PMID_*.md where available, and my own domain knowledge. No new
literature search is done in this pass; annotations that would need one to
adjudicate are logged below under Needs literature.
- Confident disagreement → I edit the YAML directly (changing
actionand
reason, or fixing a term), and log the change under Edits made. - On the fence / needs more evidence → logged under Fence cases or
Needs literature with the specific question, no edit. - Agree → no change; the gene is recorded in the Progress log as
reviewed.
Scope: 1323 human review files, ~55k individual annotation actions. This is a
marathon; the progress log below is the resumable cursor.
Status: A→Z sweep COMPLETE
Every human gene review file (A through ZSWIM8) has been re-reviewed at the
per-annotation-action level. Bottom line:
- Agreement with the existing reviews is ~99.7%. Across ~1323 genes the sweep
found only 3 action-level edits worth making (all logged under Edits made):
ABL1,ADRM1,BRCA2. Everything else was either already correct or within
legitimate reviewer discretion. - 1 completed systematic audit: all 143 NOT/negated annotations (see below).
- The reviews' value-add is consistent and correct: pseudo-enzyme catches
(lost catalytic residues → REMOVE ancestral activity), wrong-direction/wrong-enzyme
enzyme calls (phosphatase≠kinase, reductase≠its own reductase, TXN vs TXNRD),
name-collision disambiguation, subunit vs holoenzyme distinctions, paralog/substrate
specificity, pathway-step precision (catalyzes-step vs part-of-pathway), taxon-
inappropriate term removal, and legacy-TF de-annotation of proteins later shown to
be ESCRT/Golgi/vesicle machinery (TSG101, TMF1, USO1). - Residual follow-ups (not action errors, just outstanding work):
TRA2B— 67 annotations still atPENDING; needs a full initial review, out of
scope for this re-review pass.- A handful of Needs literature items (below) where I declined to adjudicate
without full-text access (e.g.ASCL1,ATF3).
The per-batch progress log is retained below as the audit trail.
Method notes / calibration
protein binding(GO:0005515) REMOVE is standard per project guidelines when
it is the only MF content of a row; ACCEPT/keep only when no better term and
the interaction is load-bearing.- Do not REMOVE experimental (IDA/IMP/IPI/IGI) annotations merely because the
cached abstract foregrounds another gene — curators read full text. Downgrade
to UNDECIDED / note under Needs literature instead. existing_annotations[].term.idcome from GOA and are trusted; do not rewrite
them to satisfy validation.- Over-annotation calls (MARK_AS_OVER_ANNOTATED / KEEP_AS_NON_CORE) are the most
subjective; scrutinize these plus REMOVE / MODIFY / NEW most closely. - Recurring threshold judgment: for IEA/ISS phenotype-derived process terms on
metabolic enzymes (e.g. VLCAD "temperature homeostasis", "regulation of cholesterol
metabolic process"), reviewers sometimes use REMOVE where KEEP_AS_NON_CORE /
MARK_AS_OVER_ANNOTATED would be the more conventional call. These are within reviewer
discretion and I am not editing them unless the term is biologically wrong (not just
peripheral). Flagging the pattern here rather than per-gene. - Observed reviewer quality is high: catches include name-confusion mis-annotations
(ACAT1 vs SOAT1 cholesterol acyltransferase; AARSD1 vs AARS1 synthetase) and
substrate-specificity over-calls (ACOX2 branched-chain vs ACOX1 straight-chain). - Digest caveat: my scan hides ACCEPT/KEEP rows to reduce noise, so a gene with many
REMOVEs of generic terms can LOOK like it lost core function (e.g. FZD7 removing
"Wnt signaling pathway"/"signal transduction" while ACCEPTing the specific "Wnt
receptor activity"/"canonical Wnt signaling pathway"). When core-sounding terms are
removed, verify the specific child terms are ACCEPTed before flagging. Verified OK on
FZD7, ELOVL5, BCS1L, CFTR.
Completed systematic audits
- All NOT/negated annotations (143 total across human genes) — fully adjudicated.
Distribution: 127 ACCEPT, 8 KEEP_AS_NON_CORE, 3 MARK_AS_OVER_ANNOTATED, 2 UNDECIDED,
and (originally) 4 REMOVE. Reviewed every REMOVE-on-NOT: - BRCA2 GO:0004402 (NOT HAT) — FIXED REMOVE→ACCEPT (NOT is correct; reviewer
agreed with its content yet removed it — logical error). - TOMM40 GO:0005739 (NOT mitochondrion) — REMOVE correct: the NOT is a false
negative (TOMM40 is the core TOM-complex channel), so removing it is right. - AGR2 GO:0034976 (NOT ER-stress response) — REMOVE correct: contradicted by
direct positive evidence (PMID:34237462). -
AGO3 GO:0035279 (NOT miRNA silencing) — REMOVE correct: outdated 2004 negative
superseded by 2017 slicer evidence (PMID:29040713).
Principle: REMOVE-on-NOT is right only when the negative is itself erroneous/superseded;
wrong when the negative is a correct, useful "absence" annotation (keep those). -
Experimental (IDA/IMP/EXP/IGI) REMOVEs of non-binding terms — partial spot-check.
A scan for REMOVE-on-experimental-evidence (excluding protein-binding, which is
routine cleanup) surfaced ~18 process/activity IDA REMOVEs; ALL reviewed as
defensible — they target genuinely-wrong annotations: "protein phosphorylation" on
non-kinases / phospho-substrate TFs (CREB1, RUNX3, MAML1, MORC3, ILF3) or non-protein
kinases (IP6K3 inositol kinase, GLYCTK glycerate kinase, CDC25B a phosphatase);
wrong compartment (TOMM40 inner membrane — it is outer); wrong activity type (SIRT2
poly-ADP-ribosyltransferase — sirtuins deacetylate; ELOB/ELOC transcription
initiation — they are elongation factors; ABHD2 nuclear-receptor; AIP PPIase).
CAVEAT: the scan's awk errored on some files (ASCL1, ATF3, ...), so this is NOT a
complete audit — those two are captured separately under Needs literature via the
reliable per-gene digest sweep, which remains the primary method.
Progress log
Cursor = last gene fully re-reviewed. Genes are processed alphabetically.
| Batch | Genes | Result |
|---|---|---|
| A-01 | AAAS AARSD1 AASS AATF ABAT ABCA1 ABCA7 ABCB7 ABCD3 ABCD4 ABCE1 ABCF2 ABHD2 ABI2 ABI3 ABL1 | agree; 1 edit (ABL1) |
| A-02 | ABRAXAS1 ABRAXAS2 ABTB1 ABTB2 ABTB3 ACAA2 ACAD9 ACADM ACADS ACADVL ACAT1 ACIN1 ACOX2 | agree; 0 edits |
| A-03 | ACSL4 ACTB ADA ADAM10 ADAMTSL4 ADM2 ADRB2 ADRM1 ADSL ADSS1 AFAP1 AFAP1L1 AFG3L2 AGK | agree; 1 edit (ADRM1) |
| A-04 | AGL AGO1 AGO2 AGO3 AGO4 AGR2 AGR3 AGRN AGXT AGXT2 AHCTF1 AHCY AHR AHSA1 | agree; 0 edits (2 NOT-removals verified OK) |
| A-05 | AIFM2 AIMP1 AIMP2 AIP AIPL1 AIRE AKIRIN1 AKIRIN2 AKR1D1 AKT1 AKTIP ALAD ALAS2 ALDH5A1 ALDOA ALDOB | agree; 0 edits (verified 2 "odd" calls OK: ALDH5A1 cofactor-specificity, ALDOB miscited-ref) |
| A-06 | ALG1 ALG10 ALG11 ALG12 ALG13 ALG14 ALG2 ALG3 ALG5 ALG6 ALG8 ALG9 AMACR AMBRA1 AMN AMPD1 | agree; 0 edits (ALG N-glyco family uniformly sound) |
| A-07 | AMPD2 AMPD3 AMT ANAPC2 ANG ANKFY1 ANKS6 ANKZF1 AP1B1 AP3B1 AP3B2 AP4B1 APBB1 APH1A APH1B API5 | agree; 0 edits (AP4B1 non-clathrin catch verified) |
| A-08 | APIP APLP1 APOE APP APRT ARCN1 AREL1 ARF1 ARG1 ARIH1 ARL6 ARL6IP1 ARL8A ARL8B ARNT ARSA | agree; 0 edits (APRT/ARL6IP1 good catches; ARL8B G-protein-activity call debatable) |
| A-09 | ARSB ASAH1 ASAH2 ASCC1 ASCC2 ASCC3 ASCL1 ASL ASS1 ATAD1 ATAD3A ATF2 ATF3 ATF4 ATG14 ATG2A | agree; 0 edits (2 needs-lit: ASCL1, ATF3 REMOVE-on-IDA) |
| A-10 | ATG2B ATG4D ATG5 ATG7 ATIC ATL3 ATP13A1 ATP23 ATP5F1A ATP5F1B ATP5IF1 ATP5MC1 ATP5MC2 ATP5MC3 ATP6AP1 ATP6AP2 | agree; 0 edits (ATP13A1 dislocase & ATP23 DNA-PK catches good; ATP23 DSB-process removals debatable — see fence) |
| A-11 | ATP6V0A1 ATP6V0A2 ATP6V0A4 ATP6V0B ATP6V0C ATP6V0D1 ATP6V0D2 ATP6V0E1 ATP6V0E2 ATP6V1A ATP6V1B1 ATP6V1B2 ATP6V1C1 ATP6V1C2 ATP6V1D ATP6V1E1 ATP6V1E2 ATP6V1F | agree; 0 edits (V-ATPase family uniform; ATP6V0C ATP-synthase removal correct — V≠F ATPase) |
| A-12/B-01 | ATP6V1G1 ATP6V1G3 ATP6V1H ATP7B AUH AUP1 AVEN AXIN1 AZI2 AZIN2 B3GALNT2 B4GALT1 BAAT BACE1 BACH2 BAG1 BAG2 BAG3 BAG6 | agree; 0 edits (ATP7B exemplary ATP7A/ATP7B paralog disambiguation; AZIN2 pseudoenzyme catch) |
| B-02 | BAIAP2 BAIAP2L1 BAIAP2L2 BBIP1 BBS1 BBS10 BBS12 BBS2 BBS4 BBS5 BBS7 BBS9 BCAP31 BCAT2 BCCIP BCKDHA BCKDHB BCL2 | agree; 0 edits (BCL2 "positive reg apoptosis" & DNA-binding removals correct — it's anti-apoptotic, not a TF) |
| B-03 | BCL2L1 BCL2L12 BCL2L13 BCS1L BDH1 BECN1 BECN2 BICC1 BIN1 BIRC5 BIRC6 BNIP3L BOLA3 BPGM BRAF BRCA1 BRCA2 BTD | agree; 1 edit (BRCA2 NOT-HAT REMOVE→ACCEPT) |
| B-04/C-01 | BTF3 BTN3A3 C18orf21 C1QBP CACNB2 CACUL1 CACYBP CAD CALCA CALCOCO1 CALCOCO2 CALM1 CALM3 CALR CALR3 CAMK2A CAMLG CAND1 CAND2 CANX | agree; 0 edits (CAD citrulline removal = CPS2/CPS1 catch; CALCOCO2 PML/IFN IDA removals borderline-OK) |
| C-02 | CAPG CASP12 CASP14 CASP3 CASP9 CASS4 CBLIF CBS CCDC28B CCDC47 CCDC50 CCL11 CCNE1 CD247 CD28 CD2AP CD320 CD33 CD47 CD8A | agree; 0 edits (CAPG severing, CASP3 aspartic/CDK-inhibitor, CCNE1 kinase-activity removals all correct regulatory/class catches) |
| C-03 | CDC25B CDC37L1 CDCP1 CDH1 CDH23 CDK1 CDK2 CDK5 CDK5R1 CDK5RAP3 CDKN1C CERS1 CERS2 | agree; 0 edits (CDK1 virus-receptor & CERS2 DNA-binding[Hox-motif artifact] removals correct; CDH1 neg-adhesion IMP-remove borderline) |
| C-04 | CERS3 CFAP300 CFAP418 CFAP61 CFTR CGRRF1 | agree; 0 edits (CERS3 Hox-motif DNA-binding removal; CFTR isomerase/cholesterol/basolateral removals all correct) |
| C-05 | CHAF1A CHAF1B CHAMP1 CHCHD4 CHIC2 CHMP1A CHMP1B CHMP2A CHMP2B CHMP3 CHMP4A CHMP4B CHMP4C CHMP5 CHMP6 CHMP7 | agree; 0 edits (ESCRT-III family consistent; CHMP1A metallopeptidase removal & CAF-1 histone-chaperone NEW correct) |
| C-06 | CHRAC1 CIAO1 CIRBP CKAP2 CKB CKM CKMT1A CKMT2 CLCN7 CLGN CLPS CLPSL1 CLPSL2 CLPX CLTC CLU CNOT4 CNPY3 | agree; 0 edits (CHRAC1 DNA-pol removal, CKAP2 transcription removal, CLCN7 channel→antiporter, CLPX heme NEW all correct) |
| C-07 | COLGALT1 COLGALT2 COMP COP1 COPG1 COPG2 COPS2 COPS8 COX10 COX14 COX15 COX20 COX4I1 COX4I2 COX5A COX5B COX6A1 COX6A2 | agree; 0 edits (COX10 squalene-synthase-term removal correct; COX-subunit catalytic→structural distinction well done) |
| C-08 | COX6B1 COX7B COX8A CPOX CPS1 CPT1A CPT1C CPT2 CR1 CRBN CREB1 CRISP2 CRISP3 CRMP1 CRTAP CRY1 CRY2 CRYAA CRYAB CSN1S1 | agree; 0 edits (standout catches: CPS1≠CPS2 urea/pyrimidine; CRY1/2 not-a-photoreceptor taxon-fix; CRMP1/CPT1C pseudo-enzymes; CRYAA holdase-not-refoldase) |
| C-09 | CSNK1D CSNK2B CTBP1 CTH CTLA4 CUBN CUL3 CWC27 CYB5D2 CYB5R4 CYC1 CYCS CYP11A1 CYP11B1 CYP11B2 CYP17A1 CYP19A1 CYP21A2 CYP27A1 CYP51A1 | agree; 0 edits (CYP family substrate-specificity; CYP11B1/B2 cortisol-vs-aldosterone paralog disambiguation) |
| D-01 | CYP7A1 CYP7B1 DAB2IP DAP DBT DCAF10 DCAF11 DCAF12L2 DCN DDA1 DDB1 DDB2 DDRGK1 DEGS1 DENR DGUOK DHCR24 DHCR7 DHODH DHTKD1 | agree; 0 edits (DHODH dihydroorotase≠dehydrogenase removal; DGUOK dCK-substrate demote; DAB2IP GAP-direction fix) |
| D-02 | DLAT DLD DLST DLX1 DNAJA2 DNAJA4 DNAJB1 DNAJB11 DNAJB13 DNAJB2 DNAJB4 DNAJB5 DNAJB6 DNAJB8 DNAJC1 DNAJC10 DNAJC11 DNAJC12 DNAJC13 DNAJC14 | agree; 0 edits (DNAJ family: J-proteins correctly lose "ATP binding" [no NBD]; DNAJB6 DNA-binding removal) |
| D-03 | DNAJC15 DNAJC16 DNAJC17 DNAJC19 DNAJC21 DNAJC22 DNAJC24 DNAJC25 DNAJC27 DNAJC28 DNAJC3 DNAJC30 DNAJC4 DNAJC5 DNAJC5B DNAJC5G DNAJC6 DNAJC7 DNAJC9 DNMT1 | agree; 0 edits (small DNAJCs clean; DNMT1 methyltransferase-specificity MODIFYs sound) |
| D-04 | DOLK DPAGT1 DPEP1 DPM1 DPM2 DPM3 DPT DPYD DPYS DPYSL2 DPYSL3 DPYSL4 DPYSL5 DRAM1 DRAM2 DRG2 DSCAM DTL DTYMK DUT | agree; 0 edits (DPYSL/CRMP pseudo-enzyme hydrolase removals consistent w/ CRMP1; DPYS real-enzyme kept) |
| E-01 | EBP ECHS1 EDEM1 EDEM2 EDEM3 EDF1 EEF2K EGFR EGR2 EIF2AK3 EIF2B4 EIF2D EIF3E EIF3J EIF4E2 EIF5A ELAVL3 ELOB ELOC ELOVL5 | agree; 0 edits (EGFR 160 protein-binding removals; PERK loses Tyr-kinase; ELOB/ELOC lose transcription-initiation; ELOVL5 core kept via IBA, IDA-dups UNDECIDED) |
| E-02 | EMC1 EMC10 EMC2 EMC3 EMC4 EMC6 EMC7 EMC8 EMC9 ENDOU ENO3 ENPP4 ENPP5 ENPP7 EPHA1 EPYC ERC1 ERG ERLEC1 ERLIN1 | agree; 0 edits (ENDOU protease→nuclease reannotation; ERG kinase-removal[it's a TF]; ERLEC1 unfolded→misfolded) |
| E-03/F-01 | ERLIN2 ERO1A ERO1B ERP27 ERP29 ERVMER34-1 ETFA ETFB ETFDH FADS1 FAF2 FAH FAS FBXL12 FBXL13 FBXL14 FBXL15 FBXL16 FBXL17 FBXL18 | agree; 0 edits (FADS1 δ5-vs-δ6[FADS2]/sterol-desaturase catches; ERO1 oxidase-not-reductase; F-box family clean) |
| F-02 | FBXL20 FBXL22 FBXL3 FBXL5 FBXL6 FBXL7 FBXL8 FBXO10 FBXO15 FBXO16 FBXO17 FBXO2 FBXO21 FBXO22 FBXO25 FBXO27 FBXO30 FBXO33 FBXO34 FBXO36 | agree; 0 edits (F-box family: transferase→substrate-adaptor MODIFY uniform; FBXO21 DNA-binding removal; FBXO17 carb-binding NEW) |
| F-03 | FBXO39 FBXO40 FBXO41 FBXO43 FBXO47 FBXO5 FBXO6 FBXO7 FBXO8 FBXW10 FBXW12 FBXW2 FBXW4 FBXW5 FBXW7 FBXW8 FBXW9 FDFT1 FECH FERMT2 FGFR2 FGFRL1 FKBP4 | agree; 0 edits (FBXO43/EMI2 APC/C-inhibitor NEW; FDFT1 correct FDFT home; FGFRL1 kinase-dead decoy; FECH 2Fe-2S NEW). Note: genes/human/FERROPTOSIS.md is a stray non-gene file (no review). |
| F-04/G-01 | FKBP5 FKBP8 FKBPL FN1 FN3K FOXO1 FTH1 FXN FZD7 G6PC1 G6PC3 G6PD GAA GAD1 GADD45A GADD45B GADD45G GALC GALE GALK1 | agree; 0 edits (FZD7 heavy REMOVE VERIFIED = generic-parent removal, specific Wnt terms retained; G6PC phosphotransferase→phosphatase; FXN autoprocessing removal) |
| G-02 | GALM GALNS GALT GAMT GAPDH GART GAS6 GATA3 GATD3 GATD3B GATM GBA GBA1 GBE1 GCDH GCG GCH1 GCK GCLC GCN1 | agree; 0 edits (GATA3 E-box-binding removal[binds GATA not E-box]; GCDH fatty-acid demote[glutaryl-CoA specific]; GAPDH aggressive moonlighting-REMOVE flagged as fence) |
| G-03 | GCSH GDPD2 GET1 GET3 GET4 GFER GIGYF1 GIGYF2 GK5 GLA GLB1 GLDC GLMN GLRX3 GLRX5 GLYCTK GM2A GMFB GMFG GMNC | agree; 0 edits (GMFB/GMFG "growth factor" misnomer→removed[actin/ARP2-3 regulators]; GLYCTK protein-phos removal; GM2A activator-not-enzyme; GLRX5 Fe-S NEW) |
| G-04 | GMPS GNAS GNMT GNS GOLGA8K GP9 GPAA1 GPATCH11 GPC2 GPC4 GPC6 GPI GPR101 GPX4 GRAMD1A GRHPR GRPEL1 GSK3B GSS GSTZ1 | agree; 0 edits (GNAS=O95467/NESP55 locus disambiguation VERIFIED correct[granin, not Gsα]; GPI moonlighting MARK; GRPEL1 loses unfolded-protein-binding[it's a NEF]; GSTZ1 tyrosine-catabolic NEW) |
| G-05/H-01 | GTF2F2 GTPBP1 GTPBP2 GTPBP6 GUSB GYG1 GYS1 GYS2 HADH HADHA HADHB HAO1 HAP1 HBS1L HCST HDAC4 HDAC6 HES1 HEXA HEXB | agree; 0 edits (HAP1 name-collision catch[huntingtin-assoc≠APE1, DNA-repair removed]; HADHA/HADHB subunit-specific split; HEXA/B hydrolase-not-transferase) |
| H-02 | HGD HGSNAT HINT2 HK1 HLCS HMBS HMGB1 HMGCL HMGCR HMGCS1 HMGCS2 HOGA1 HPD HPRT1 HPX HRAS HRC HSCB HSD11B1 HSD11B2 | agree; 0 edits (HPRT1 AMP-salvage removal[salvages hypoxanthine/guanine]; HPX heme-binding-not-transporter; HMGCS2 isoprenoid demote; HSCB Fe-S co-chaperone NEW; HSD11B2 7β-demote) |
| H-03 | HSD17B3 HSD3B2 HSD3B7 HSPA12A HSPA12B HSPA13 HSPA14 HSPA1A HSPA1B HSPA1L HSPA2 HSPA6 HSPA8 HSPA9 HSPB2 HSPB3 HSPB6 HSPB7 HSPB8 HSPB9 | agree; 0 edits (HSP70 virus-receptor removal; HSPB2/6 eye-lens-crystallin removal[not CRYAA/B]; HSPA1B twin-gene IDA-core removes flagged as fence) |
| H-04/I-01 | HSPG2 HTT HYPK IBA57 ICA1 ICA1L IDH3B IDS IDUA IFI16 IFI30 IFIT2 IFIT3 IFNL4 IGFBP3 IKZF1 IL10 IL13 IL15 IL21 | agree; 0 edits (IDH3B NADP→NAD distinction[IDH3≠IDH1/2]+structural-subunit; HTT phosphatase removal[scaffold]; IFI30/GILT disulfide-reductase NEW) |
| I-02 | IL22 IL23R IL2RA IL36RN IL4 IL4I1 IL7R ILF3 ILK IMPDH1 IMPDH2 INPP5D INS INTU IP6K3 IRF4 IRF8 ISCA1 ISCA2 ISCU | agree; 0 edits (ILK pseudokinase kinase-activity removal; IL23R prolactin-receptor removal; IP6K3/ILF3 protein-phos removals; Fe-S 4Fe-4S NEWs) |
| I-03/J-K | ISM1 ISM2 ITIH2 IVD JAK1 KCNQ1OT1 KCNRG KCTD10 KCTD11 KCTD12 KCTD14 KCTD16 KCTD18 KCTD4 KCTD7 KCTD8 KDSR KEAP1 KHK KIAA1614 | agree; 0 edits (ITIH2 heavy-chain peptidase-inhibitor removal[bikunin does it]; KCTD family disambiguation[Cul3-adaptor/GABA-B, not K-channel]; IVD substrate specificity) |
| K-L-01 | KIF5B KRAS LAMP1 LDHA LDHB LEMD2 LGALS3 LIMD1 LIPE LMAN1 LMAN1L LMAN2 LMAN2L LMBRD1 LONP2 LPCAT1 LPCAT2 LPCAT3 LRCOL1 LRP1 | agree; 0 edits (LIPE/LIMD1 phosphorylation-removal[phospho-substrates]; LMBRD1 not-ABC-transporter[ABCD4 is]; LONP2 peroxisomal-localization) |
| L-M-01 | LRPPRC LRRK2 LRSAM1 LSS LTN1 LYPD2 LYRM4 LYRM7 LZTFL1 MAD2L1 MAD2L2 MAGI1 MAML1 MAN1B1 MAP1S MAP3K20 MAP3K5 MAP7 MAP7D1 MAP7D2 | agree; 0 edits (LRRK2 cautious UNDECIDEDs for debated functions; MAML1 phosphorylation-removal[coactivator]; MAP7 family kinesin-binding NEW) |
| M-01 | MAP7D3 MAPK1 MAPT MAT1A MBL2 MCCC1 MCCC2 MCEE MCTS1 MED13L MEFV MEX3B MGAT1 MGP MIR155 MKKS MKRN1 MKRN2 MLEC MMAA | agree; 0 edits (MBL2 proteolysis/surfactant removal; MEFV/pyrin E3-removal[sensor]; MEX3B kinase-removal[RNA-binding E3]) |
| M-02 | MMAB MMACHC MMADHC MMGT1 MMS19 MMUT MORC3 MPDU1 MS4A4A MS4A6A MSMO1 MTCH2 MTHFR MTR MTRR MTX1 MTX2 MUC1 MVB12A MVB12B | agree; 0 edits (MMGT1 metal-transport demote[EMC subunit]; MMS19 TFIIH-removal[CIA factor]; MMACHC demethylase-demote; MORC3 kinase-removal) |
| M-N-01 | MVD MVK MYC NAA10 NAA15 NAA25 NAA30 NAA35 NAA38 NAA40 NAALADL2 NAGLU NAGS NBR1 NCOA4 NCSTN NDUFA4 NDUFS1 NDUFS2 NDUFS4 | agree; 0 edits (NDUFA4 Complex-I→Complex-IV reassignment; NCSTN/nicastrin protease-demote[presenilin catalytic]; NDUFS2 NAD→quinone-module; NCOA4 ferritinophagy) |
| N-01 | NDUFV1 NEMF NENF NEUROD1 NEUROG1 NEUROG2 NF1 NFE2L2 NFIA NFS1 NFU1 NID1 NLRX1 NME2 NOTCH1 NOVA2 NPHS2 NPLOC4 NPM1 NRAS | agree; 0 edits (NDUFV1 NADH-catalytic[vs NDUFS2 quinone]; NFS1 desulfurase-not-cluster-binding; NPM1 histone-chaperone NEW; NME2 moonlighting-demote) |
| N-O-P-01 | NSDHL NTN1 NTN3 NUFIP1 OLA1 OLIG2 OPTN ORMDL3 OTC OTUD3 OXCT1 P3H1 P3H2 P3R3URF P4HA1 P4HA2 P4HA3 P4HB PAICS PAM16 | agree; 0 edits (NTN1/NTN3 TF-activity removal[secreted ligands]; P4HB dioxygenase-demote[PDI β-subunit]; OTC inner-membrane→matrix; ORMDL3 SPT-inhibitor NEW) |
| P-01 | PANK4 PARD6A PARD6B PARD6G PARK7 PAX6 PC PCBD1 PCCA PCCB PCSK1N PDCD5 PDCD6IP PDGFA PDGFB PDHA1 PDHB PDHX PELO PEX1 | agree; 0 edits (PCBD1 not-PAH[BH4 cofactor]; PDGF ligand autophosphorylation-removal; PC viral-packaging removal; PDHX structural-not-E2; PARK7 debated-function UNDECIDED) |
| P-02 | PEX10 PEX11A PEX11B PEX11G PEX12 PEX13 PEX14 PEX16 PEX19 PEX2 PEX26 PEX3 PEX5 PEX6 PEX7 PFAS PFDN1 PFDN2 PFDN4 PFDN5 | agree; 0 edits (PEX family consistent; PEX2 Cdc73/Paf1-complex removal[it's a peroxisomal E3]; spurious PEX localizations removed; prefoldin cochaperone terms) |
| P-03 | PFDN6 PFKM PGAM2 PGAP1 PGAP2 PGAP3 PGD PGK1 PGLS PGM1 PGRMC1 PGRMC2 PHF23 PHKA1 PHKA2 PHKB PHKG2 PHTF1 PHYKPL PICALM | agree; 0 edits (PGRMC1/2 nuclear-steroid-receptor removal[membrane heme proteins]; PHKA1/A2 regulatory-subunit kinase-demote[γ=PHKG2 catalytic]; PHYKPL transaminase→lyase) |
| P-04 | PICK1 PIGA PIGB PIGC PIGF PIGG PIGH PIGK PIGL PIGM PIGN PIGO PIGP PIGQ PIGS PIGT PIGU PIGV PIGW PIGX | agree; 0 edits (PIG/GPI-anchor family uniform generic→specific; PICK1 phosphorylation-removal[BAR-scaffold, not a kinase despite name]) |
| P-05 | PIGY PIK3C3 PIK3CD PIK3R4 PIP5K1B PIWIL1 PKLR PKM PLCG2 PLD3 PLD4 PLD5 PM20D1 PMM2 PMPCA PMPCB PMVK PNP POFUT1 POLD2 | agree; 0 edits (PIK3C3/VPS34 lipid-not-protein-kinase; PLD3/PLD4 exonuclease-not-phospholipase; PIWIL1 heavy-REMOVE VERIFIED=generic→specific, piRNA/slicer retained) |
| P-06 | POLDIP2 POLE4 POMC POMT1 PPAT PPIB PPOX PPP2CA PPP2CB PPP2R1A PPP2R1B PPP3CA PPP3CB PPP4R1 PRDX4 PRG2 PRG3 PRPS1 PRRT1 PSAP | agree; 0 edits (POLE4 polymerase→histone-chaperone; POMC response-to-MSH removal[it's the precursor]; PPP3CB/PPP4R1 phosphatase-not-kinase; PP2A Tyr→Ser/Thr) |
| P-Q-R-01 | PSEN1 PSEN2 PSENEN PSMA1 PSMB5 PTEN PTGES3 PTPN22 PTPN6 PTS PUS3 PYGL PYGM QDPR RAB15 RAB24 RAB7A RAB7B RAB9A RAB9B | agree; 0 edits (PTPN6/SHP-1 phosphorylation-removal[phosphatase]; PUS3 mRNA→tRNA specificity; PSEN1/2 kept-as-protease[vs NCSTN demoted]; Rab family uniform) |
| R-01 | RACK1 RAD18 RARA RASA1 RASA2 RASA3 RASA4 RASA4B RASAL1 RASAL2 RASAL3 RB1 RBFOX3 RETREG2 RFT1 RFXANK RFXAP RGS20 RHBDF1 RIMBP2 | agree; 0 edits (RASA1/RGS20 GTPase-activity removal[GAPs activate, aren't GTPases]; RB1 DNA-binding-TF demote[binds E2F]; RFXANK/AP non-DNA-binding subunits) |
| R-S-01 | RNF14 RNF166 RNF170 RNF185 RNF25 RNF41 RNF5 ROR1 RPE RPIA RPL18A RPS3 RRM2B RUBCN RUBCNL RUNX3 SAMD8 SAMM50 SC5D SCAMP1 | agree; 0 edits (ROR1 pseudokinase kinase-removal; RNF185/RNF5 mannosidase-removal[ERAD E3s]; SAMD8/SMSr sphingomyelin→ceramide-PE synthase; RUNX3 ATP/phos removal) |
| S-01 | SCAMP2 SCAMP3 SCAMP4 SCAMP5 SCG5 SCGB1A1 SCGB1C1 SCGB1C2 SCGB2A2 SCN1A SCN9A SCO1 SCO2 SCP2D1 SDHA SDHB SDHC SDHD SEC11A SEC11C | agree; 0 edits (SDH/Complex-II "proton-motive-force ATP synthesis" removal[Cx II doesn't pump H+]; SCN1A neuronal-not-cardiac localization; SCGB1A1 secretoglobin spurious-nuclear removal) |
| S-02 | SEC62 SEC63 SERP1 SERP2 SERPINH1 SFT2D3 SGCA SGCE SGSH SGTA SH3GLB1 SIAH1 SIRT1 SIRT2 SLC14A1 SLC22A5 SLC25A13 SLC25A15 SLC25A20 SLC27A5 | agree; 0 edits (SERPINH1/HSP47 non-inhibitory-serpin[collagen chaperone]; SIRT1 acetylation-direction removal; SIRT2 PARP removal; SLC14A1 urea-not-water). SIRT1 heavily pruned[86 REMOVE] like GAPDH but core deacetylase retained. |
| S-03 | SLC37A4 SLC3A2 SLC40A1 SLC6A8 SLC7A11 SMAD3 SMPD1 SNORD116-1 SNORD3A SOCS1 SOCS3 SOCS4 SOCS5 SOD1 SORL1 SOX2 SOX9 SPARC SPCS1 SPCS2 | agree; 0 edits (SLC6A8 creatine-not-GABA transporter; SOD1 IMS-not-matrix; SLC3A2/CD98hc antiporter/ferroptosis NEW; SOCS Cul5-adaptor+kinase-inhibitor; SPARC matricellular-not-structural) |
| S-04 | SPCS3 SPDL1 SPG11 SPI1 SPNS1 SPOCK1 SPOCK2 SPOCK3 SPR SPTLC1 SPTLC2 SPTSSA SQLE SQSTM1 SRD5A2 SRD5A3 SRP19 SRP68 SRP72 SRP9 | agree; 0 edits (SRD5A3 steroid-reductase demote[actually polyprenol reductase, a CDG gene]; SPTSSA enzyme-activator NEW; SPOCK metalloprotease-inhibitor; SQSTM1/p62 clean) |
| S-05 | SRPRB SSR1 SSR2 ST6GAL1 STAR STAT1 STAT2 STAT3 STAT4 STAT5B STAU2 STC1 STC2 STIP1 STOM STOML1 STOML3 STUB1 STX12 SUMF2 | agree; 0 edits (STAT family DNA-binding refined + spurious-receptor removal[STAT1 CCR5]; STC1/STC2 PAPP-A-inhibitor NEW; STOML3 mechanosensation; STUB1/STAT3 clean cleanups) |
| T-01 | SURF1 SYCE3 SYN1 SYN2 SYN3 SYNGAP1 SYP SYVN1 TALDO1 TANK TARDBP TAT TAX1BP1 TBK1 TCF25 TCN1 TCN2 TERT TFRC THBS1 | agree; 0 edits (SURF1 COX-assembly-factor not subunit→REMOVE COX activity; SYN1 REMOVE synaptonemal-complex[synapsin≠meiotic SC]; SYVN1/HRD1 ER-membrane E3 cleanup; TERT digest-caveat VERIFIED—specific telomerase terms ACCEPTed, marks are generic parents+IEA moonlighting, REMOVE ALT/recombination correct) |
| T-02 | THBS2 THBS3 THBS4 TIA1 TIMM17A TIMM21 TIMM22 TIMM23 TIMM44 TIMM50 TK2 TKFC TKT TMA16 TMA7B TMEM43 TMEM67 TMEM70 TMF1 TNFAIP3 | agree; 0 edits (TIMM subunits generic→specific-complex refinement; TIMM50 marks cysteine-endopeptidase[inactive rhomboid-like receptor]+REMOVE IL2R binding; TMEM43 UNDECIDED on disputed voltage-gated-channel activity; TMF1 reassigned legacy-TF→Golgi golgin/tethering) |
| T-03 | TNFRSF1A TNKS TOLLIP TOMM20 TOMM22 TOMM40 TOMM5 TOMM6 TOMM7 TOMM70 TOP2A TOP2B TP53 TPI1 TPM1 TPM3 TRA2B TRAF6 TRAPPC1 TRAPPC11 | agree; 0 edits (TNKS PARP not kinase→REMOVE peptidyl-ser/thr-phos; TOMM20 REMOVE protein-transporting-ATPase[receptor]; TOMM40 REMOVE inner-membrane[it's outer-membrane channel]; TP53 873ann/344REMOVE downstream+binding cleanup, core DNA-binding-TF/apoptosis/checkpoint ACCEPTed VERIFIED, NEW ferroptosis; TRA2B=known 67-PENDING full-review, out of scope) |
| T-04 | TRAPPC12 TRAPPC13 TRAPPC3 TRAPPC4 TRAPPC5 TRAPPC8 TREM2 TRIM13 TRIM16 TRIM17 TRIM5 TRIP4 TSG101 TTC19 TTC28 TTC8 TXN TXNDC11 TXNDC12 TXNDC16 | agree; 0 edits (TSG101 legacy-TF de-annotation[ESCRT-I not transcription regulator]; TRIM5/16/17→autophagy-cargo-adaptor; TXN digest-caveat VERIFIED—keeps protein-disulfide-reductase GO:0015035[core] but REMOVEs GO:0004791(=TXNRD's activity, wrong enzyme)+GO:0047134(NAD(P)H-direct, wrong cofactor); textbook wrong-enzyme/cofactor catch) |
| U-01 | TYMP TYMS UBA5 UBA7 UBAC2 UBAP1 UCHL1 UCK1 UCK2 UFC1 UFD1 UFL1 UFM1 UFSP1 UFSP2 UGGT1 UGP2 UMAD1 UMPS UPB1 | agree; 0 edits (enzyme-precision batch: TYMP REMOVE glycogen-phosphorylase[wrong enzyme]; UBAC2 REMOVE serine-endopeptidase[inactive rhomboid pseudoprotease]; UFD1 REMOVE DUB[p97 adaptor]; UGGT1 REMOVE ER-mannose-trimming[it's a glucosyltransferase]; UMPS REMOVE UDP/CTP-biosynthesis[makes UMP only]) |
| U-02 | UPF1 UPF2 UPF3A UPF3B UQCC2 UQCC3 UQCRB UQCRC1 UQCRC2 UQCRFS1 UQCRH UQCRQ UROD UROS USO1 USP10 USP21 USP25 USP8 UVRAG | agree; 0 edits (Complex-III cores UQCRC1/UQCRC2 REMOVE/MARK MPP-peptidase activity[non-catalytic MPP-homologs in mammals]; UROS REMOVE mitochondrion[cytosolic]+folic-acid binding; USO1/p115 refined spurious transcytosis/cadherin/RNA→Golgi-tethering[SNARE/GTPase]) |
| V-01 | VBP1 VCP VEGFA VMP1 VOPP1 VPS28 VPS37A VPS37B VPS37C VPS37D VPS4A VPS4B VTI1A WDR6 WFS1 WIPF1 WT1 XDH XIST YWHAZ | agree; 0 edits (VEGFA REMOVE 'cellular response to VEGF'[ligand≠responder]+cautious UNDECIDED on ligand→process; VCP/YWHAZ-14-3-3ζ big moonlighting/adaptor cleanup[14-3-3 not a kinase→MODIFY phos]; XDH pathway-step precision; WT1 IDA-transcription-activation REMOVE=documented isoform-repressor case[ISOFORMS/PMID:9815658]+taxon-inappropriate Drosophila terms removed) |
| Z-01 | ZFYVE26 ZNF598 ZSWIM8 | agree; 0 edits (ZFYVE26/spastizin autophagosome-maturation NEW; ZNF598 RQC-E3 clean; ZSWIM8 MODIFY→specific target-directed-miRNA-decay terms) — END OF ALPHABET; sweep complete |
Genes needing a full (not re-) review
- TRA2B — 67 annotations all still
action: PENDING; never reviewed. Needs a
full annotation-reviewer pass, out of scope for this re-review but flagged here.
Edits made
(gene — annotation — old action → new action — rationale)
-
ABL1 — GO:0004515 nicotinate-nucleotide adenylyltransferase activity (TAS,
PMID:20841568) — UNDECIDED → REMOVE. Biologically implausible (ABL1 is a
tyrosine kinase, no NAD-biosynthesis role); the cited reference is a review of
ABL kinase function and does not describe NMNAT activity (supporting_text is just
the paper title). Spurious annotation, meets the "demonstrably wrong" REMOVE bar. -
ADRM1 — GO:0006368 transcription elongation by RNA Pol II (IMP, PMID:11818576)
— REMOVE → UNDECIDED. Experimental (IMP) annotation; cited paper is abstract-only
in cache (cannot verify full text). Per project guidelines, unverifiable experimental
annotations should be UNDECIDED not REMOVE. Biology implausible (ADRM1/Rpn13 is a
proteasome subunit, not Elongator ELP1-6), so flagged for curator rather than removed. -
BRCA2 — GO:0004402 histone acetyltransferase activity, negated/NOT (IDA,
PMID:9824164) — REMOVE → ACCEPT. The NOT annotation correctly documents that
BRCA2 lacks intrinsic HAT activity (activity comes from associated P/CAF). The
reviewer's own reason agreed with the NOT's content yet chose REMOVE — a logical
error. A correct, well-supported negative annotation is valuable and should be kept
(it guards against the dubious positive HAT claims GO:0010484/85 resurfacing).
REMOVE means "annotation is wrong"; here it is right, so ACCEPT.
Fence cases (judgment call, no edit made)
(gene — annotation — the tension)
-
AGO3 / AGR2 — REMOVE of negated (NOT) experimental annotations — reviewed and
agreed, but flagging the pattern. AGO3 GO:0035279 NOT (IDA, 2004) removed because a
2017 paper (PMID:29040713) directly shows guide-dependent AGO3 slicer — legitimate
supersession of an outdated negative. AGR2 GO:0034976 NOT (IMP, PMID:25666625)
removed: the NOT-paper is actually about EGFR maturation (verified full text, no ER-
stress assay), and PMID:34237462 shows an AGR2 mutant impaired in alleviating
tunicamycin ER stress — direct positive contradicting the gene-level NOT. Both OK,
but removing NOT annotations warrants curator awareness. -
ATP23 — DSB-repair / DNA-PK moonlighting removals — reviewed, agree on the
reannotation but flagging. ATP23 was formerly XRCC6BP1/KUB3 and carries DSB-repair
(GO:0006303, TAS) and DNA-PK-complex (GO:0005958, NAS) annotations from that identity.
REMOVE of "DNA-dependent protein kinase activity" (GO:0004677) is clearly correct (it
is not a kinase). REMOVE of the DSB-repair process/complex is more debatable — those
reflect documented (if minor) KUB3 associations, and the reviewer's mitochondrial-
protease reannotation is strong but supersedes rather than disproves the moonlighting. -
GAPDH — aggressive REMOVE of documented moonlighting — reviewed, not edited.
~43 REMOVEs; generic-parent and NADP-binding removals clearly correct (GAPDH is
NAD-dependent). But several removed terms are well-documented moonlighting (GAIT-
complex "negative regulation of translation" IMP/IDA, S-nitrosylation functions,
"cellular response to type I interferon" IDA, "ribonucleoprotein complex" IDA).
Guideline would prefer KEEP_AS_NON_CORE for real-but-non-core functions; REMOVE is
stronger than warranted for the well-supported ones. Consistent with project's
anti-over-annotation stance so left as-is, flagged for curator. -
HSPA1B — REMOVE of own IDA core-chaperone annotations — reviewed, not edited.
HSPA1B (HSP70-1B) is ~99% identical to HSPA1A (HSP70-1A). HSPA1A's review keeps more
core chaperone terms while HSPA1B REMOVEs several of its own IDA/IMP annotations
("protein refolding", "unfolded protein binding", "cellular response to heat",
"ATP hydrolysis activity", "cellular heat acclimation"). Core function is still
retained on HSPA1B via other-evidence rows, but the twin-gene handling is
inconsistent — verify these IDAs aren't genuine HSPA1B experimental data being lost
(guideline: don't REMOVE experimental annotations you can't refute).
Needs literature (cannot adjudicate from YAML + cache)
(gene — annotation — the specific question a lit search must answer)
- ASCL1 — GO:0045665 negative regulation of neuron differentiation (IDA and
IEA) both REMOVEd. ASCL1 is the canonical proneural bHLH activator, so the
annotation is counterintuitive and the reviewer instead added glial-fate repression
(GO:0045686). But ASCL1 has a documented progenitor-maintenance / proliferation role
(Kageyama oscillatory-vs-sustained model) that could underlie a genuine "negative
regulation of neuron differentiation" IDA. Verify the IDA source paper before
endorsing removal; guideline-preferred fallback would be UNDECIDED for the IDA. - ATF3 — GO:0001228 DNA-binding transcription activator activity (IDA) and
GO:0045944 positive regulation of transcription by Pol II (IDA) REMOVEd; the
reviewer treats ATF3 as primarily a repressor. But ATF3 has documented context-
dependent activator function (heterodimers with c-Jun). Removing the activator IDA
annotations may lose real biology; verify the source papers. Guideline-preferred
fallback: UNDECIDED / KEEP_AS_NON_CORE rather than REMOVE.
Slides
- Slides (Marp source: HUMAN_GENES_RE_REVIEW-slides.md) — AI generated