GANAB catalyzes early ER N-glycan glucose trimming as the alpha subunit of glucosidase II. The review separates measured enzyme chemistry, curated ancestry, original target-table observations and client-protein phenotypes.
PMID:10929008(https://doi.org/10.1093/glycob/10.8.815) was read as an official abstract and selected original Results/Methods/captions from the author-hosted paper. Both human alpha forms work with beta; alpha1 D542 and alpha2 D564 are the corresponding catalytic positions in current Q14697-1 (944 aa) and Q14697-2 (966 aa). Both complexes cleave Glc1/Glc2 high-mannose substrates under the reported conditions; Glc3 was not cleaved in that assay. Single-human-subunit baculovirus preparations form human/insect heteroenzymes, so their activity does not establish an isolated human alpha-only enzyme. Retention-signal manipulation of beta affects secretion/ER retention. No whole-paper or original-image inspection is claimed.
The current official GO:0106407 definition describes Glc2-to-Glc1 hydrolysis. The original human experiments support this and the subsequent Glc1-to-Man9 reaction, represented in two chemistry cores. The second core and five existing MODIFY recommendations retain the conditional proposal “Glc1Man9GlcNAc2 oligosaccharide glucosidase activity”. The preferred resolution is to reconcile the existing term with its documented broader usage, rather than infer an overlooked activity from a null ontology search.
The actual production FlyBase GO-CAM 680ad14200001222 has two GO:0106407 activities enabled by Gcs2alpha (FB:FBgn0027588): 680ad14200001292 directly provides input for 680ad14200001231. Both enabled_by evidence records cite human Q14697 through ISS/GO_REF:0000024. Their placement between glucosidase I and mannosidase is consistent with the two successive glucose removals established experimentally; the intermediate glycans are not explicit molecule nodes. The ER-lumen assertions separately use yeast P38138, so they must not be described as location evidence from Q14697. The human-accession index search alone missed Q14697 in these donor fields.
The local Drosophila module represents both inner-glucose removals under GO:0106407, while the generated CAZy mapping records EC:3.2.1.207 → GO:0106407 with ec2go provenance. The module is corroborating curation, not another primary experiment. The generated mapping is evidence of retained mapping usage; the attempted fresh official ec2go download returned HTTP 403. GH31 is polyspecific, so the mapping does not assign this activity to every family member. RHEA:56000 independently has the same GO crossreference despite representing the final-glucose reaction. Reconciling the existing term would preserve these model/module assignments. If curators instead retain its Glc2-only definition, a separate final-step term would require retyping the downstream GO-CAM activity and representing that step separately in the combined module, while retaining GO:0106407 for Glc2 hydrolysis. No model, module or upstream mapping is changed here.
The current graph explicitly has GO:0106407 is_a GO:0090599; calling the corresponding annotation change a refinement describes that verified relation. The alpha-D-glucose wording of GO:0090599 conflicts with the beta-D-glucose product in GO:0106407/RHEA and is treated as an unresolved representation issue, not a missing edge or disproved GANAB function. The official NCBI GH31 description identifies a retaining mechanism. That family-level evidence motivates the stereochemistry question, but the inspected human assays do not establish the released product anomer. The conditional definition therefore says D-glucose and retains generic parent GO:0015926 without asserting a RHEA defect. The single 10929008 excerpt in the first core summarizes both removals; the second core reuses that primary reference without repeating the quotation.
GO:0033919 instead names alpha-1,3-glucan polymer cleavage; its replacement remains a source-substrate-based lateral correction, not a descendant refinement or a claim of target lineage loss. GO:0004573 describes glucosidase-I Glc3 processing and is not used for GANAB. Both incoming RHEA identifiers remain unchanged. GO:0140277 is the ER-trimming process directly performed by GANAB.
All three current products were independently matched to direct FASTA and coordinate reconstruction. Q14697-3 is a 52-aa product with a premature stop/NMD caution in UniProt, not a demonstrated miniature enzyme. Do not transfer activity or experimental engineered-mutant localization to it.
PMID:33097077(https://doi.org/10.1186/s13023-020-01585-4): authentic full XML, selected WT/mutant Results/Figures 3–4 captions and Methods were read. Human N-terminal FLAG GANAB constructs were tested in HEK293/HeLa systems with GIIbeta/PDI localization reagents. WT distribution supports ER localization. Truncated/misspliced mutant patterns are not natural-product localization evidence; predicted activity effects are not a direct enzyme assay.
HPA target page: supported ER staining with HPA061426 in MCF-7, Rh30 and U2OS, with ER staining in serum-starved hTERT-RPE1. The textual record was read, not independently re-scored images. This organelle assay does not itself resolve lumen topology.
The Golgi ISS donor P79403 is pig GANAB. Its direct ER/Golgi evidence is PMID:2404989(https://doi.org/10.1083/jcb.110.2.309), now included as an assessed reference and linked from both Golgi rows; the authentic abstract and official donor record support a kidney tubular secretory-route distribution. The abstract does not independently identify the animal species; species comes from the official donor record. This is not direct human Golgi imaging. The donor's reverse human-derived melanosome inference is a separate source chain.
PMID:27259053(https://doi.org/10.1016/j.ajhg.2016.05.004): selected original functional Results/Figures 4–5 captions support effects on polycystin maturation and trafficking. The paper's 966-aa long and 944-aa short forms map to current -2/-1. Cilia still form despite loss of ciliary PC2, so neither GANAB cilium residence nor a direct transport function follows. Complete Methods/images/supplements were not read.
PMID:28375157(https://doi.org/10.1172/JCI90129): the normal cache and official PubMed record are abstract-only. Correct original routes failed, so the exact activity experiment remains unread; independent 10929008 establishes the activity refinement. The official corrigendum corrects mislabeled Pkd1 RT-PCR primers, not the GANAB findings. It was read directly. An initial wrong 90183 route is preserved privately as a failed lookup, not biological evidence.
The portable original-table evidence gives the exact target joins and stable primary locators for melanosome, urinary/salivary vesicle and HeLa RNA-contact studies. Positive detection is retained with specific limits. Historical isoform names and accession rows are not substitutes for peptide mapping; neither vesicle topology/biogenesis nor a melanogenesis activity is asserted. RNA contact after UV crosslinking does not identify a regulatory mechanism. GANAB meets the ion-count route, whereas its DESeq adjusted P is 0.930 and it is also detected in whole-cell lysate. Its signal peptide and mature ER-lumen location leave the mechanism of RNA encounter unresolved; these limits alone do not demonstrate a false positive or abundance artifact.
PMID:19946888 remains a source-specific uncertainty: its YTS membrane-proteome abstract is available, but the exact GANAB table was not recovered. Attempted spreadsheets are HTML errors. The membrane definition includes attached proteins, so known lumenal activity or lack of a TM segment does not establish that the annotation is false.
Official reactions 532667 and 548890 explicitly use glucosidase II complex 532671; component 532674 is human Q14697, chain 29–944, in the ER lumen. The viral models 9683663/9694364 and inhibitor-binding 9686790 reach this component through ER-alpha-glucosidase set 9682983. The CDH1 models 9932352/9932913 use the same catalyst and explicitly say the specific processing was not studied. Location is supported through actual membership and independent human evidence; modeled client functions are not assigned to GANAB. Old catalyst GO:0033919 labels in these records do not override the current chemical definition. Official HTML was read; failed ContentService and PathwayBrowser responses are not presented as JSON evidence.
The reused authentic PTHR22762 tree/MSA maps exact 944-aa human AN77 to PTN002503652. PTN002734202/AN68 supports enzyme/complex inheritance; PTN000509742/AN3 supports N-glycan processing. Actual IBDs, donor records and conserved catalytic positions were checked. Human self-donor evidence is legitimate; no target-path loss/negation was found. Chemical term scope and process specificity are corrected independently of sound descent. The unrelated open family-review YAML remains excluded from native source scope.
The official response and genuine 16-column export contain 39 original records. Two 10929008/GO0006491 IDA records differ in original source metadata but collapse into one native key, producing the actual 38-row seed. Both originals are represented in final 39, restoring one source record without a NEW annotation. Two literal extension atoms are restored on the urine and saliva rows. The compact source evidence contains original contexts, products and actual native joins; current judgments are only in the review YAML.
The initial publication used 34 normal sources: 19 PMIDs, 12 Reactome records and 3 immutable family files. Fresh bounded ownership/base checks passed; 25 existing sources were reused unchanged and 9 missing sources were acquired without force. All native source bytes, 39 original joins, 3 product records and 5 exact quote locations passed reconciliation. Two 10929008 anchors use the genuine cache line breaks with identical words; the 27259053 title uses native typography, and its full original is now cached without expanding the selected reading claim. No normal source was manually rewritten. Brief quotations are centralized with reference-only reuse; repeated occurrences count toward the 25-word per-source aggregate. Default validation passed with zero curation advisories. Strict GOA matched all current originals, native history passed, and the status audit computes COMPLETE while the source-specific membrane judgment remains UNDECIDED. Private rendering, links and scoped whitespace checks passed.
The source union is now 35 after one qualified, missing-only acquisition of PMID:2404989. All previous 34 scientific sources and the generated GO cache remain unchanged. The new cache is abstract-only, matching the scope claimed above. All 39 original machine contexts, three products, two extension atoms, 17 ACCEPT/12 MODIFY/9 KEEP_AS_NON_CORE/1 UNDECIDED actions and original seed remain unchanged. The review has 26 assessed references and five active short excerpts; the previous client quote remains historically locatable but is no longer attached to a molecular-activity claim. Default validation passed with zero curation advisories. Strict GOA matches all 39 original contexts; native codex EDIT history, computed COMPLETE status, private rendering, links and scoped whitespace checks passed. The one source-specific UNDECIDED judgment is retained.