The proposal reviews all 45 current annotation contexts individually: 18 ACCEPT, 19 KEEP_AS_NON_CORE, six MODIFY, one UNDECIDED and one MARK_AS_OVER_ANNOTATED. It adds no annotation. The 45 original raw objects, their literal donor order, the actual exporter/parser forecast and the single urine extension remain unchanged. A genuine no-seed target fetch produced the same 45-context, 16-column export and 952-residue sequence. The inherited review was not regenerated. The one urine atom omitted by the parser is restored explicitly at raw41/native38; all original fields remain in the evidence archive.
Two catalytic cores describe the work GAA performs in lysosomal glycogen degradation: alpha-1,4 hydrolysis and complementary alpha-1,6 cleavage. The slower isomaltose reaction does not make branch cleavage incidental. The reported 32-fold specificity-constant comparison is between maltose and isomaltose under particular assay conditions, not a universal rate ratio for glycogen linkages. No independent role in exocytosis, membrane organization or extracellular-vesicle function is assigned.
PMID:29061980(https://pmc.ncbi.nlm.nih.gov/articles/PMC5653652/) is available as an immutable full-text cache. The selected Results, kinetic Methods and captions establish CHO-derived recombinant human GAA chemistry. The structural preparation was treated with alpha-chymotrypsin and retained comparable activity. Maltose and isomaltose assays measured glucose release at acidic pH; glycogen assays measured reducing equivalents. The enzyme is human even when the glycogen substrate is bovine or rabbit. Isomaltose accommodation was modeled by superposition with another enzyme, rather than observed in a human isomaltose co-crystal. A second acarbose site was observed near the experimentally removed G116–M122 loop, but its endogenous processivity role remains a hypothesis. The supplementary kinetic table and structural images were not independently inspected for this proposal.
PMID:1856189(https://pubmed.ncbi.nlm.nih.gov/1856189/) is abstract-only in the normal cache. Its original abstract reports active-site-directed conduritol-B-epoxide inhibition, glycogen/artificial-substrate competition, covalent labeling, and human GAA cDNA mutagenesis in COS cells. It distinguishes demonstrated catalytic importance from the initial mechanistic assignment of the catalytic carboxylate. PMID:18429042(https://pubmed.ncbi.nlm.nih.gov/18429042/) and PMID:7717400(https://pubmed.ncbi.nlm.nih.gov/7717400/) are also abstract-only; their human variant/function and residual-enzyme observations are retained without inventing unavailable assay conditions.
PMID:9505277(https://pubmed.ncbi.nlm.nih.gov/9505277/) explicitly reports CHO-produced human GAA uptake, lysosomal immunofluorescence, precursor processing and enzyme/glycogen correction in deficient human fibroblasts. Quail recipient experiments are stated separately. The abstract does not establish maltose or sucrose flux in those cells. Maltose retention instead combines the existing curator inference, direct GAA substrate chemistry and the official GO:0000023 definition, which explicitly includes maltose as an intermediate in glycogen/starch catabolism. The broad disaccharide context is non-core. Sucrose metabolism remains over-annotated because alpha-1,4 activity does not establish the distinct sucrose chemistry; no negative sucrose assay is invented.
PMID:24417(https://pubmed.ncbi.nlm.nih.gov/24417/) and PMID:5264799(https://pubmed.ncbi.nlm.nih.gov/5264799/) remain metadata/title-only after targeted access attempts. Their historical source identifiers remain literal. The two linkage activities and broad alpha-glucosidase classification can be supported independently by the later human biochemical study. The particular glycophagy experiment attributed to 5264799 cannot be reconstructed from its title. Its action nevertheless retains the combined gene-level glycophagy claim: GAA catalyzes the terminal degradation of autophagically delivered glycogen. The lysosome-organization claim is similarly retained from combined glycogen-clearance and verified ancestral evidence while the precise 7717400 endpoint remains unverified. The diaphragm-contraction annotation attributed to PMID:16917947(https://pubmed.ncbi.nlm.nih.gov/16917947/) remains UNDECIDED: the accessible abstract neither establishes direct participation in force generation nor exposes the underlying diaphragm measurements. No unavailable experiment is represented as read.
The normal PMID:17897319(https://doi.org/10.1111/j.1600-0854.2007.00643.x) cache is abstract-only. Separately retained original publisher text and selected Methods/Results/Table 1 identify human placental lysosomal alpha-glucosidase IPI00293088 with 24 peptides, 36.2% coverage and 112 spectral counts. The enriched lysosomal-membrane fraction was compared with dense-pool material. Peripheral proteins were not stripped. This supports non-core membrane association; the historical 957-residue database entry and predicted transmembrane count do not prove an integral topology or a present-day natural isoform. This external reading does not upgrade the normal cache's access flag.
The normal PMID:23533145(https://pmc.ncbi.nlm.nih.gov/articles/PMC3773505/) cache contains original main text. Its Methods and Results, together with the retained authentic Supplemental Table 2, establish P10253/GAA in SI Table 2 row 440, with 15, 14, 29 and 28 unique peptides across the pooled measurements. Protein-group labels mentioning processed forms do not establish form-specific peptide localization. Differential ultracentrifugation, a DTT/sucrose wash and an identification FDR below 1% were used. The paper explicitly notes soluble material and possible contamination.
The normal PMID:19056867(https://pubmed.ncbi.nlm.nih.gov/19056867/) cache contains only the abstract, despite a Full Text heading. The separately retained original NHLBI urinary-exosome table lists GAA NP_000143 and NP_001073272, with four and five peptides respectively and an explicit reference-2 link to PMID:19056867. Its imported GO columns are not additional experiments. Historical accessions do not determine current UniProt products.
These positive target observations support non-core retention of the two exosome-fraction annotations, while purity, intravesicular topology and biogenesis remain unresolved. The complete October 10 manual decision is retained in the inherited archive: it confirmed MO using the soluble-protein/EV-surface-receptor criterion. The new proposal explicitly reassesses that criterion. A demonstrated cognate EV-surface receptor is not a prerequisite for retaining positive preparation membership, and solubility does not negate the identified peptides. This does not assign GAA a receptor function or a physiological role in extracellular vesicles.
PMID:19946888(https://pubmed.ncbi.nlm.nih.gov/19946888/) remains abstract-only. The YTS membrane methods are described, but no retained authentic GAA-specific row or peptides were found. The general membrane claim is refined to the already-supported lysosomal-membrane descendant on independent human placental evidence. That gene-level refinement does not verify the unavailable YTS target observation or claim lysosomal resolution for that assay.
Successful official ContentService JSON responses establish target membership; failed direct HTML responses are retained access diagnostics and excluded as evidence. The six membrane TAS rows are supported by these literal pairs:
| Event | Granule input and human GAA | Plasma-membrane output and human GAA |
|---|---|---|
| R-HSA-6798739 | 6798738 → 6800877 | 6806267 → 6801041 |
| R-HSA-6798747 | 6799718 → 6801021 | 6806434 → 6801041 |
| R-HSA-6800426 | 6800437 → 6800913 | 6806514 → 6801041 |
All GAA members resolve to human P10253. These are specialized, non-core location assertions. No GAA catalystActivity occurs on the exocytosis events. The original granule-paper peptide/topology experiments were not read, and the modeled 70–952 range is not a separately measured natural isoform.
R-HSA-9036727 assigns normal GAA catalysis through 5357535 to the mature GAA set 5357562 in the lysosomal lumen. R-HSA-9036729 is a FailedReaction involving mutant set 5357599 (E521K/M318T); its loss-of-function status and normalReaction link identify the corresponding normal set/reaction. A failed mutant reaction is not a negative wild-type activity experiment. R-HSA-70221 includes the normal lysosomal reaction as a direct child; the broad pathway's cytosol field does not relocate that child.
The retained actual PTHR22762 tree/alignment places canonical P10253 at AN454/PTN002503674. Its ungapped 952 residues exactly match the current sequence. The alpha-1,4 IBD at AN318/PTN002333501 and the glycogen-catabolism/lysosome-organization IBDs at AN444/PTN002734140 lie on its ancestral path. Incoming donor sets match the retained IBD and main PAINT records. No target-path loss was found in those records. A human self-donor is a legitimate experimental descendant, not circular support; donor count is not a strength metric.
All six GO_REF assessments use previously retained authentic official method descriptions: InterPro, phylogenetic inference, location mapping, Ensembl transfer, ARBA, and combined methods. The literal donor GOA records were inspected, but their cited papers are not represented as newly read primary sources. Valid ancestry and chemical function do not turn all descendant phenotypes into core activities.
The current official UniProt record supplies one canonical 952-residue sequence and processing features, without an ALTERNATIVE PRODUCTS block. The signal peptide, propeptide and mature chain ranges are processing states. Experimental proteolysis, Reactome sequence ranges and historical proteomic accessions are kept separate from natural splice-product claims.
The complete original 45-row review, old journal and all three history records remain unchanged in GAA-inherited-evidence.json, including all 55 original claim/quotation associations. The quote audit found 54 literal matches and one whitespace-only match; the latter remains exactly historical rather than being silently repaired. Two existing exact snippets are reused in the current cores. The other 53 occurrences remain historical-only, including title-only snippets, the fragment “we used,” and broad lysosome quotations formerly attached to specific membrane claims. Four new snippets quote the selected original data rows reproduced in GAA selected original proteomics evidence. They add no experiment and do not alter any inherited quote. The two original core snippets remain exact. Both original core snippets pass exact substring checks against the actual native caches; their byte offsets and source identities are recorded in the evidence file.
There are 26 manually assessed original references plus one derivative selected-evidence record. The initial 78-source acquisition union is broader because native UniProt bibliography and family/pathway requirements differ from the set of sources supporting authored scientific claims. Unread bibliography is not labeled manually verified. Native acquisition refreshed only the target UniProt and GOA files. It preserved 28 existing sources and acquired only the 50 qualified absent sources. No provider output was requested.
The initial five parent-term refinements follow the current project rule; the follow-up extends the conservative membrane refinement to the second membrane row. The general O-glycosyl hydrolase and both alpha-glucosidase parent rows propose the already-asserted alpha-1,4 and alpha-1,6 activity terms. Carbohydrate metabolism proposes the existing glycogen-catabolic term. The subsequent source-review follow-up makes both broad membrane rows consistent by proposing only the established lysosomal-membrane descendant. Specialized Reactome rows retain their separate model evidence; neither broad row is used to reconstruct the unread YTS assay. All replacement paths were verified from retained official QuickGO typed edges, not inferred from names. Every path used here is is_a. This refinement adds no annotation or source and does not say the general claims are false. The initial version-2 changes remain archived; the current follow-up separately records the three action changes and scoped evidence/prose corrections. Both catalytic cores and their two exact quotations remain unchanged.
The initial 78 sources comprise 67 publication records, eight Reactome records and three existing family source files. Missing-only native acquisition added 48 publications and two Reactome records with force=False. All 28 existing source bytes remain exact. Native full text is available for PMID23533145, PMID29061980 and the additional bibliography record PMID22676651; this last availability is not a new manual reading or scientific claim. Failed full-text fallbacks for eight other newly fetched publications remain abstract-only, including the 54-character PDF extraction for PMID9529346. All actual stdout/stderr was read. The existing access limits and separate original-source readings above remain governing. All 25 reference titles were compared with the native records or unchanged method descriptions.
The owned PTHR22762 family review is outside this change. Only its three existing source inputs are used. The exact old review, journal and three histories remain in the inherited archive, and the current notes retain the entire old journal as a prefix. Native EDIT history and all target checks passed. At the initially published head, native status computation returned DRAFT because of three mixed-action advisories; the earlier COMPLETE sign-off remains exact history. Strict GOA conservation passed for all 45 contexts. The current follow-up passes target validation without advisories and strict conservation of all 45 contexts. Native status computation now returns COMPLETE; history, whitespace, source conservation and private rendering passed. The earlier DRAFT outcome remains a dated result, and neither historical status was silently rewritten.
Selected original proteomics data now reproduce four selected source tables with original URLs, artifact hashes and locators. This makes the external reading reviewable without changing any publication cache or its access flag. Re-reading original experimental data is direct source verification, not circular reuse of a GO annotation. Fraction purity and topology remain bounded limitations rather than claims of demonstrated contamination.
Carbohydrate-binding retention rests on the observed second acarbose-derived ligand site about 25 Å from the catalytic-site ligand in PMID:29061980. The broad GO definition does not require a lectin, adhesion or regulatory role. The processed preparation and hypothetical endogenous processivity role remain explicit limitations. No extra core, process annotation or natural product is proposed. One qualified missing publication was acquired natively as described below.
PMID:27805059(https://pmc.ncbi.nlm.nih.gov/articles/PMC5090863/) was independently read in selected original Methods/Results and Table 2, then acquired once through the native missing-only reference fetcher. CD9/CD63/CD81 immunocapture, reported TEM morphology, and short detergent-free on-bead digestion provide an independent urinary EV preparation. Table 2 identifies GAA/P10253 with 23–39% coverage among 49 selected proteins found across all three captures. In this source, extraexosomal describes outer-surface-associated material; it does not mean a nonvesicular fraction. Carryover/aggregate controls and the absence of a target-specific physiological EV-function experiment remain explicit limits. The historical B7Z5V6 entry is not a natural-product assignment.
The new cache contains full XML text; all 78 prior normal source bytes remain exact. The current union is 79 normal sources, with 27 manually assessed reference entries comprising 26 original sources and the derivative selected-evidence file. No other bibliography was fetched, refreshed or removed. Original images, raw spectra and full supplementary control tables were not independently assessed.