The historical decision summary below is superseded by the 2026-09-27 audit appended
at the end, particularly the NAD(P)+ parent-term, glutamate-pathway, homotetramer and
clinical-source assessments.
UniProtKB: P51649 (SSDH_HUMAN). HGNC:408. EC 1.2.1.24. 535 aa precursor with an
N-terminal mitochondrial transit peptide (1..47); mature chain 48..535.
ALDH5A1 encodes succinate-semialdehyde dehydrogenase, mitochondrial (SSADH), the
NAD+-dependent enzyme catalysing the final step of GABA degradation (the GABA shunt):
oxidation of succinate semialdehyde -> succinate, which then enters the TCA cycle,
routing GABA carbon into central energy metabolism.
Mitochondrion / mitochondrial matrix. TransitPeptide 1..47. UniProt subcellular location:
Mitochondrion. Reactome places the reaction in the mitochondrial matrix (R-HSA-888548).
Confirmed by IDA (HPA), HDA/HTP proteomics (PMID:20833797 muscle mito phosphoproteome;
PMID:34800366 MitoCoP high-confidence mito proteome), and IBA.
SSADH deficiency (SSADHD; MIM:271980) = 4-hydroxybutyric (gamma-hydroxybutyric, GHB)
aciduria. Autosomal recessive; accumulation of GABA and GHB; developmental delay,
hypotonia, intellectual disability, ataxia, seizures, behavioural disturbance
[PMID:9683595, PMID:12208142, PMID:14635103, PMID:15037717].
Approved HGNC:408 symbol ALDH5A1, UniProt P51649, aliases SSADH/SSDH were checked.
The five canonical files matched main ba3ff58d7d2de76dbe3c24b16e05e12369f463fc;
canonical and alias PR searches and alias-directory checks found no overlap. The
25 original annotation source objects, two alternative products, 16 original
reference identities, GOA and UniProt are preserved. The completed decisions are
22 ACCEPT, two MODIFY and one KEEP_AS_NON_CORE, with no NEW annotation. Four donor
references were added and all 20 references manually assessed.
The original PMID:19300440 full Results/Methods
were recovered through indexed primary PMC text although direct opens were challenged.
SSA is the substrate in the ligand complex; only the ADP portion of soaked NAD+ is
resolved. The structures are not a succinate-product complex or evidence of an ADP
cofactor. NADH-formation assays support catalysis. The oxidant/reductant work includes
recombinant protein and overexpressing HEK293 cells followed by lysate measurements;
quantitative control of endogenous neural GABA flux remains an open question. Crystal
assembly and the biochemical tetramer observation support oligomerization, while
disease-mutant assembly effects inferred from structure are not direct measurements.
The local publication remains abstract-only.
The full original PMID:9683595 institutional PDF
was independently read by the annotation-reviewer peer and this author. Human
recombinant GST-SSADH, patient/relative cell activity and splice-genotype segregation
support enzyme identity. Clinical developmental/speech delay grounds the existing
non-core developmental association; the paper does not assay a CNS morphogenetic or
differentiation step. Its discussion reports no established residual-activity/GHB-to-
clinical-severity correlation. Historical ALDH4A1 allele labels here denote SSADH,
not the modern ALDH4A1 protein. The short full-text quotation in row 22 was checked
against page 407 and is stored as supporting_text_fulltext; the cache stays unchanged.
PMID:7814412 experimentally validates rat cDNA through bacterial activity and purified
rat-brain sequence, alongside human partial cDNA and sequence comparisons. The ISS
donor is rat P51650. Later human expression studies independently establish the
conserved reaction; the rat expression experiment is not relabeled as human.
All nine original GOA PMID caches were read. Local full-text flags follow machine
cache availability: seven are abstract-only, while PMID:20833797 and PMID:34800366
have full-text metadata. The former extraction actually contains abstract/Discussion
and omits important experimental sections/tables. The latter main article was read,
but neither individual ALDH5A1 supplementary identification was re-extracted. Both
curated mitochondrial annotations are accepted with independent human localization
support, without inferring contamination or unmeasured submitochondrial resolution.
GO_REF:0000052 verifies the immunofluorescence curation method; its original antibody
images were not independently re-scored. Matrix localization in the integrated core
is independently supported by the human literature and cached Reactome R-HSA-888548.
The cached PTHR43353 PAINT records identify PTN008681047 for SSADH activity/GABA
catabolism and PTN000192583 for mitochondrial localization. These are ancestral IBD
assertions, not a count of similar extant proteins. Human experimental descendants
are valid ancestral grounding; no circularity or function loss is inferred. The full
historical tree/MSA was not reconstructed. Twelve propagation blocks identify the
PAINT ancestral nodes for IBA and the traced donors or mapping identifiers for other
inferences, with rule predicates left UNRESOLVED where not inspected.
Mouse Q8BWF0 is Aldh5a1 (MGI:MGI:2441982). The cached GO-CAM
gocams/68d5ebd600002976/68d5ebd600002976-src.yaml models SSADH catalysis and GABA-shunt
participation with IMP PMID:11544478 and matrix context with IDA PMID:14651853.
The model supplies a positive pathway role rather than a gap warranting NEW.
The MGI donor graph
is labeled generated 2023-03-10; it is a historical source snapshot, not a fresh GOA
export. MouseMine
resolves its reference IDs: J:125589 is PMID:12065715 for succinate-process IMP,
J:128716 is PMID:17854388 for glutamate-process IMP, and J:86816 is PMID:14651853 for
mitochondrial evidence. J identifiers are not PMIDs.
One genuine Falcon request with automatic perplexity-lite fallback was launched
concurrently with standard GOA-publication caching, using isolated writable temporary
UV directories. Both providers failed during dependency retrieval from PyPI due to
DNS errors before provider contact; no provider artifact was produced or authored.
Logs: /tmp/ALDH5A1-fresh-research.log and /tmp/ALDH5A1-fetch-goa.log (9/9 original
PMIDs already cached). This is a manual primary-source review.
The three newly traced donor PMIDs 11544478, 17854388 and 12065715 were submitted to
the normal fetcher; /tmp/ALDH5A1-donor-fetch.log records DNS failures and 0/3 cached.
They remain required cache gates for a future recovery batch. PMID:14651853 was already cached. No cache
was manually fabricated or edited. Status is DRAFT until the required sources and
validation warnings are resolved. The notes-inclusive authored PMID set is the nine
original publications plus these four donor papers; there are no provider-only PMIDs.
The parent independently reviewed every annotation, reference assessment, core and
question and found no biological blocker. Source preservation and all 43 cached
supporting snippets were independently checked with case-sensitive whitespace
normalization. The clinical full-text excerpt was checked against the institutional
PDF, page 407. No action was changed merely to eliminate a validation warning.
PR #3268 remained at exact head 8d48f6840058c4387f535908b1eb1baedd1804b8; all five canonical file blobs matched before authoring. The three required donor records are now genuine normal-fetch artifacts imported by the parent without overwrites from source run 36297910960, attempt 1, head 41e41a94fb65b65ddc78627208d955fa1e7eb1c7. The source archive and staged record hashes were verified separately; each canonical publication byte sequence matches tmp/source6-canonical-import-receipt.json. All three are abstract-only, not newly available full papers. This supersedes the earlier operational missing-cache gate, not the full-source limitations.
Each recovered reference assessment now states actual local access and includes an exact abstract finding. No annotation reason/action, original source assertion, alternative product, core, original title, machine source, or published history has been rewritten. The publication census covers the entire gene tree, including DOI strings, institutional PDF and primary-page links, and confirms 13 required PMIDs and the reviewed Reactome event R-HSA-888548 are cached. There are no provider artifacts or new DOI-only publication gates. Nine raw UniProt bibliography entries remain separately inventoried because they are not invoked for retained substantive claims; the homotetramer claim is independently grounded in the existing human biochemical/structural references. The broader UniProt Reactome cross-reference R-HSA-916853 is likewise not an additional reviewed source assertion. These inventories do not replace source-specific evidence judgments.
Status is COMPLETE after closure of the three operational cache gates and warning-free targeted validation; the documented limits of abstract-only donor evidence remain explicit. The normal provider attempts from the full review are preserved as failed pre-contact attempts, with no fabricated report and no unnecessary repeat.
The targeted validator completed successfully with no review warnings. Exact preservation checks retain all 25 complete annotation objects, both alternative products, the core and 20 reference identities/titles. All 46 cached snippets are case-sensitive verbatim matches after whitespace normalization, and the unchanged external clinical excerpt matches its retained primary-PDF receipt. New-history validation and HTML rendering also passed.
At published head 643b13985aa097cf259d1e0804fc59b0c01d6cad, all five canonical gene-file blobs matched before editing. Formal review 5329963687 and full comment 5855089062 were read. The proposed downgrade of glutamate metabolism was independently assessed against the live ontology and positive human enzyme evidence.
The live GABA-shunt definition explicitly places SSADH at the final SSA-to-succinate step of the glutamate-to-GABA-to-succinate pathway and lists glutamate metabolic process and succinate metabolic process as is_a parents. ALDH5A1 performs that step; it is not merely required for the pathway through an indirect disease effect. Accordingly, the existing glutamate-process ACCEPT and NO_FAILURE_CORE remain appropriate, without asserting that glutamate is the immediate substrate. The integrated core is a compact synthesis of that catalytic work, not an exhaustive list of all valid pathway labels. Its GABA-catabolism description therefore does not negate the broader core-process annotation. The annotation-reviewer peer and parent independently agreed with this participation distinction. Succinate-process acceptance also remains grounded in actual product formation. All 25 decisions and source assertions remain unchanged: 22 ACCEPT, two MODIFY and one KEEP_AS_NON_CORE; no NEW assertion.
Live GO:0004777 directly confirms GO:0009013 as a parent, so the cofactor-specific refinement and parent wording stand. An exact cached UniProt reaction/EC/RHEA excerpt now accompanies the electronic reaction row. The same record and live GO reaction participants support explicit core substrate entries for 4-oxobutanoate (succinate semialdehyde) and NAD(1-); these are the measured substrates in PMID:16199352, not additional activities. The standalone description now includes aldehyde-dehydrogenase family membership and mitochondrial targeting; UniProt's 1–47 transit-peptide boundary is predictive (ECO:0000255) and no newly measured cleavage site is claimed. Redox inhibition remains biological prose without assay-workflow framing.
The earlier Nature URL is corrected to ng727, matching the unchanged normal PMID:11544478 record. The provenance summary now distinguishes the IBA ancestral-node entries from traced ortholog/mapping sources. Published history remains untouched. The public institutional PMID:9683595 clinical excerpt remains documented by its existing primary-PDF receipt and page 407; its external-public provenance now appears in the reason instead of the field reserved for nonshareable full text. The developmental judgment and the abstract-only local availability flag are unchanged.
No new source was introduced. The recursive authored/provider DOI, PMID, PDF and link census retains the same 13 required cached publications and reviewed Reactome R-HSA-888548. Bare unused UniProt bibliography entries remain outside the required functional-source set. There are no provider artifacts, new retrievals, source edits or cache gates in this follow-up.
Targeted validation completed without review warnings; history validation and rendering passed. All 47 cached supporting excerpts are exact after case-sensitive whitespace normalization. Integrity checks preserve all 25 source rows/actions, both isoforms, all 20 reference objects and immutable raw-source bytes. The bounded YAML delta changes only the description, three annotation-review evidence/rationale blocks and two core substrate entries. No new source gate was introduced.