Normal fetch seeded the original GOA assertions; all source fields are retained. Reviewed the full primary Carassius study PMID:28801642(https://pubmed.ncbi.nlm.nih.gov/28801642/) and the source record. The record-to-publication sequence linkage was checked against the ENA accessions KF960826 (E1alpha3), KF960828 (E1beta2, partial CDS) or JX975104 (ADH8a3). Source files were not edited.
The muscle pathway has strong physiological support, while the individual paralog enzyme assignment uses expression, sequence, localization and structural modeling. No purified alpha3/beta2 reconstitution or ADH8a3 substrate panel is claimed. Conserved PDH-domain membership does not settle coupling to E2 after beta2 interface divergence. The retained conventional-PDH claims therefore remain UNDECIDED rather than being removed. ADH8a comparisons use PMID:15231826(https://pubmed.ncbi.nlm.nih.gov/15231826/) as homolog evidence, not as a direct carp assay. No new process rows were proposed.
Falcon and fallback provider attempts failed; manual primary research was used and no provider report was fabricated. GOA contains only GO_REF sources, so publication caching from GOA yielded no target paper; relevant primaries were fetched separately. No matching cached GO-CAM activity was found for the three accessions. Exact family labels are reported separately for module integration; no unverified PTN origin is asserted.
Independent annotation-reviewer follow-up: distinguish E1alpha antibody localization from alpha3-specific evidence and general cytosolic ADH background from target localization. Zebrafish ADH8A had no detectable HMGSH activity in the published assay, not low measured activity; target carp specificity remains unresolved.
The NON_TER sequence feature describes the submitted UniProt sequence, not demonstrated truncation of the native protein. The paper reports full-length beta2 RACE; the discrepancy remains documented rather than silently repairing the downloaded record.