GOA changes (refresh in commit a3cf70b6d):
- Two new ISO rows split by donor identifier: GO:0005739 mitochondrion and GO:0005777 peroxisome, each with WITH/FROM UniProtKB:Q9UHK6|UniProtKB:Q9UHK6-1 (human AMACR, protein plus canonical isoform). Protein-level siblings already exist.
- No rows retired.
PENDING resolved:
- Both rows -> KEEP_AS_NON_CORE, each with its own propagation_review naming the isoform donor. The human enzyme is dual-targeted PMID:11060344, and rat has direct fractionation data [PMID:11060359 "Subcellular fractionation experiments, however, revealed that both in humans and rats alpha-methylacyl-CoA racemase is bimodally distributed to both the peroxisome and the mitochondrion."; PMID:7649182 "the rat enzyme co-distributed exclusively with mitochondrial marker enzymes"].
Actions changed:
- GO:0005102 signaling receptor binding (ISO, human AMACR): MARK_AS_OVER_ANNOTATED -> REMOVE. The only receptor interaction reported for AMACR is PTS1 recognition by the peroxisomal import receptor Pex5p PMID:11060344. Pex5p is a cargo-import receptor, not a signaling receptor in the sense of the GO:0005102 definition. QuickGO (2026-10-04) returns no GO:0005102 annotation for Q9UHK6 or Q9UHK6-1, so the donor appears to have been withdrawn (propagation_review: SOURCE_STALE_OR_MISSING).
Support strengthened (actions unchanged): EXP/IDA rows that cited only paper titles now also carry claim-bearing abstract quotes:
- PMID:9106621 (R)-ibuprofenoyl-CoA inversion assay
- PMID:7649182 rat mitochondrial co-distribution
- PMID:11060359 bimodal distribution
- PMID:11964182 racemase step in cholic acid synthesis
- PMID:8020470 trihydroxycoprostanoyl-CoA substrate
Description: removed the curation commentary ("The review accepts..."). Added the biology: (2S)-requirement of beta-oxidation, dual targeting, and the ibuprofen epimerase role.
Open questions:
- PMID:14561759 (rat liver peroxisome proteomics) supports the peroxisome IDA only through the curator's reading of the full text; the abstract does not name Amacr.
- core_functions lists both GO:0008206 bile acid metabolic process and its descendant GO:0006699 bile acid biosynthetic process. This is redundant but harmless; left as is.
Stale UniProt quotes (2026-10-10): replaced 15 UniProtKB:P70473 supporting_text quotes (paraphrased FUNCTION text no longer in the refreshed flat file) with verbatim FUNCTION and SUBCELLULAR LOCATION lines from the current Amacr-uniprot.txt, matched to each row (location rows quote SUBCELLULAR LOCATION, activity/process rows quote FUNCTION).
Review follow-up 2026-10-17 (PR #4617 reviewer findings): Dropped GO:0008206 bile acid metabolic process from core_functions directly_involved_in because it is the ancestor of the GO:0006699 bile acid biosynthetic process listed with it. CC convention for this batch: a location row is ACCEPTed only when that location is part of a core_functions entry (Pnlip extracellular region for luminal digestion, Arsb lysosome for GAG degradation); otherwise location rows are KEEP_AS_NON_CORE. Batch policy for broad parents (applied to all eight genes in PR #4617): an uninformative broad parent of a specific activity or process the gene already carries is MODIFIED to that specific term; experimental rows stay at the curator's breadth unless the cited assay itself identifies the specific substrate. (Amacr's GO:0003824 MODIFY already followed it.)