col1a1b notes

2026-09-28 (claude-opus-5-5, DANRE_DUPLICATION pair review col1a1a/col1a1b)

• Deep research failed. Edison returned 402 Payment Required and the OpenAI key is invalid. It was not retried. The literature research was done by hand (Europe PMC REST search plus cached publications).
• Accession Q6PEI9 (1449 aa, TrEMBL, "Collagen, type I, alpha 3", synonym col1a3, RefSeq NP_958886). It holds all 9 GOA rows. There are no GOA rows from PMID:30082390 or PMID:26876635, although both study col1a1b directly.
• Protein: 78% identical to alpha1(I) PMID:26876635. The pair script gives 76.3% (global BLOSUM62). One inter-chain C-propeptide Cys is lost PMID:26876635. Confirmed on the UniProt sequences as col1a1a C1265 vs col1a1b S1267 (col1a1b-bioinformatics/RESULTS.md). Gly-Gly content is higher PMID:26876635.
• Protein-level presence: PMID:26876635. The chains are roughly 1:1:1 in adult bone, skin and scales, with alpha3/alpha1 higher in skin and scales PMID:26876635.
• Genetics (PMID:30082390): the sa12931 nonsense null is viable, with no alpha3(I) peptides PMID:30082390. This is a PTC/NMD allele, and paralog upregulation (transcriptional adaptation) was not tested. The double heterozygote with col1a1a is fragile. The dominant dmh29 allele is severe.
• Comparative: alpha3 is present in skin collagen of most teleosts but absent from some PMID:3677606. It is reported in sturgeon PMID:1617936; this is my caveat, since sturgeons have their own polyploidy. It is the fastest-evolving chain PMID:34430038. In trout, skin collagen is a heterotrimer and muscle collagen is (alpha1)2alpha2 (PMID:11358497).
• Decisions:
• skeletal system development IMP (dmh29): ACCEPT.
• ECM structural constituent conferring tensile strength (IBA), collagen fibril organization (IBA), ECM (IBA), and the IEA rows: ACCEPT.
• cytoplasm IBA: REMOVE. Its sole source is the col1a1a mRNA-based IDA.
• skin development and skeletal system morphogenesis (IBA): non-core, the same as col1a1a.
• NEW GO:0005584 collagen type I trimer (IDA, PMID:26876635), the same as col1a1a.
• No col1a1b-specific process term. Nothing col1a1b does is absent from col1a1a.
• Pair page: projects/DANRE_DUPLICATION/pairs/col1a1a_col1a1b/col1a1a_col1a1b.md