Focused function hypothesis
Hypothesis: The horse protein F6S899 catalyzes NAD-dependent lysine desuccinylation.
Target: Equus caballus (NCBITaxon:9796), UniProt F6S899. Gene label: SIRT5; verify identity independently rather than treating the label as proof.
Decisive question
Evaluate the structural and cofactor requirements for this single catalytic activity in the exact sequence. Compare characterized mammalian proteins and relevant public structures. Distinguish a functional target protein from a family label or a different transcript product.
Identity and sequence inputs
- Target record: https://www.uniprot.org/uniprotkb/F6S899/entry
- Human comparison lead: https://www.uniprot.org/uniprotkb/Q9NXA8/entry (SIRT5). Establish the relevant orthology/isoform relationship rather than assuming it.
- Frozen current UniProt sequence: 282 residues; SHA-256
61bce25d76191fa8ce4d18a700c25beab8f0bcb857f8af2f8eb343b76bfc0fd0. - These are current sequences downloaded for the cohort on 2026-09-08. Identity with the original prediction-time input has not been established. Evaluate the supplied sequence explicitly; document any different sequence used.
>F6S899 Equus caballus SIRT5
MRPLQIVHSRLISRLCCGLKSAASTQTKICLTMARPSSNMADFRKFFAKAKHIVVISGAG
ISAESGVPTFRGAGGYWRKWKAQDLATPQAFARNPSQVWEFYHYRREVVQTKEPNPGHLA
IAQCEARLHKQGRRVVVITQNIDELHRKAGTKNLLEIHGSLFKTRCTSCGVVAENYKSPI
CPALSGKGSPDPETQSARIPAENLPRWEHPLWSILPPCLPPRCLPGEFQWPNSTWKPPQP
QADSGFISRGPVVRLFLKPLLTKPKLFLNCPGEGRNYSASVY
Evidence and deliverable
Use primary literature and public sequence, structural and genomic resources. Select analyses that answer the decisive question; this is not a general gene review. Assess support and contrary evidence, and allow an unresolved outcome. Distinguish directly observed horse evidence, justified mammalian transfer, and results for a different protein model.
Do not consult the ai-gene-review repository's existing judgments, research syntheses or local bioinformatics analyses. Those are held out for comparison. Do not use agreement with ARBA or another prediction as biological validation. Preserve reproducible methods, accessions/versions, actual computation outputs and primary-source URLs/DOIs/PMIDs. Report the decisive findings and limitations, not just a verdict.