CG31606 / Q8IPG8

Q8IPG8 is the 197-residue CG31606-PB product linked to FBgn0051606 and RefSeq NP_723348.1. The completed Falcon report was inspected and found no focused functional literature; its uncontextualized expression-table lead does not establish a mechanism. Unlike the report’s supplied prompt, the full selected sequence record contains a SignalP-derived signal peptide. The cleaved N-terminus and soluble mature sequence support a reasonable extracellular-location inference.

Donor provenance assessment identifies P0DTS2 as northern elephant seal ApoA-V and Q55286 as Shigella IpgD. Partial structural matching is not equivalent to orthology or biochemical specificity. Three lipid claims remain UNC; extracellular region is COR from sequence topology, not from the bacterial donor. All three original ND GOA rows are preserved and reviewed. No core molecular function can be specified.

Sequence and research provenance

The selected frozen UniProt record is retained in CG31606-uniprot-source.json, with the complete original prediction metadata in CG31606-predictions-source.json. It maps the exact accession to this FlyBase gene; current sequence identity does not establish the historical predictor input. Gene-level experiments support conserved functions unless a relevant isoform difference is established. Falcon research was requested with perplexity-lite fallback alongside publication caching. Primary publications and sequence observations, rather than AI summaries or ARBA assertions, support the curated conclusions.

2026-10-20 — Full re-review, independent of the ProtNLM2 predictions

This re-review did not use ProtNLM2 output or prediction-donor reasoning. The evidence is the UniProt sequence record, the Falcon literature report, and a new reproducible homology analysis (rereview/RESULTS-rereview.md, script rereview/homology_rereview.py, raw rereview/homology_rereview.json).

• Identity: Q8IPG8 (197 aa) is the only UniProtKB entry for FBgn0051606, and OMA marks it as the main isoform. Two cDNA fragments from the same locus (G2J5W9, 198 aa; D9PTT6, 142 aa) are identical over their aligned length. The benchmark accession is the canonical full-length product, not a fragment, isoform or alternative gene model.
• Family: InterPro has 0 entries and UniProt lists no PANTHER family, so there is no PAINT/IBA node. The only feature is the SignalP signal peptide (1-18).
• Phylogeny: The OMA HOG is restricted to the melanogaster group, with 1:1 orthologs in sec/sim/ere/yak/ana/bip/suz. blastp found no Swiss-Prot hit, and every UniProtKB Eukaryota hit is from Drosophila. The only paralog-level signal is marginal (E ~ 1e-4) partial similarity to CG42763 and TsetseEP-domain-labelled ana/yak proteins. CG31606 is a lineage-restricted orphan, so no function can be transferred. The "vertebrate homolog: Yes" flag in the Tran 2010 expression table (via Falcon) is not supported.
• Annotations: MF ND is ACCEPTED and BP ND is ACCEPTED. CC ND is MODIFIED to GO:0005576 extracellular region, based on the signal peptide, the soluble mature chain and the absence of a KDEL/HDEL motif. This is a sequence inference, not observed localisation. core_functions is left empty on purpose because no activity is known. A BIOLOGY knowledge gap (WHOLLY_DARK) was added.
• Relative: No characterised ortholog or PANTHER subfamily exists. The nearest detectable relative is the uncharacterised D. melanogaster paralog CG42763 (M9MRG9), which has no review in the repo. It is too weakly related and too poorly characterised to be a useful reference.
• Modules / PANTHER family review: None apply (no PANTHER family, no hits in modules/ or gocams/index.tsv).