The naked mole rat

Testing ortholog-projected GO annotations against a species that diverged

AI Gene Review · projects/NAKED_MOLE_RAT · 2026

Bottom line

  • Of ~335,000 GO annotations on naked mole rat proteins, exactly one is experimental; the rest are projected from orthologs or sequence models.
  • We reviewed all 169 existing annotations on 8 landmark genes (180 rows with the 11 NEW) against the species' own literature (37 papers).
  • Projection is mostly correct but unfocused: 4 removed, 84 demoted, 11 new. The removals include a sign reversal in Cgas and a Hyal2 term propagated past a curated NOT.

Why this species

  • Long-lived, cancer-resistant, insensitive to some pain stimuli, hypoxia-tolerant.
  • Only six HETGA entries are Swiss-Prot reviewed; everything else is TrEMBL.
  • An ortholog projection is a hypothesis. Here it is testable, because the literature exists because the proteins differ.
  • Two stories chosen for batch 1:
    • Hyaluronan axis: Has2, Hyal2, Cd44 (very-high-molecular-mass hyaluronan)
    • Nociception: Scn9a, Ntrk1, Trpv1, Tac1
    • plus Cgas, all 13 original rows from one PANTHER node

Three ways a projection resolves

Where propagation fails

What a review row looks like

Cgas review page: the ancestral innate-immune signalling row is accepted, with the PANTHER node and both seeds checked. The homologous-recombination row is removed further down.

New terms GO cannot yet express

Gene Proposed term
Has2 high molecular mass hyaluronan biosynthetic process
Hyal2 regulation of hyaluronan polymer size
Scn9a proton-inhibited voltage-gated sodium channel activity
Trpv1 vanilloid-gated monoatomic cation channel activity
Cd44 positive regulation of basal ATF6-mediated signalling

Polymer length as the functional property; a channel closed by protons; a sensor's resting set-point.

An expressivity gap

  • Has2: the naked mole rat sequence in HEK293 cells makes high-molecular-mass hyaluronan; shRNA knockdown in naked mole rat fibroblasts reduces it.
  • That is direct and mutant-phenotype evidence for terms carried only as ISS.
  • The validator rejects action: NEW for any term already in GOA, so "right term, evidence code understates what is known" cannot be recorded.
  • For an all-electronic species this is the most valuable recommendation a review could make.

Status and next steps

  • ✅ Batch 1: 8/8 reviews, 180 rows, none pending; TreeGrafter rows re-reviewed Sept 2026.
  • ⬜ No further batches started.
  • ⬜ A way to propose an evidence-code upgrade for an existing term.
  • Tip: the decisive paper is often titled for another gene (e.g. TMEM2 for hyaluronan degradation).

Read more: projects/NAKED_MOLE_RAT.md · genes/HETGA/<Gene>/