Neuron or glia?

Reviewing GO annotations for the genes that decide neural cell fate

AI Gene Review · projects/NEURON_DEVELOPMENT · 2026

Bottom line

  • GO describes fate choice with specification → determination → commitment terms from classical embryology; we asked whether the evidence behind each annotation supports the rung it claims.
  • 13 of 38 planned human genes reviewed (all 12 master regulators + DLX1): 1,652 annotations, 960 accepted, 138 removed.
  • Fate-stage terms mostly held up; the removals were almost all generic protein binding (129 of 138). Priorities 2–4 (25 genes) are not started.

The decision being annotated

Why these terms are a curation risk

  • Telling specification from commitment needs transplant or neutral-environment assays, which most papers lack.
  • Single-cell studies (Velten 2017, Weinreb 2020) show continuous trajectories, not discrete stages.
  • Some cell-type commitment terms were added on request for one gene, e.g. GO:0072154 proximal convoluted tubule segment 1 cell fate commitment.
  • Neuron vs glia is the clearest binary choice (Notch lateral inhibition), so it is the right test case.

What the reviews decided on the fate ladder

Review actions

What a review looks like

ASCL1 review page: each GOA row gets an action and a written reason with quoted evidence.

Gene-level findings

  • HES1 is a repressor: projected positive regulation of transcription rows marked over-annotated.
  • ASCL1: negative regulation of glial cell differentiation (GO:0045686) proposed as NEW.
  • OLIG2: specification accepted; oligodendrocyte cell fate commitment (GO:0021779) added as NEW.
  • NEUROD1: endocrine pancreas development accepted alongside neuronal roles; insulin-secretion and glucose rows kept as non-core.
  • STAT3 (456 rows) and NOTCH1 (378 rows): 66 and 33 removals, all generic protein binding.

Status and next steps

  • ✅ Priority 1 (12 master regulators) and DLX1 reviewed; all validate.
  • ⬜ Priority 2 subtype genes: DLX2, LHX6, TBR1, NR4A2, PITX3, ISL1, MNX1, OLIG1, SOX10.
  • ⬜ Priority 3–4 signalling and oligodendrocyte genes (SHH already reviewed under the cerebellum work).
  • ⬜ Decide whether granular commitment terms should give way to general terms plus Cell Ontology extensions.

Read more: projects/NEURON_DEVELOPMENT.md · genes/human/<GENE>/