zic1 (XENLA, UniProt O73689) — Hypothesis Review

Focus: proposed GO term "neural plate border formation"; seed = pax3+zic1 border co-expression (scRNA)

Executive Judgment

Verdict: Supported (core claim) / Partially supported (specific details), with important dynamic qualifiers.

The seed hypothesis's central biology — that at the Xenopus neural plate border (NPB) pax3 and zic1 are
co-expressed in the same cells, that the pax3+zic1 double-positive population is the one that progresses to
neural crest (snai2/sox10/foxd3), and that zic1-only cells adopt preplacodal/anterior-neural fate — is
strongly and directly supported by causal, functional primary literature in Xenopus (PMID:15843410;
PMID:17409353). This is stronger than the correlative single-cell co-expression test the seed proposes.

Importantly, the specific single-cell experiment the seed asks for has effectively already been performed in
the correct organism at higher spatial fidelity by quantitative HCR in situ hybridization (Montequin &
LaBonne, PMID:41718037 / preprint PMID:41256729). That work confirms co-expression and the double-positive→
crest logic but qualifies the static snapshot model: pax3/zic1 onset is broad, non-uniform and partially
overlapping; relative (not merely binary) pax3:zic1 levels tune crest-gene choice (snai2 vs sox8); and at later
stages NPB factors are inversely correlated with crest identity because pax3/zic1 are downregulated as crest
cells commit. A naïve "double-positive cells = crest" readout from a whole-embryo scRNA atlas (Briggs et al.
2018) is therefore liable to stage-dependent misinterpretation.

For curation: the underlying gene function (zic1 is a core NPB transcription factor whose combinatorial activity
specifies NPB derivatives) is well supported. A live GO/QuickGO check (this iteration) shows this biology is
already annotated — zic1 (O73689) already carries GO:0014034 neural crest cell fate commitment (IMP),
GO:0014033 neural crest cell differentiation (IMP), GO:0001840 neural plate development (IMP) and GO:0014029
neural crest formation (IEA). There is no dedicated "neural plate border" GO term, but GO:0014029's own
definition literally describes formation of the border region. So the proposed new term "neural plate border
formation" is biologically apt yet largely redundant with an existing term's definition; its real value is
as an ontology-structure refinement (splitting border-territory formation from crest formation), not as a
missing gene-level annotation. See GO Curation Implications.

Evidence Matrix

GO Curation Implications (leads — require curator verification)

Verified against the live GO ontology (EBI OLS) and QuickGO on 2026-10-09 — direct queries, see Provenance.

Key ontology finding: GO has no dedicated "neural plate border" term (searches for "neural plate
border", "…specification", "…formation" return none). The closest existing terms are GO:0014029 neural crest
formation, the neural-plate-regionalization family (GO:0060896 neural plate pattern specification; GO:0021998
mediolateral; GO:0021999 A/P; GO:0060897 regionalization), and GO:0021502 neural fold elevation formation.
Critically, GO:0014029's own definition already is the border: "The formation of the specialized region of
ectoderm between the neural ectoderm (neural plate) and non-neural ectoderm. The neural crest gives rise to the
neural crest cells that migrate away from this region…" So the existing term labelled "neural crest formation"
is biologically a neural-plate-border-formation term.

Key annotation finding: Xenopus zic1 (O73689) is already annotated to the relevant processes, several
with experimental (IMP) evidence:

Curation decision table (leads — require curator verification):

• Aspect: this is a Biological Process matter (plus the already-annotated MF transcription-factor terms); not a CC claim.
• Avoid "protein binding" as the curation outcome — specific developmental BP terms are well supported.
• Recommended qualifiers if any annotation is added/upgraded: experimental evidence (IMP/IGI) from
  PMID:15843410 and PMID:17409353; taxon = Xenopus; annotate with awareness that the function is
  combinatorial with pax3 and Wnt-dependent, and that zic1 alone biases to preplacodal fate.
• Bottom line for the curator: the seed's underlying biology is real and already represented in zic1's GO
  record (GO:0014034/GO:0014033 IMP; GO:0014029 IEA). The proposed new term adds value mainly as an ontology
  refinement (separating border formation from crest formation), not as a missing gene annotation.

Mechanistic Scope

• Immediate molecular function: zic1 is a zinc-finger (ZIC/GLI family) transcription factor acting at the
  NPB. Its direct, cell-autonomous activity is transcriptional regulation of border/fate genes.
• Direct cellular process: specification of neural-plate-border cell fates — with pax3 (+ Wnt) it specifies
  neural crest; alone it biases toward preplacodal ectoderm.
• Downstream / not the direct activity being annotated: overt neural crest EMT/migration, craniofacial and
  placodal derivatives, cerebellar patterning, prdm12 induction — these are downstream consequences, not the
  immediate molecular function. Curation of the proposed term should rest on the specification activity, not on
  the distal phenotypes.

Conflicts and Alternatives

1. Dynamics vs static snapshot (most important). The seed frames double-positive cells as "the" crest
   progenitors across stages. Montequin & LaBonne (PMID:41718037) show pax3/zic1 are downregulated as crest
   identity consolidates, producing an inverse NPB-vs-crest correlation at later stages. A Briggs-atlas
   per-cell co-expression readout could therefore under-count committed crest cells at later stages and should
   be stage-stratified.
2. zic1 is not crest-specific. zic1 is required for preplacodal fate too (PMID:17409353; PMID:30012125).
   Treating zic1 primarily as a crest gene would be too narrow.
3. Paralog claim now supported at the domain level (Iteration 3). The seed asserts zic2, zic3, zic4, zic5 are
   also in the pax3+zic1+ border population. zic2 (PMID:32891623), zic3 (PMID:40939754), and critically zic4
   (PMID:16871625, "neural plate border, dorsal neural tube… similar to zic1") are documented at the Xenopus
   border; all five zic paralogs co-regulate neural/crest development (PMID:16871625). Caveat: profiles are
   "significantly diverged" with two functional groups (zic1/2/3 vs zic4/5), so strict per-cell co-expression
   of all five in the identical double-positive cells is an over-simplification not formally shown at single-cell
   resolution. Potential paralog cross-reactivity/mis-mapping is a real caveat for any probe- or read-mapping-
   based quantification (X. laevis is allotetraploid with L/S homeologs; zic paralogs are closely related and
   zic1/zic4 are chromosomally adjacent).
4. Organism specificity. The combinatorial Pax3+Zic1 crest model is best established in Xenopus; its
   mammalian conservation is explicitly "open to question" (PMID:34638777). This matters only if the annotation
   is projected beyond Xenopus.

Knowledge Gaps

• Briggs 2018 (PMID:29700227) per-cell quantification not run here. The whole-embryo X. tropicalis scRNA
  atlas was not downloaded/analyzed in this review (dataset retrieval/compute not performed; not necessary given
  the superior targeted Xenopus HCR data already available). Gap matters because the seed specifically names it;
  resolving it would require loading the atlas, identifying NPB/early-crest clusters, and computing stage-
  stratified per-cell co-expression of pax3, zic1–5 and snai2/sox10/foxd3 — explicitly separating early
  (double-positive→crest) from late (inverse-correlation) stages.
• zic paralog claim — RESOLVED at domain level (Iteration 3). zic4 is explicitly a neural-plate-border gene
  like zic1 (PMID:16871625); zic2 (PMID:32891623), zic3 (PMID:40939754) documented at the border; all five zic
  genes co-regulate neural/crest development. Remaining gap: strict per-cell co-expression of all five (esp.
  zic5) in the identical double-positive cells is not shown at single-cell resolution, and profiles are
  "significantly diverged."
• Whether GO already contains a suitable term — RESOLVED this iteration (direct OLS/QuickGO query): no
  dedicated neural-plate-border term exists, but GO:0014029's definition already describes border formation, and
  zic1 is already annotated to GO:0014029/GO:0014034/GO:0014033. The remaining gap is an ontology-design
  question (should "border formation" be split from "crest formation"?) for the GO consortium, not a missing
  gene annotation.

Discriminating Tests

1. Re-analyze Briggs et al. 2018 (PMID:29700227) stage-by-stage: per-cell co-expression of pax3 × zic1–5 vs
   snai2/sox10/foxd3, testing the prediction that double-positive fraction co-expressing crest specifiers rises
   early then falls (inverse correlation) as in PMID:41718037.
2. Lineage tracing / live imaging of pax3+zic1+ cells to confirm they are the cells that become crest (co-
   expression alone cannot establish progression).
3. Quantify relative pax3:zic1 ratio per cell against snai2 vs sox8 to test the dose-tuning model.
4. Paralog-resolved probes/reads (handle L/S homeologs and zic2–5) to test the multi-zic co-expression claim.

Curation Leads (require curator verification)

• References to attach: PMID:15843410 and PMID:17409353 (primary functional support for the border-
  specification/combinatorial model); PMID:41718037 (direct quantitative single-cell-resolution test + dynamic
  qualifier); supporting: PMID:30012125, PMID:40939754, PMID:32891623.
• Candidate GO action: support a BP annotation for zic1 at the neural plate border. Prefer an existing GO BP
  term if one covers "neural plate border / neural crest specification"; if the proposed new term "neural plate
  border formation" is adopted, consider wording it as border cell-fate specification and apply combinatorial/
  contributes_to framing (zic1 + pax3 + Wnt), not sole-driver.
• Snippet to verify (PMID:17409353): "At the end of gastrulation, Pax3 and Zic1 are coexpressed in the
  neural crest forming region ... Zic1 is detected in the preplacodal ectoderm ... their combined activity is
  essential to specify the neural crest."
• Snippet to verify (PMID:41718037): "later stages display an inverse correlation between neural crest and
  neural plate border factors, suggesting that pax3 and zic1 initially promote neural crest gene activation but
  are downregulated as neural crest identity emerges."
• Suggested curator questions: (a) Does GO already have a neural-plate-border term? (b) Should the annotation
  be crest-specific or border-level (recommend border-level for accuracy)? (c) Is the combinatorial/Wnt-dependent
  nature captured? (d) Keep annotation taxon-scoped to Xenopus given mammalian uncertainty.
• Suggested experiment: the stage-stratified Briggs re-analysis above, to directly close the seed's named gap.

Provenance (analyses actually run)

• GO ontology check (EBI OLS4 API, 2026-10-09): searched GO for "neural plate border", "neural plate border
  specification/formation", "neural crest formation", "neural plate pattern", "roof plate". Result: no dedicated
  neural-plate-border term; retrieved definitions of GO:0014029, GO:0060896, GO:0021998/0021999/0060897,
  GO:0021502. GO:0014029 definition verbatim = "The formation of the specialized region of ectoderm between the
  neural ectoderm (neural plate) and non-neural ectoderm…".
• Existing-annotation check (QuickGO API): retrieved BP annotations for zic1 O73689 (22 BP terms incl.
  GO:0014029 IEA, GO:0014034 IMP, GO:0014033 IMP, GO:0001840 IMP, GO:0021904 IMP, GO:0030177 IGI, GO:0045944/
  0045893 IMP) and human ZIC1 Q15915 (20 BP terms) for cross-species comparison.
• These queries are reproducible with the requests calls shown in the executed code cells this iteration. No
  values were hand-entered; term labels were resolved programmatically from OLS.

Limitations

This review is literature- and reasoning-based (plus the live GO/QuickGO queries above). The named scRNA atlas (Briggs 2018) was not computationally
re-analyzed in this run; conclusions rest on causal functional studies and a direct quantitative in-situ study
in Xenopus that supersede a pure co-expression correlation. No results were fabricated; where a check was not
performed it is stated plainly.

Artifacts

• OpenScientist final report
• OpenScientist final report