NTR: crypt top colonocyte (CL:9900000) — fixes #3776 by ai4c-agent[bot] · Pull Request #3777 · obophenotype/cell-ontology

ai4c-agent · 2026-10-02T13:47:19Z

Summary

Adds a new term for crypt top colonocyte as requested in #3776.

Label: crypt top colonocyte
Temporary ID: CL:9900000 (from the Temporary IDs idrange 9900000–9999999; expected to be replaced by a definitive ID during the standard reserialisation step)
Parent: colonocyte (CL:1000347) via SubClassOf
Definition: "A colonocyte of the human colon located at the crypt top, representing the most differentiated, mature stage of the colonocyte lineage. Distinguished by high expression of the water channel AQP8 together with CEACAM5, CEACAM7, FABP1, TFF1, SLC26A3 and GUCA2A, consistent with peak absorptive and secretory capacity at the luminal surface. It sits at the opposite end of the crypt-villus differentiation axis from early colonocyte (CL:4047018), the earliest post-mitotic stage of the same lineage, and is distinct from BEST4+ colonocyte (CL:4047052), a separate ion-channel-expressing subtype scattered through the colon epithelium." [PMID:35530046]
Related synonyms: late colonocyte, surface colonocyte (both xref PMID:35530046)
term_tracker_item: CL new term request: crypt top colonocyte #3776

Notes for the reviewer

Overlap with PR NTR: crypt top colonocyte (closes #3756) #3757 (closes [NTR] crypt top colonocyte #3756). PR NTR: crypt top colonocyte (closes #3756) #3757 adds the same term (crypt top colonocyte) from the parallel request [NTR] crypt top colonocyte #3756 with a shorter definition. This PR uses the richer wording supplied in CL new term request: crypt top colonocyte #3776, which explicitly contrasts the new term against its sibling colonocyte subtypes (CL:4047018 early colonocyte, CL:4047052 BEST4+ colonocyte). Please pick one PR, or ask me to merge the two definitions into NTR: crypt top colonocyte (closes #3756) #3757; I'll close the other.
Anatomical placement. Kept implicit via the colonocyte parent (already part_of UBERON:0000397 colon). No separate part_of axiom to a "crypt top of colon" UBERON class was added — I could not identify a specific UBERON term for the colonic crypt-top / surface epithelium region without fabricating one. Happy to amend if the reviewer can point to the right UBERON class.
Reference. No PMID was given in CL new term request: crypt top colonocyte #3776; I reused PMID:35530046 from the sister request [NTR] crypt top colonocyte #3756/PR NTR: crypt top colonocyte (closes #3756) #3757, which the previous reviewer appears to have accepted as the HGCA v1 reference for this term. If a different reference is preferred, let me know.
ORCID. No ORCID was provided in CL new term request: crypt top colonocyte #3776, so no terms:contributor was added; terms:creator "Claude" records me as the author of the term stanza.

Test plan

Reviewer confirms the richer #3776 definition is preferred over the #3757 version (or asks for a merge).
robot reason on the ODK pipeline classifies the new term as expected.
Temporary CL:9900000 ID is replaced with a definitive ID during reserialisation.

Closes #3776.

🤖 Generated by @ai4c-agent

Model: claude-opus-4-7
Agent harness: claude-code
Triggered by: @ct31-sanger
Run: View workflow run

🤖 Generated with Claude Code

Adds the mature, terminally differentiated crypt-top colonocyte of the human colon, as requested in issue #3776. - Label: crypt top colonocyte - Parent: SubClassOf colonocyte (CL:1000347) - Definition sourced from the NTR text in issue #3776 (CAP / HGCA v1 curation), citing PMID:35530046 (also used for the parallel request in issue #3756 / PR #3757). Mentions the lineage relationship to early colonocyte (CL:4047018) and the distinction from BEST4+ colonocyte (CL:4047052). - Related synonyms: "late colonocyte", "surface colonocyte" (xref PMID:35530046). - term_tracker_item: #3776 - Temporary CL:9900000 ID (idrange:81 Temporary IDs); expected to be replaced with a definitive ID at the next reserialisation. Note: there is an existing open PR #3757 adding the same term from sister issue #3756 with a shorter definition. This PR uses the richer definition supplied in #3776; the reviewer should pick one or merge the two definitions and close the other request. Validated with `robot convert` (no syntax errors). Signed-off-by: Claude Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ct31-sanger · 2026-10-02T13:56:10Z

/clara #gogoeditdiff

github-actions · 2026-10-02T14:02:39Z

Here's a diff of how these changes impact the classified ontology (on -base file):

Ontology comparison

Left

Ontology IRI: http://purl.obolibrary.org/obo/cl/cl-base.owl
Version IRI: http://purl.obolibrary.org/obo/cl/releases/2026-10-02/cl-base.owl
Loaded from: file:/__w/cell-ontology/cell-ontology/src/ontology/cl-base-master.owl/cl-base.owl

Right

Ontology IRI: http://purl.obolibrary.org/obo/cl/cl-base.owl
Version IRI: http://purl.obolibrary.org/obo/cl/releases/2026-10-02/cl-base.owl
Loaded from: file:/__w/cell-ontology/cell-ontology/src/ontology/cl-base-pr.owl/cl-base.owl

Ontology imports

Ontology annotations

crypt top colonocyte `http://purl.obolibrary.org/obo/CL_9900000`

Added

Class: crypt top colonocyte
crypt top colonocyte IAO_0000233 "CL new term request: crypt top colonocyte #3776"
crypt top colonocyte creator "Claude"
crypt top colonocyte date "2026-10-02T13:40:00Z"^^dateTime
crypt top colonocyte IAO_0000115 "A colonocyte of the human colon located at the crypt top, representing the most differentiated, mature stage of the colonocyte lineage. Distinguished by high expression of the water channel AQP8 together with CEACAM5, CEACAM7, FABP1, TFF1, SLC26A3 and GUCA2A, consistent with peak absorptive and secretory capacity at the luminal surface. It sits at the opposite end of the crypt-villus differentiation axis from early colonocyte (CL:4047018), the earliest post-mitotic stage of the same lineage, and is distinct from BEST4+ colonocyte (CL:4047052), a separate ion-channel-expressing subtype scattered through the colon epithelium."
- hasDbXref "PMID:35530046"
crypt top colonocyte hasRelatedSynonym "late colonocyte"
- hasDbXref "PMID:35530046"
crypt top colonocyte hasRelatedSynonym "surface colonocyte"
- hasDbXref "PMID:35530046"
crypt top colonocyte label "crypt top colonocyte"
crypt top colonocyte SubClassOf colonocyte

github-actions · 2026-10-02T14:02:42Z

Here's a diff of your edit file (unreasoned)

Ontology comparison

Left

Ontology IRI: http://purl.obolibrary.org/obo/cl.owl
Version IRI: None
Loaded from: file:/__w/cell-ontology/cell-ontology/master/src/ontology/cl-edit.owl

Right

Ontology IRI: http://purl.obolibrary.org/obo/cl.owl
Version IRI: None
Loaded from: file:/__w/cell-ontology/cell-ontology/branch/src/ontology/cl-edit.owl

Ontology imports

Ontology annotations

crypt top colonocyte `http://purl.obolibrary.org/obo/CL_9900000`

Added

Class: crypt top colonocyte
crypt top colonocyte term tracker item "CL new term request: crypt top colonocyte #3776"
crypt top colonocyte creator "Claude"
crypt top colonocyte date "2026-10-02T13:40:00Z"^^dateTime
crypt top colonocyte definition "A colonocyte of the human colon located at the crypt top, representing the most differentiated, mature stage of the colonocyte lineage. Distinguished by high expression of the water channel AQP8 together with CEACAM5, CEACAM7, FABP1, TFF1, SLC26A3 and GUCA2A, consistent with peak absorptive and secretory capacity at the luminal surface. It sits at the opposite end of the crypt-villus differentiation axis from early colonocyte (CL:4047018), the earliest post-mitotic stage of the same lineage, and is distinct from BEST4+ colonocyte (CL:4047052), a separate ion-channel-expressing subtype scattered through the colon epithelium."
- database_cross_reference "PMID:35530046"
crypt top colonocyte has_related_synonym "late colonocyte"
- database_cross_reference "PMID:35530046"
crypt top colonocyte has_related_synonym "surface colonocyte"
- database_cross_reference "PMID:35530046"
crypt top colonocyte label "crypt top colonocyte"
crypt top colonocyte SubClassOf colonocyte

ct31-sanger · 2026-10-02T14:39:03Z

/clara

github-actions · 2026-10-02T14:42:10Z

Claude finished @ct31-sanger's task in 5m 23s —— View job

CLARA routed verification — complete

Ran the routed CLARA review per clara_workflow_ref/clara_workflow/agent_instructions.md against routing.json (verification-only; no ontology edits, commits, or pushes).

Todo

Read agent_instructions.md, routing.json, changes.json
Process term group CL:9900000 (targets: ntr:CL:9900000, synonym:CL:9900000:1, synonym:CL:9900000:2)
- Decompose textual/relationship changes into 19 atomic assertions
- Stage B: snippet-search verification against PMID:35530046
- Stage C: full-text fallback for unresolved core assertions
- Wrote runs/CL_9900000/verdicts.json, tool_calls.jsonl, report.md
Post final summary

Result for `crypt top colonocyte` (CL:9900000): needs revision

Reference checked: PMID:35530046 (Serigado et al., Ulcerative Colitis: Novel Epithelial Insights Provided by Single Cell RNA Sequencing) — this is the paper that actually coins "Crypt Top Colonocytes (CTC)".

Core assertions: 16 total — 8 pass, 1 fail, 7 uncertain
Background assertions: 3 total — 1 pass, 2 uncertain

Key finding: the definition's marker panel only partially matches this reference.

The paper's actual CTC/mature-colonocyte marker list is ABCB1, OAS1, ISG15, SLC20A1, CDHR2, CLCA4, MS4A12, SLC9A3R1, SLC26A3/DRA, CDHR5 — only SLC26A3 overlaps with the definition's panel.
AQP8, CEACAM5, CEACAM7, FABP1, TFF1 do not appear anywhere in PMID:35530046 (confirmed via snippet search and full-text check) — unsupported by this specific citation.
GUCA2A is misattributed (fail): the paper assigns GUCA2A/GUCA2B expression to the BEST4+/OTOP2+ cell cluster, not to Crypt Top Colonocytes — and the definition itself explicitly calls BEST4+ colonocyte a separate, distinct subtype. Citing this paper for "crypt top colonocyte expresses GUCA2A" directly contradicts what the paper says.
The synonym "late colonocyte" is not attested anywhere in this reference (only "surface colonocyte" is, via the paper's own phrase "fully differentiated surface colonocytes").
The core identity/location/differentiation-stage claims (is a colonocyte, located at crypt top, most differentiated stage, distinct from BEST4+ colonocyte) are all well supported.

These markers may well be supported by other literature (e.g. a different intestinal single-cell atlas), but as cited, PMID:35530046 does not back most of the marker list and actively contradicts the GUCA2A claim. Recommend the PR author either swap in a reference that actually supports AQP8/CEACAM5/CEACAM7/FABP1/TFF1/GUCA2A, or trim the definition to the markers this paper does support (SLC26A3) plus the structural/comparative claims.

Full assertion-by-assertion detail: runs/CL_9900000/report.md (machine-readable verdicts in runs/CL_9900000/verdicts.json, tool call log in runs/CL_9900000/tool_calls.jsonl).

No source files, routing.json, or changes.json were modified; no commits or pushes were made.

github-actions · 2026-10-02T14:47:51Z

CLARA review for PR #3777

Base SHA: b7cb31d4d3eae2790cb9631fd86f3ee70e1fe963
Head SHA: 0c22f363b05dc5735150814ab4aca60aa18323f9
Terms touched: 1
Total extracted changes: 9
Reviewable changes: 4
Decomposable changes: 1
Routed validation targets: 3
Routed NTR bundles: 1
Routed revised-text targets: 0
Routed ref-only-addition targets: 0
Routed relationship targets: 0
Routed synonym targets with refs: 2

Stage 2/3 verification ran for 1 routed term(s).

Routed verification results

CL:9900000 — crypt top colonocyte: FAIL (core: 16, fail: 1, uncertain: 7, background warnings: 2)
- Failed core assertion: Crypt top colonocyte has high expression of GUCA2A.

Artifacts uploaded:

changes.json
routing.json
runs/

ct31-sanger · 2026-10-05T10:11:25Z

@ai4c-agent please use these edits to amend the definition and references used:

There are two papers (one mouse model and one organ-on-a-chip model) to support GUCA2A expression in crypt top colonocytes: PMID: 31666564 and PMID: 38876106, respectively.
The is strong evidence for high AQP8 expression in crypt top colonocytes with ref: PMID: 38876106
Please remove CEACAM5 marker. There is less consistent evidence to support high expression of this marker in crypt top colonocyte.
Please remove the late colonocyte reference. Keep the synonym.
AQP8, CA1, CA2, SLC26A3 have the strongest, most repeated direct evidence (use ref PMID:22081019).

ai4c-agent · 2026-10-05T10:11:42Z

🤖 Working on it...

Follow along: View workflow run

— @ai4c-agent

…ct31-sanger

…ct31-sanger Updates to the term per feedback on PR #3777: - Remove CEACAM5 from the marker list (less consistent evidence). - Add CA1 and CA2 to the marker list; promote AQP8, CA1, CA2 and SLC26A3 as the markers with the strongest and most consistent evidence (PMID:22081019, Dalerba et al. 2011). - Add supporting references for GUCA2A expression at the crypt top: PMID:31666564 (mouse, Dye et al. 2019) and PMID:38876106 (bioengineered human colon organoids, Mitrofanova et al. 2024). - Add PMID:38876106 as additional support for high AQP8 expression. - Reword the definition so each marker-expression claim is tied to the reference that supports it, addressing the CLARA failure where the original PMID:35530046 did not back the full marker panel. - Remove the PMID:35530046 xref from the "late colonocyte" synonym (kept the synonym itself). Definition cross-references updated to: PMID:22081019, PMID:31666564, PMID:35530046 (retained for structural / crypt-position / sibling-type claims and the "surface colonocyte" synonym), PMID:38876106. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ai4c-agent · 2026-10-05T10:16:07Z

Thanks — amended the term stanza on branch ai4c-agent-issue-3776-run1011 (commit 5f3f213) per all five edits:

Marker panel

Removed CEACAM5 from the definition (less consistent evidence).
Added CA1 and CA2 to the marker list.
Promoted AQP8, CA1, CA2 and SLC26A3 as the markers with the strongest, most consistent evidence of high expression at the crypt top, cited to PMID:22081019 (Dalerba et al. 2011, Nat Biotechnol).
Kept CEACAM7, FABP1, TFF1 and GUCA2A as additional reported markers.

Reference additions

PMID:31666564 (Dye et al. 2019, Sci Rep) — mouse proguanylin / GUCA2A reference.
PMID:38876106 (Mitrofanova et al. 2024, Cell Stem Cell) — bioengineered human colon organoid reference, cited both for GUCA2A at the crypt top and for strong AQP8 expression.
Retained PMID:35530046 (Serigado et al. 2022) on the definition for the structural / crypt-position / sibling-type claims and on the surface colonocyte synonym.

Synonyms

late colonocyte: kept the synonym, removed its PMID:35530046 xref as requested.
surface colonocyte: unchanged (still cites PMID:35530046).

Rewording. Each marker-expression claim is now explicitly tied in-text to the reference that supports it, which should address the CLARA failure where the original PMID:35530046 did not back the full marker list. The structural / comparative claims (crypt-top location, most-differentiated stage, contrast with early colonocyte CL:4047018 and BEST4+ colonocyte CL:4047052) are unchanged.

robot convert -vvv round-trips the edited file cleanly. Happy to re-run /clara once you're ready, or iterate further on wording.

🤖 Generated by @ai4c-agent

Model: claude-opus-4-7
Agent harness: claude-code
Triggered by: @ct31-sanger
Run: View workflow run

ct31-sanger · 2026-10-05T10:22:18Z

/clara

github-actions · 2026-10-05T10:25:03Z

Claude finished @ct31-sanger's task in 5m 43s —— View job

CLARA routed verification — complete

Ran the routed CLARA review per clara_workflow_ref/clara_workflow/agent_instructions.md against routing.json (verification-only; no ontology edits, commits, or pushes), re-checking the term after the 2026-10-05 definition/reference amendment (commit 5f3f213).

Todo

Read agent_instructions.md, routing.json, changes.json
Process term group CL:9900000 (targets: ntr:CL:9900000, synonym:CL:9900000:1)
- Decompose amended definition into 18 atomic assertions (16 core, 2 background)
- Stage B: snippet-search verification against routed refs (PMID:35530046, PMID:38876106, PMID:22081019, PMID:31666564)
- Stage C: full-text fallback for unresolved core assertions
- Wrote runs/CL_9900000/verdicts.json, tool_calls.jsonl, report.md
Post final summary

Result for `crypt top colonocyte` (CL:9900000): no outright core failures; 6 of 16 core assertions uncertain

This is an improvement over the previous run (1 core fail, the GUCA2A/BEST4 misattribution) — that specific problem is now resolved. Two things worth your attention before considering this closed:

CA2 marker (new finding): PMID:22081019 contains internally conflicting statements. One passage lists CA2+ among the bulk "top-of-the-crypt" mature-enterocyte markers (alongside CA1, SLC26A3, AQP8), but a finer SINCE-PCR breakdown in the same paper assigns OLFM4+/CA2-high to an immature progenitor compartment preferentially found at the bottom of the crypt — the opposite pole from the CA1+/SLC26A3+ top-of-crypt population. Can't resolve this from snippets alone; marked uncertain, flagged for your judgement call.
"...absorptive and secretory capacity" claim: PMID:35530046 explicitly frames colonocytes as the absorptive lineage and treats "absorptive" vs. "secretory" as separate cell categories throughout (secretory = goblet/enteroendocrine cells). The absorptive half is well supported; no citation here backs the secretory half for this specific cell type.

Other results:

AQP8, CA1, SLC26A3 — pass, clearly supported by PMID:22081019 (explicit "top-of-the-crypt" marker lists).
GUCA2A — mouse-model half (PMID:31666564) passes clearly; the human-organoid half (PMID:38876106) stays uncertain — get_europepmc_full_text returned no content for this identifier, so it couldn't be checked beyond the abstract.
CEACAM7, FABP1, TFF1 — still uncertain; confirmed absent from the full text of PMID:35530046, no snippet hits in the other three references, and PMID:38876106 full text wasn't retrievable to double-check.
Structural/comparative claims (colonocyte subclass, crypt-top location, most-differentiated stage, distinct from BEST4+ colonocyte, "surface colonocyte" synonym) — all pass, well supported by PMID:35530046.

Full assertion-by-assertion detail: runs/CL_9900000/report.md (machine-readable verdicts in runs/CL_9900000/verdicts.json, tool call log in runs/CL_9900000/tool_calls.jsonl).

No source files, routing.json, or changes.json were modified; no commits or pushes were made.
·

github-actions · 2026-10-05T10:31:06Z

CLARA review for PR #3777

Base SHA: b7cb31d4d3eae2790cb9631fd86f3ee70e1fe963
Head SHA: 5f3f2138b26f76ddc73979f0ab5ee7ddc9c59820
Terms touched: 1
Total extracted changes: 9
Reviewable changes: 4
Decomposable changes: 1
Routed validation targets: 2
Routed NTR bundles: 1
Routed revised-text targets: 0
Routed ref-only-addition targets: 0
Routed relationship targets: 0
Routed synonym targets with refs: 1

Stage 2/3 verification ran for 1 routed term(s).

Routed verification results

CL:9900000 — crypt top colonocyte: UNCERTAIN (core: 16, fail: 0, uncertain: 6, background warnings: 1)

Currently ignored by routing

Synonym changes without refs: 1

Artifacts uploaded:

changes.json
routing.json
runs/

ct31-sanger · 2026-10-05T10:51:46Z

@ai4c-agent please remove the CA2, CEACAM7, FABP1, TFF1 markers from the definition. Not strong enough evidence to support this.

ai4c-agent · 2026-10-05T10:52:03Z

🤖 Working on it...

Follow along: View workflow run

— @ai4c-agent

Per reviewer feedback on PR #3777: evidence is not strong enough to support including CEACAM7, FABP1 or TFF1 in the marker list for crypt top colonocyte (CL:9900000). Removed the sentence that referenced those markers; the strongly-supported AQP8/CA1/CA2/SLC26A3 and GUCA2A claims remain unchanged, as do the structural/comparative statements. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ai4c-agent · 2026-10-05T10:54:28Z

🤖 Working on it...

Follow along: View workflow run

— @ai4c-agent

ai4c-agent · 2026-10-05T10:54:36Z

Done — removed CEACAM7, FABP1, and TFF1 from the definition on branch ai4c-agent-issue-3776-run1011 (commit b234911).

I dropped the full sentence ("Additional reported markers include CEACAM7, FABP1 and TFF1, consistent with peak absorptive and secretory capacity at the luminal surface.") rather than leaving a dangling fragment. The amended definition now reads:

A colonocyte of the human colon located at the crypt top, representing the most differentiated, mature stage of the colonocyte lineage. AQP8, CA1, CA2 and SLC26A3 are the markers with the strongest and most consistent evidence of high expression in crypt top colonocytes [PMID:22081019], with high AQP8 expression further supported in bioengineered human colon organoid models [PMID:38876106]. GUCA2A expression at the crypt top is supported by both mouse [PMID:31666564] and human colon organoid [PMID:38876106] models. It sits at the opposite end of the crypt-villus differentiation axis from early colonocyte (CL:4047018), the earliest post-mitotic stage of the same lineage, and is distinct from BEST4+ colonocyte (CL:4047052), a separate ion-channel-expressing subtype scattered through the colon epithelium.

All four PMID xrefs on the definition axiom (22081019, 31666564, 35530046, 38876106) are retained — each still backs at least one remaining claim. Synonyms, parent, metadata and the GUCA2A reference set are unchanged. robot convert -vvv round-trips the file cleanly. Ready for another /clara pass whenever you are.

🤖 Generated by @ai4c-agent

Model: claude-opus-4-7
Agent harness: claude-code
Triggered by: @ct31-sanger
Run: View workflow run

@ct31-sanger

Per @ct31-sanger review on PR #3777: CA2 does not have strong enough evidence to support inclusion in the marker list. PMID:22081019 contains internally conflicting signals for CA2 (listed with CA1/SLC26A3/AQP8 as a top-of-crypt marker in one place, but assigned to a crypt-bottom immature progenitor compartment in another), so it has been dropped. The remaining strongest-evidence marker list on the definition is now AQP8, CA1 and SLC26A3 (cited to PMID:22081019). All other claims and references are unchanged. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

ai4c-agent · 2026-10-05T10:57:08Z

Done — removed CA2 from the definition on branch ai4c-agent-issue-3776-run1011 (commit 3589d7d).

Note: CEACAM7, FABP1 and TFF1 were already dropped in the previous commit (b234911), so this latest change is CA2-only — the "additional reported markers" sentence is already gone.

The strongest-evidence marker clause now reads:

AQP8, CA1 and SLC26A3 are the markers with the strongest and most consistent evidence of high expression in crypt top colonocytes [PMID:22081019], with high AQP8 expression further supported in bioengineered human colon organoid models [PMID:38876106].

Everything else (GUCA2A sentence, structural/comparative claims, parent, synonyms, metadata, PMID xref set) is unchanged — all four definition PMIDs still back at least one remaining claim. robot convert -vvv round-trips the file cleanly. Ready for another /clara pass whenever you want to re-verify.

🤖 Generated by @ai4c-agent

Model: `claude-opus-4-7`
Agent harness: claude-code
Triggered by: @ct31-sanger
Run: View workflow run

ct31-sanger · 2026-10-05T12:28:48Z

/clara

github-actions · 2026-10-05T12:31:17Z

Claude finished @ct31-sanger's task in 3m 31s —— View job

CLARA routed verification — complete

Ran the routed CLARA review per clara_workflow_ref/clara_workflow/agent_instructions.md against routing.json (verification-only; no ontology edits, commits, or pushes), re-checking CL:9900000 after the 2026-10-05 marker-trim commits (b234911 removed CEACAM7/FABP1/TFF1, 3589d7d removed CA2).

Todo

Read agent_instructions.md, routing.json, changes.json
Process term group CL:9900000 (targets: ntr:CL:9900000, synonym:CL:9900000:1)
- Decompose current definition into 14 atomic assertions (12 core, 2 background)
- Stage B: snippet-search verification against routed refs (PMID:35530046, PMID:38876106, PMID:22081019, PMID:31666564)
- Stage C: full-text fallback for the one unresolved core assertion
- Wrote runs/CL_9900000/verdicts.json, tool_calls.jsonl, report.md
Post final summary

Result for `crypt top colonocyte` (CL:9900000): PASS — core: 12, fail: 0, uncertain: 1; background warnings: 1

This is an improvement over the previous run (core: 16, fail: 0, uncertain: 6) — trimming CEACAM7, FABP1, TFF1 and CA2 resolved every previously-uncertain marker claim.

AQP8, CA1, SLC26A3 — pass, explicitly confirmed by PMID:22081019 (Dalerba et al. 2011), which lists CA1+/SLC26A3+/AQP8+/MS4A12+ as the defining marker set of the "top-of-the-crypt" mature-enterocyte population. The CA2 conflict flagged previously is now moot since CA2 was removed.
GUCA2A — pass, clearly supported by the mouse model (PMID:31666564): Guca2a-Venus fluorescence in colon was "largely restricted to the surface epithelium at the top of the crypts," dominated by mature colonocytes.
AQP8 — organoid-specific half (PMID:38876106), uncertain: snippet search returned only title/abstract with no explicit AQP8 mention, and get_europepmc_full_text returned no content for this identifier (same retrieval gap as the prior run). Doesn't affect the term-level pass since PMID:22081019 independently supports the AQP8 claim (any-reference-supports rule).
Structural/comparative claims (subclass of colonocyte, crypt-top location, most-differentiated stage, distinct from BEST4+ colonocyte, "surface colonocyte" synonym) — all pass, strongly supported by PMID:35530046, the paper that coined "Crypt Top Colonocytes (CTC)" as a cluster distinct from BEST4+ cells.
Background (uncertain): "early colonocyte (CL:4047018) is the earliest post-mitotic stage" — subject is a different term, not phrased that way in PMID:35530046; flagged as a background warning only, doesn't affect the term-level result.

Full assertion-by-assertion detail: runs/CL_9900000/report.md (machine-readable verdicts in runs/CL_9900000/verdicts.json, tool call log in runs/CL_9900000/tool_calls.jsonl).

No source files, routing.json, or changes.json were modified; no commits or pushes were made.

github-actions · 2026-10-05T12:35:05Z

CLARA review for PR #3777

Base SHA: b7cb31d4d3eae2790cb9631fd86f3ee70e1fe963
Head SHA: 3589d7d84d5272d5c4ac5c99413b772d05d1b9d7
Terms touched: 1
Total extracted changes: 9
Reviewable changes: 4
Decomposable changes: 1
Routed validation targets: 2
Routed NTR bundles: 1
Routed revised-text targets: 0
Routed ref-only-addition targets: 0
Routed relationship targets: 0
Routed synonym targets with refs: 1

Stage 2/3 verification ran for 1 routed term(s).

Routed verification results

CL:9900000 — crypt top colonocyte: UNCERTAIN (core: 12, fail: 0, uncertain: 1, background warnings: 1)

Currently ignored by routing

Synonym changes without refs: 1

Artifacts uploaded:

changes.json
routing.json
runs/

This was referenced Oct 2, 2026

CL new term request: crypt top colonocyte #3776

Open

NTR: crypt top colonocyte (closes #3756) #3757

Closed

[NTR] crypt top colonocyte #3756

Closed

This was referenced Oct 2, 2026

NTR: mid crypt colonocyte (CL:9900000) — fixes #3778 #3779

Open

CL new term request: mid crypt colonocyte #3778

Open

ct31-sanger assigned dosumis and mkeays Oct 2, 2026

Conversation

ai4c-agent Bot commented Oct 2, 2026

Summary

Notes for the reviewer

Test plan

Uh oh!

ct31-sanger commented Oct 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented Oct 2, 2026

Ontology comparison

Left

Right

Ontology imports

Ontology annotations

crypt top colonocyte http://purl.obolibrary.org/obo/CL_9900000

Added

Uh oh!

github-actions Bot commented Oct 2, 2026

Ontology comparison

Left

Right

Ontology imports

Ontology annotations

crypt top colonocyte http://purl.obolibrary.org/obo/CL_9900000

Added

Uh oh!

ct31-sanger commented Oct 2, 2026

Uh oh!

github-actions Bot commented Oct 2, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CLARA routed verification — complete

Result for crypt top colonocyte (CL:9900000): needs revision

Uh oh!

github-actions Bot commented Oct 2, 2026

CLARA review for PR #3777

Routed verification results

Uh oh!

ct31-sanger commented Oct 5, 2026

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ct31-sanger commented Oct 5, 2026

Uh oh!

github-actions Bot commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CLARA routed verification — complete

Result for crypt top colonocyte (CL:9900000): no outright core failures; 6 of 16 core assertions uncertain

Uh oh!

github-actions Bot commented Oct 5, 2026

CLARA review for PR #3777

Routed verification results

Currently ignored by routing

Uh oh!

ct31-sanger commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ai4c-agent Bot commented Oct 5, 2026

Uh oh!

ct31-sanger commented Oct 5, 2026

Uh oh!

github-actions Bot commented Oct 5, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CLARA routed verification — complete

Result for crypt top colonocyte (CL:9900000): PASS — core: 12, fail: 0, uncertain: 1; background warnings: 1

Uh oh!

github-actions Bot commented Oct 5, 2026

CLARA review for PR #3777

Routed verification results

Currently ignored by routing

Uh oh!

This branch has not been deployed

ct31-sanger commented Oct 2, 2026 •

edited

Loading

crypt top colonocyte `http://purl.obolibrary.org/obo/CL_9900000`

crypt top colonocyte `http://purl.obolibrary.org/obo/CL_9900000`

github-actions Bot commented Oct 2, 2026 •

edited

Loading

Result for `crypt top colonocyte` (CL:9900000): needs revision

github-actions Bot commented Oct 5, 2026 •

edited

Loading

Result for `crypt top colonocyte` (CL:9900000): no outright core failures; 6 of 16 core assertions uncertain

ct31-sanger commented Oct 5, 2026 •

edited

Loading

github-actions Bot commented Oct 5, 2026 •

edited

Loading

Result for `crypt top colonocyte` (CL:9900000): PASS — core: 12, fail: 0, uncertain: 1; background warnings: 1