Repository navigation
Dr. Diehl 2026-10-05: NK species audit, CL issue drafts, group 1 ILC hierarchy proposal - #25
Draft
s28ale-maker wants to merge 11 commits into
Draft
s28ale-maker wants to merge 11 commits into
s28ale-maker wants to merge 11 commits into
Conversation
…10-05) Task A + D of the 2026-10-05 meeting plan. - pro_marker_species_support.tsv: new columns human_status / mouse_status (supported, contradicted, no_evidence_found, not_applicable), counter_citation / counter_quote (other side of a contradiction) and audit_batch. Legacy 10 columns unchanged on all 87 original rows. - 66 directly asserted pairs on the 28 marker-bearing terms under CL:0001067 (Ubergraph, 2026-10-05) plus the 4 plasmablast pairs; 57 new rows. 34 quotes, all checked verbatim against Europe PMC full text (abstract for Chiossone 2009). - HLA-DRA: CL:0000938 and CL:0000939 both assert "lacks HLA-DRA"; no CL term asserts HLA-DR+ NK; neither definition source supports the axiom; human evidence contradicts it for a minority subset. One-sheet export in nk_hladr_conflicts_2026-10.tsv. - nk_marker_audit_2026-10.md: counts per status, requested cases, 9 other issue candidates, open questions. - plasmablast_marker_audit_2026-10.md: lineage (inherits in-taxon Mammalia via CL:0001200), verdicts, proposed human/mouse children. - Skill and reports README document the new vocabulary. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ehl, 2026-10-05) Task B of the 2026-10-05 meeting plan. Nothing filed. - gaps.tsv: 9 appended rows (existing rows untouched). 7 draft_for_review (B1 lacks HLA-DRA on CL:0000938; B3a plasmablast lacks CD138; B3b plasmablast markers to a human child; B6a DX5+ on CL:0002345; B6b PMID:1994922 typo; B6c "NK1.1-positive T cell" definitions; B6d galectin-1 relation) and 2 on_hold pending Dr. Diehl (B2 opposing HLA-DR definition not found in CL; B6e pre-NK markers). - gap-issue-filing skill: never file draft_for_review / on_hold rows. - docs/protege_ncam1_query_repro.md: expected classes from Ubergraph, NCAM1/CD56 are PRO synonyms not labels, high/low-amount sub-properties and PR:P13591 as likely causes, what to record. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ch on Windows core.autocrlf=true checked these files out with CRLF, so test_benchmark_reserve failed locally although the committed content matched the frozen hashes. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…aw CD56 draft - docs/diehl_review_2026-10.md: batch 1 (B1 HLA-DRA on CL:0000938 and CL:0000939, B3a plasmablast lacks CD138 with quotes, B3b plasmablast child with both human and mammal options), batch 2 (B6a-d), on-hold items. Nothing filed. - gaps.tsv: B1 now one issue for both NK terms; B2 not_a_cl_issue (the conflict was in Dr. Diehl's old spreadsheet); old "add CD56 to CL:0000623" row withdrawn (generic NK keeps only lack statements). - cl_term_issues.md, audit report and gap-issue-filing skill updated to match (withdrawn / not_a_cl_issue are closed statuses). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…sign only) - Before/after trees for the 37-term CL:0001067 subtree (Ubergraph 2026-10-05; parents match CL 2026-03-26). ILC1 and NK kept apart under group 1; NK gets generic, human and mouse branches. - 6 new grouping terms (natural killer cell / immature NK / mature NK, each ", human" and ", mouse") plus 3 optional (uterine NK, human; group 1 ILC, human / mouse, the latter matching issue #14). - 15 re-parentings incl. CL:0002442 under CL:0002438; 4 terms with unclear species and 5 questions flagged for Alex and Dr. Diehl. - TSV lists every term: assignment, current/proposed parents, new term needed, reason. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Proposal: "Why a parallel hierarchy"; none of NEW:1-9 exists in CL incl. obsolete terms (52 label variants vs CL 2026-03-26 and Ubergraph); NEW:7 and NEW:8 (issue #14) recommended, NEW:9 optional; logical definitions with in_taxon as the suggested two-parent default. - Recommended defaults for the four unclear-species terms with sources: mice have uNK (Shreeve 2021); hepatic pit cell first described in rat (Wisse 1976, PMID:991252); cycling NK source is human PBMC (Hao 2021). - CL:0002349 note kept with the exact Chiossone 2009 four-stage quote and Andrews 2010 (both abstracts; Chiossone full text not accessible). - gaps.tsv: draft C1, liver-resident NK human sentences to a human child. - diehl_review_2026-10.md: numbered "Decisions needed" list up top. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…(Task E) - pro_species_support: SPECIES_STATUSES, HUMAN_ONLY_MARKERS (HLA-DRA, NCAM1), not_applicable_reason(), species_status() (a blank becomes not_applicable when the term/protein is limited to the other species or a human-only marker is on the mouse side; never overrides a recorded value) and species_scope(). - soulcap-audit: per-species status + basis on every species row; summary counts human/mouse statuses; only "contradicted" raises a finding (species_marker_contradicted); unknown values are flagged; dashboard shows Human / Mouse badges and defines the vocabulary. - Regenerated audit_* (6 contradicted findings). The matcher evaluation shows stale_or_unverified: its source fingerprints predate later commits to cl_match.py / marker_resolution.py as well as this audit.py change. - Tests for the helpers, the audit behaviour and the curated TSV (vocabulary, quotes behind verdicts, no failure where a species can't apply). mapping-audit and pro-marker-species-support skills updated. - marker_validation.md is grammar-only (no evidence statuses); unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…k F) - M2: human/mouse marker validity vocabulary and the NK / group 1 ILC and plasmablast audit (done), remaining pairs not assessed. - M3: soulcap-audit keeps contradicted / no evidence / not applicable apart; 6 contradicted assertions in the review queue; blocker unchanged. - M4: group 1 ILC parallel hierarchy proposal (design only). - CL term corrections: issue drafts awaiting Dr. Diehl, the on-hold pre-NK item, the Protégé repro note, the withdrawn CD56 draft. - roadmap-status-sync after a fresh soulcap-sync (sheet unchanged since 2026-09-18): M1 75/56, M2 8/14, M3 sheet columns still absent, M4 80/127, #13/#14 and CL#3663/#3664 still open. No status marker changed. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…issue - CLAUDE.md: never send Alex's email or other personal information to an outside service without asking. - ROADMAP.md (cross-cutting): matcher evaluation source fingerprints hash raw bytes, so CRLF checkouts on Windows show the evaluation as stale. Noted for a later fix. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Only that folder is ignored, so open-access PDFs can still be committed. No paywalled PDF is tracked (the only tracked PDF is our own Fall 2026 planning draft). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It is a personal preference, now in the user-level ~/.claude/CLAUDE.md; this repo is shared with David's team. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to the 2026-10-05 meeting with Dr. Diehl. Stacked on #23 (
pr-09-species-pilot), so this diff shows only this work.For review
What changed
reports/pro_marker_species_support.tsvgainshuman_status/mouse_status(supported,contradicted,no_evidence_found,not_applicable). It covers 66 marker assertions on the group 1 ILC / NK subtree plus 4 on plasmablast. Every quote is checked word for word against Europe PMC. Summaries:reports/nk_marker_audit_2026-10.md,reports/plasmablast_marker_audit_2026-10.md, and the HLA-DR sheetreports/nk_hladr_conflicts_2026-10.tsv.reports/gaps.tsv. Thegap-issue-filingskill never files draft, on-hold, withdrawn or not-a-CL-issue rows.soulcap-auditkeepscontradicted,no_evidence_foundandnot_applicableapart. Onlycontradictedraises a finding, and human-only markers (HLA-DRA, NCAM1) on the mouse side count asnot_applicable. Tests cover this and check the curated TSV.docs/protege_ncam1_query_repro.md, for Dr. Diehl to reproduce the NCAM1 query problem..gitignore:literature/paywalled/holds local library copies of paywalled papers for quote checks; it is never committed. Open-access PDFs can still be committed elsewhere.Notes
.gitattributesoverlaps with Force LF on the mapping TSVs whose hashes tests check #24.d1d0a2caddsmappings/*.tsv text eol=lf, and Force LF on the mapping TSVs whose hashes tests check #24 adds the same fix for two named files. Whichever merges second needs a one-line conflict resolution; keeping the glob covers both.cl_match.py. This is noted in ROADMAP for a later fix and not changed here.🤖 Generated with Claude Code