[codex] Review yeast CCT4 EGD1 YAR1 NMD3 CCT8

[cmungall]

Summary

Adds completed yeast gene reviews for CCT4, EGD1, YAR1, NMD3, and CCT8.

This batch integrates Falcon deep research into the review YAMLs, refreshes rendered HTML, and adds directly cited PANTHER family context for CCT/TRiC, NAC/BTF3, NMD3, and YAR1 family evidence.

Validation

• just validate yeast CCT4
• just validate yeast EGD1
• just validate yeast YAR1
• just validate yeast NMD3
• just validate yeast CCT8
• just render yeast CCT4
• just render yeast EGD1
• just render yeast YAR1
• just render yeast NMD3
• just render yeast CCT8
• git diff --check

Curator QA subagent approved all five genes with no blocking issues; nonblocking term-specificity suggestions for CCT4/CCT8/NMD3 were applied before this PR.

[github-actions]

Claude finished @cmungall's task in 5m 46s —— View job

---

PR Review: Yeast CCT4, EGD1, YAR1, NMD3, CCT8

Review checklist:

• Schema validation passes (per PR description; uv/just not available in this environment)
• All PMIDs verified as real (all 21 cited PMIDs found in publications cache)
• Supporting text matches publication content (spot-checked key claims)
• GO terms correct and specific
• Review actions well-justified
• Core functions accurately identified
• Over-annotation issues handled

---

Summary

Five well-documented yeast gene reviews. All PMIDs are real and key supporting-text quotes verified against cached publications. GO term aspect assignments are correct throughout. The REMOVE decisions for YAR1's inappropriate IBA transfers are particularly strong. One 🟡 IMPORTANT issue found in EGD1; two 🔵 SUGGESTIONs for NMD3.

Verdict: Approvable. The EGD1 issue below should be addressed before or shortly after merge, but is not a hard block.

---

🟡 IMPORTANT — EGD1: cotranslational protein targeting to membrane should be KEEP_AS_NON_CORE, not ACCEPT

File: genes/yeast/EGD1/EGD1-ai-review.yaml, annotations for GO:0006613 (IEA ARBA and IGI PMID:10518932 entries) and core_functions.directly_involved_in

Issue: Both GO:0006613 annotations are ACCEPT'd, and GO:0006613 appears in core_functions.directly_involved_in. However, the only experimental evidence for this function in yeast (PMID:10518932) is purely in vitro — the study uses a cell-free translation system with yeast microsomes: "we use an in vitro translation system in combination with isolated microsomes."

PMID:26618777 — already cited in this review for the cotranslational folding annotation — explicitly states: "this function is less well understood in yeast where in vivo studies showed no aberrant translocation phenotype upon NAC deletion." This makes the in vivo relevance of ER-targeting gating in yeast genuinely uncertain.

Suggested fix: Change both GO:0006613 annotation actions to KEEP_AS_NON_CORE with a reason acknowledging the in vitro support (PMID:10518932) but lack of in vivo phenotype in yeast (PMID:26618777). Remove GO:0006613 from core_functions.directly_involved_in. The cotranslational folding annotation (GO:0051083) is unaffected.

---

🔵 SUGGESTION — NMD3: NES position in reason text is off by two residues

File: genes/yeast/NMD3/NMD3-ai-review.yaml, nucleus IBA annotation review.reason

The reason states NES at positions 493-502. PMID:11086007 maps the NES sequence INIDELLDEL to aa 491–500 and the NLS to aa 387–435 (not 399-415). If these numbers come from a secondary source, that source should be cited; otherwise correct to match the primary reference.

---

🔵 SUGGESTION — NMD3: Core function supporting text for rRNA binding doesn't quote the rRNA evidence

File: genes/yeast/NMD3/NMD3-ai-review.yaml, second core_functions entry (GO:0070180)

The supporting text for PMID:24240281 reads "Nmd3 is an essential nuclear export factor" — which describes the overall role but not the rRNA binding. A quote already used in existing_annotations for the same PMID is much more precise: "binding sites, which were found to lie in H38, H69 and H89 of 25S rRNA." Swapping this in would better support the rRNA-binding core function.

---

Gene-by-gene notes (no blocking issues)

CCT4 / CCT8: Internally consistent. The three MODIFY decisions replacing GO:0051082 with GO:0140662 are correct for all evidence types (IBA, IEA, IDA). MARK_AS_OVER_ANNOTATED on protein-binding IPI entries is well-justified — interactome studies don't define a specific binding function beyond chaperonin complex membership. PMID:19536198 supporting text ("indirect TAP-tag based interactions") verified in the cached publication.

YAR1: The REMOVE decisions for inappropriate IBA transfers (GO:0001228, GO:0045944, GO:0030907, GO:0033309) are a highlight. PTHR24198 is confirmed as a generic 76,778-member ankyrin-repeat family with no connection to MBF/SBF transcription. PMID:22570489 positively establishes Yar1 as an Rps3 carrier chaperone. GO:0140597 (protein carrier chaperone) as the MODIFY replacement for GO:0051082 is well-chosen.

NMD3: Solid throughout. IDA-supported ACCEPT of GO:0030674 (protein-macromolecule adaptor activity) is well-justified by PMID:11086007. MODIFY of GO:0015031 → GO:0000055 correctly tightens the broad protein-transport annotation.

---

Note: GitHub Actions cannot submit formal PR approvals. This review is posted as a comment. A human reviewer should approve.

[github-actions]

Claude finished @cmungall's task in 2m 1s —— View job

---

PR Re-review: Yeast CCT4, EGD1, YAR1, NMD3, CCT8

Review checklist:

• Schema validation passes (per PR description)
• All PMIDs verified as real (PMID:10518932, 26618777, 11086007, 22570489, 16762366, 15704212, and others confirmed in publications cache)
• Supporting text matches publication content
• GO terms correct and specific
• Review actions well-justified
• Core functions accurately identified
• No over-annotation issues missed

---

Verdict: APPROVE ✅

All three issues raised in the prior review have been resolved in commit 0205b6cb. No new issues found. This PR is biologically and curatorially sound.

---

Previous Issues — All Fixed

🟡 EGD1: GO:0006613 ACCEPT → KEEP_AS_NON_CORE — Fixed. Both the IEA (ARBA) and IGI (PMID:10518932) entries for cotranslational protein targeting to membrane are now KEEP_AS_NON_CORE. The reason correctly cites in vitro evidence (PMID:10518932) against lack of in vivo translocation phenotype (PMID:26618777). GO:0006613 also removed from core_functions.directly_involved_in.

🔵 NMD3: NES position off by two residues — Fixed. Updated to aa 491–500 (INIDELLDEL) and NLS to aa 387–435, matching PMID:11086007.

🔵 NMD3: rRNA binding core function supporting text — Fixed. PMID:24240281 supporting text in both the core_functions rRNA binding entry and the GO:0070180 IDA existing annotation now reads: "binding sites, which were found to lie in H38, H69 and H89 of 25S rRNA."

---

Gene-by-gene notes (no issues found)

CCT4 / CCT8: Clean and internally consistent. Three MODIFY decisions replacing GO:0051082 (unfolded protein binding) with GO:0140662 (ATP-dependent protein folding chaperone) are correct across all evidence types (IBA, IEA, IDA). MARK_AS_OVER_ANNOTATED on all protein binding IPI entries is well-justified — interactome studies don't define a specific binding activity beyond chaperonin complex membership. For CCT8, the SAGA/ADA paper (PMID:21734642) is correctly flagged.

YAR1: The four REMOVE decisions for inappropriate IBA transfers (GO:0001228, GO:0045944, GO:0030907, GO:0033309) are a highlight of this batch. PTHR24198 is correctly identified as a broad ankyrin-repeat family with no MBF/SBF relevance in yeast. MODIFY of GO:0051082 → GO:0140597 (protein carrier chaperone) is well-chosen and supported by PMID:22570489 and PMID:26112308.

EGD1: Now clean. Cotranslational folding (GO:0051083) correctly ACCEPT'd. Non-core roles (mitophagy, macroautophagy) appropriately held as KEEP_AS_NON_CORE.

NMD3: Solid throughout. IDA-supported ACCEPT of GO:0030674 (protein-macromolecule adaptor activity) is the strongest annotation in this gene — Crm1-adapter function precisely supported by PMID:11086007 and PMID:11313466. MODIFY of GO:0015031 → GO:0000055 correctly tightens the broad protein-transport annotation.

---

Note: GitHub Actions cannot submit formal PR approvals. This review is posted as a comment. A human reviewer should click Approve.