ClinGen Mendelian review progress

Project and complete gene inventory

Process the nuclear protein-coding Definitive, Strong, Moderate, and Limited tiers
in order, alphabetically within each tier, followed by mitochondrial protein genes,
RNA genes, other HGNC locus types, and undetermined-inheritance follow-ups. Keep
RNA genes in scope and use just fetch-ncrna human SYMBOL (RNAcentral identifiers)
before reviewing their functional literature. The archived HGNC subset records both
locus_group and locus_type: 2,837 protein-coding genes, 35 non-coding RNAs, and
4 other loci (readthrough, immunoglobulin, or T-cell receptor genes).

Existing reviews receive a substantive audit; a COMPLETE status does not certify
assessment against this campaign's current evidence and curation rules. The first
batch already demonstrates the value: AARS1's unconditional monomer claim needed
new disease-mechanism evidence, while AARS2 had untraced propagation judgments and
modeled variant effects presented as measured results. Prioritize new substantive
findings, evidence access, and rule compliance; do not manufacture changes to a
sound review merely to enlarge a PR.

Each gene has one dedicated PR containing its review, supporting research and
references, notes, and session history. Address biological review feedback and
in-scope validation failures on that PR. Mark a gene complete only after the final
review is validated, all actionable PR feedback is addressed, and its PR is merged.
Preserve justified UNDECIDED annotations when evidence remains inaccessible; record
the limitation rather than replacing it with an unsupported decision.

Ownership of shared files

Gene PRs change only the gene, its required cached sources, and gene history.
They do not change this log, the shared inventory, or generated project pages.
The coordinator updates those shared files serially in the project branch after
checking each PR's authoritative state. Only after merge does the coordinator tick
the gene's inventory checkbox and record the merged PR and final validation.
The current concurrency is three gene reviewers, not one worker per inventory row.

Symbol migration checks

Before fetching a gene, check its HGNC ID and both current and historical symbols
against existing review directories. These six source-to-approved mappings are
verified by HGNC prev_symbol, not guessed from aliases:

ClinGen source symbol	Approved symbol	Existing review at campaign start
BVES	POPDC1	Neither symbol has a review
CCDC115	VMA22	Neither symbol has a review
CENPJ	CPAP	Neither symbol has a review
NAT8L	ASPNAT	Neither symbol has a review
NDUFA4	COXFA4	`genes/human/NDUFA4/NDUFA4-ai-review.yaml`
TRAF3IP1	IFT54	Neither symbol has a review

COXFA4 must reuse and migrate the existing NDUFA4 review in its dedicated gene PR;
do not fetch a second review independently under COXFA4. Preserve history through
target.superseded_by as documented in docs/history.md. Recheck the directories
at assignment time because other work may have added or renamed a review.

Human-qualified frontmatter deliberately avoids linking same-symbol reviews in
other species. The initial renderer check demonstrated that bare ALB and CDT1
were ambiguous without a human review, and a symbol found in only one nonhuman
species can be linked there despite the human hint. Missing-human-review warnings
are therefore expected; existing human reviews still link normally.

Campaign status — 2026-09-26 UTC

Gene	Tier	Starting review status	Current state	Branch	PR
A4GALT	Definitive	INITIALIZED	Merged; final validation and CI passed	`cmungall/clingen-a4galt`	#3127
AARS1	Definitive	COMPLETE	Merged; final validation and CI passed	`cmungall/clingen-aars1`	#3129
AARS2	Definitive	COMPLETE	Merged; final validation and CI passed	`cmungall/clingen-aars2`	#3128
AASS	Definitive	INITIALIZED	Merged; final validation and CI passed	`cmungall/clingen-aass`	#3133
ABCA3	Definitive	No review	Merged; final validation and CI passed	`cmungall/clingen-abca3`	#3134
ABCA4	Definitive	No review	Merged; final validation and CI passed	`cmungall/clingen-abca4`	#3132
ABCB4	Definitive	No review	Merged; final approval and required CI passed	`cmungall/clingen-abcb4`	#3135
ABCC6	Definitive	No review	Merged; final approval and required CI passed	`cmungall/clingen-abcc6`	#3138
ABCC8	Definitive	No review	Draft; current head approved; PMID:16924481 cache required	`cmungall/clingen-abcc8`	#3136
ABCC9	Definitive	No review	Merged; final approval and required CI passed	`cmungall/clingen-abcc9`	#3148
ABCD1	Definitive	No review	Draft; biological feedback resolved; PMID:16213491 cache required	`cmungall/clingen-abcd1`	#3150
ACAD8	Definitive	INITIALIZED	Merged; final approval and required CI passed	`cmungall/clingen-acad8`	#3151
ACAD9	Definitive	COMPLETE	Merged; final approval and required CI passed	`cmungall/clingen-acad9`	#3152
ACADM	Definitive	COMPLETE	Draft; biological blocker resolved; PMID:18459129 cache required	`cmungall/clingen-acadm`	#3156
ACADS	Definitive	COMPLETE	Merged; final approval and required CI passed	`cmungall/clingen-acads`	#3155
ACADSB	Definitive	INITIALIZED	Merged; final approval and required CI passed	`cmungall/clingen-acadsb`	#3154
ACADVL	Definitive	COMPLETE	Merged; final-head approval and required CI passed	`cmungall/clingen-acadvl`	#3157
ACAN	Definitive	COMPLETE	Draft; biological feedback resolved; PMID:11222505 cache required	`cmungall/clingen-acan`	#3153
ACAT1	Definitive	COMPLETE	Draft; current head approved; eight donor caches required	`cmungall/clingen-acat1`	#3158
ACOX1	Definitive	INITIALIZED	Draft; all biological feedback resolved; PMID:16672280 cache required	`cmungall/clingen-acox1`	#3159
ACOX2	Definitive	INITIALIZED	Draft; reaction and preparation-scope follow-up published; re-review and six caches pending	`cmungall/clingen-acox2`	#3161
ACSL4	Definitive	COMPLETE	Draft; current head approved; notes-only PMID:23766516 cache required	`cmungall/clingen-acsl4`	#3160
ACTA1	Definitive	COMPLETE	Published; review follow-up in progress	`cmungall/clingen-acta1`	#3163
ACTA2	Definitive	COMPLETE	Published; current-head review and CI pending	`cmungall/clingen-acta2`	#3164
ACTB	Definitive	COMPLETE	Substantive audit in progress	—	—
ADA	Definitive	INITIALIZED	Substantive audit in progress	—	—

Branches for work without a PR may exist only in a local isolated checkout.
Published check and approval states are the last verified states from this session,
not a live dashboard. ABCC6 merged through protected auto-merge on September 26
at 21:22:34 UTC. A separate checkout inside the writable workspace supports local
Git commits while the original worktree's shared Git metadata remains read-only.
Publication uses GitHub's API with expected-head guards and exact tree verification.
Direct Git transport and normal publication-source downloads still encounter DNS
failures. No local or published draft is counted as merged or complete.

The publication queue records the exact local file sets
and SHA-256 hashes for the separately published gene revisions. Recheck each remote head
before publishing a follow-up, and keep project tracking in its own PR. ABCC8
still needs the normal cache fetch for PMID:16924481; ABCD1 needs PMID:16213491.
ACAN also needs PMID:11222505, and ACADM needs PMID:18459129. Their targeted validators pass with these warnings, but the reference-cache
requirement remains outstanding. ABCG5 initialization also failed at the normal
UniProt fetch because DNS access was unavailable; no seed files were invented.
ABCG5, ABCG8, ABHD12 and ABHD5 remain queued in the normal alphabetical order.
While those source downloads are unavailable, workers are auditing the next
available cached reviews, now reaching ACTA1 and ACTA2. This temporary
scheduling does not mark the earlier unseeded genes complete. The concurrent cached-review batch assigned ACAN before ACADVL; that was a scheduling exception to strict alphabetical order, and ACADVL has a dedicated audit PR with a published follow-up. ACAD8's genuine
Falcon/fallback launch attempts failed at dependency retrieval before provider
contact; its manual primary-source research is documented without a fabricated report.

The complete log of ABCB4's second review attempt reports the review account's
session limit, with reset at 2026-09-27 01:00 UTC. That earlier comment accepted the biological fixes, but did not record formal approval at the time. ABCB4 later received current-head approval and merged.
The subsequent ABCB4 and ABCC9 reviews completed with concrete changes requested. Later ACAD8 and ACAD9 workflow runs completed successful reviews before that reset
time, demonstrating that the earlier failure was no longer universal. Reran the
earlier failed ABCB4 and ABCC9 reviews after that new evidence. Keep the normal
approval and CI gates and inspect each current run rather than assuming a fixed
outage window.

Merged project setup PR: #3126.

The earliest unseeded nuclear protein-coding Definitive gene is ABCG5. 14 of
2,876 genes are complete for this campaign. ABAT is in the Moderate tier.

Verification log

2026-09-25: Checked open repository PRs before assigning the first batch; no open
A4GALT, AARS1, or AARS2 PR was found. Started a repository-wide baseline
just validate-all; gene-specific validation remains required on each branch.
2026-09-25: A4GALT targeted validation and history validation passed. Independent
biological review found no blocking issue; final research artifact and GitHub
review/checks remain pending on PR #3127.
2026-09-25: AARS2 targeted and history validations passed; independent biological
review found no blocker. PR #3128 awaits GitHub review and required checks.
2026-09-25: AARS1 validation and history checks passed and independent biological
feedback was addressed. PR #3129 includes a genuine late Falcon report; the
wrapper had already timed out, and the report remains unverified as direct evidence.
2026-09-25: Repository-wide baseline just validate-all passed for 4,975 reviews
and 54 pathway files. No blocking schema, ontology, reference, or best-practice
failures remained; advisory warnings were retained.
2026-09-25: AARS2 received approval on commit 3b3d7cd710. All three nonblocking
suggestions were answered with source/schema evidence; required CI remains pending.
2026-09-25: Project review caught unexpanded RNA/other-locus count placeholders.
Fixed the generator and authored page, checked the complete generated output for
unresolved placeholders, and regenerated the affected project pages.
2026-09-26 UTC: Verified AARS2 PR #3128
merged on 2026-09-25 at 22:40:15 UTC as 66219c3f16cd45254d0ed751869341c1f0047855.
The final gene and history validations passed, reviewer approval covered the final
head, all actionable feedback was answered, and required CI passed. Checked its
inventory row complete; the gene review's remaining evidence limitations are
documented rather than treated as unsupported annotations.
2026-09-26 UTC: Project setup PR #3126 passed review and required CI and merged
as 2bbd35ae337fbd15918289caa9e7ddc78c4b1eb8. A4GALT and AARS1 remain open with
further evidence-framing feedback being addressed; neither is counted complete.
ABCB4 is now assigned, with 68 source annotations seeded, publication caching
complete, and a genuine Falcon report available for independent verification.
2026-09-26 UTC: AASS, ABCA3, and ABCA4 have dedicated PRs (#3133, #3134,
and #3132), with targeted validation, history validation, rendering, and source
assertion preservation checks passed. Independent coordinator inspection found
no blocking biological issue; external review and current-head CI remain pending.
Reserved ABCC6 and ABCC8 for the next reviews. A4GALT is approved on
cc2386416e; AARS1's biological fixes were approved on f85785312d, followed
by a history-only wording correction on f0f9e75cd4. Neither is counted complete
before merge. The earlier "CI passed" status referred to superseded heads and
is replaced with current-head pending states.
2026-09-26 UTC: A4GALT PR #3127 merged at 02:50:20 UTC as
2de175d73c1746bedc98d758d5932643af176805; AARS1 PR #3129 merged at
02:54:38 UTC as 5b2c8193f4bc9875f17b0845e5f3165680e0c697. Both had
approval on their final heads and successful required CI, with targeted gene
and history validation already passed. Their two inventory checkboxes are now
complete, bringing the campaign total to three.
2026-09-26 UTC: ABCB4 PR #3135 contains 68 adjudicated assertions and 23
propagation assessments, preserving both NOT flags. Targeted validation,
history validation, rendering and source-object comparison pass; three evidence
questions remain explicitly UNDECIDED. AASS and ABCA3 follow-ups were approved
and await required checks. ABCC6/ABCC8 reviews are underway, ABCC9 is assigned,
and ABCD1 has been initialized after checking its former symbol ALD and alias
ALDP for existing reviews. The next unassigned Definitive gene is ABCG5.
2026-09-26 UTC: Opened progress PR #3137
for the three confirmed completions. ABCC8 PR #3136 reviews 64 source assertions
with 44 propagation assessments; ABCC6 PR #3138 reviews 69 assertions. The
ABCC6 follow-up distinguishes mineral-deposition assays from ion-homeostasis
evidence. ABCB4's local follow-up preserves all 68 assertions and both NOT
flags, leaves clathrin-vesicle localization unresolved, and supplies the
condition-specific cholesterol evidence from its original reference. Gene,
history and rendering checks passed for ABCB4; the revision remains unpublished.
2026-09-26 UTC: Completed independent audits and local revisions for ABCB4,
ABCC6 and ABCC8 after changes were requested on their published PRs. All 68,
69 and 64 source assertions respectively remain intact. ABCC6 now removes
uninformative generic binding terms without asserting absence of interaction,
and distinguishes cellular ATP release from proven direct ATP cargo transport.
ABCC8 now incorporates the mouse GO-CAM insulin-secretion evidence and direct
SUR1 ATPase results, with PMID:16924481 still awaiting its normal cache fetch.
Targeted gene and history validation and rendering pass for all three local
follow-ups; publication remains pending.
2026-09-26 UTC: ABCC9's 53 annotations and ABCD1's 135 annotations received
complete local reviews and independent biological audits. All seeded fields
were preserved; gene validation, history validation and rendering pass. ABCD1
retains eight evidence-limited UNDECIDED decisions and documents the missing
PMID:16213491 cache. Neither gene has a published PR yet. The earlier ABCC8
starting-state placeholder was resolved by checking for an existing human
review before seeding: none was present, so its baseline is "No review".
2026-09-26 UTC: Public GitHub PR, commit and test pages confirm three further
merges, all dated September 26 (exact clock times were not exposed): AASS
#3133 as
5f44a3fc98ffd3fa68a8ded0105cf83803e030b5; ABCA3
#3134 as
52382b633dc98f8f5b12d5f35c3cf3761343d359; and ABCA4
#3132 as
cbb8c133015f10ed37968a1ca71207c95dbd11b7. Approval and successful current-head
tests were verified for each. Their inventory checkboxes are now complete,
bringing the campaign total to six. Tracking PR
#3137 is approved with
its test passed but remains open; these newer local tracking updates have not
been published to it.
2026-09-26 UTC: Recovered local Git operations in an independent checkout at
the ignored workspace path tmp/clingen-git; the original Orca worktree and
source files were preserved. Both follow-ups were committed locally and then
published through GitHub's API with expected-head checks: ABCB4 #3135 at
5fa6ee26208562402e4295703a80ea7cf569b0e2 and ABCC6 #3138 at
681dc1cf8acc6a0487294b8841b0479b617e5fac. GitHub tree hashes exactly match the
corresponding validated local commit trees. Updated both PR descriptions.
Direct Git HTTPS transport still cannot resolve its host; GitHub API access
works. Normal fetches of PMID:16924481 and PMID:16213491 were retried once each
and still failed DNS resolution, so their cache requirements remain open.
2026-09-26 UTC: ABCC6 PR #3138 merged at 21:22:34 UTC as
21121fc735d20bcb2dbf8328aa82d5da30185e5e, after final-head approval and
required CI. Checked its inventory row complete, bringing the total to seven.
Published ABCC8's validated follow-up at 1bf52d95a381d852b6a2a1dc5b44c2bceaefd329
and kept PR #3136 in draft pending PMID:16924481. Opened ABCC9 PR #3148 at
dd674e49837004266dafbafccd62b0b98a2c7a98 and ABCD1 draft PR #3150 at
8090c5a86ab995ecdcc191e17e71f723853ede17; the latter still requires the
normal PMID:16213491 cache. The complete source manifests and remote/local tree
equality checks are recorded in the publication queue. ABCC9's required test passed.
2026-09-26 UTC: Audited the next available cached reviews while earlier gene
initialization remained unavailable. ACAD8 PR #3151 preserves 23 source rows,
corrects substrate-specificity reasoning and passes validation without warnings.
ACAD9 PR #3152 preserves 40 source rows, withdraws one redundant authored NEW
row and retains directly supported beta-oxidation in its core; validation passes
with one intentional advisory. ACAN draft PR #3153 preserves 40 source rows,
corrects matrix-location and binding evidence, and withdraws an unsupported
authored assembly row; its three warnings include the missing PMID:11222505
cache. Independent biological inspection, history validation and rendering pass
for all three. Existing source caches were preserved. ACADM, ACADS and ACADSB
audits are now underway; no unmerged review is checked complete.
2026-09-26 UTC: Inspected the complete failed reviewer log rather than assuming
a biological or Git failure. ABCB4's second attempt reports the reviewer account
session limit, resetting September 27 at 01:00 UTC; no formal current-head
approval was recorded. Further attempts are deferred until reset. Required
review gates remain intact. The project branch's sole main-merge conflict is
its generated inventory HTML; regenerate it from the updated source to retain
the newly available gene links and the seven confirmed completion checkboxes.
2026-09-26 UTC: Published ACAD8 feedback commit
acd39fb8cbb021bbee3841ca977b247b8ac3f802: broad mitochondrial rows retain their
original source resolution, the encompassing short-chain GO term is accepted
without changing its specific propionyl Rhea reaction, and full-text/pathway
wording is corrected. Published ACAD9 feedback commit
53b3e3c398f59a40419f7c080954fd71d2de4b2e: refine the supported beta-oxidation
process, attribute neuronal details to externally read primary text and record
the unresolved assembly molecular function. Both targeted validators now pass
without warnings; all 23 and 40 source assertions respectively are preserved.
2026-09-26 UTC: Published ACADSB #3154 (28 source rows) and ACADS #3155 (30
source rows), each with independent biological inspection and clean targeted
validation. Published ACADM draft #3156 (50 source rows), whose original mouse
process donor was recovered as PMID:18459129; its required normal cache fetch
failed DNS. The review distinguishes metabolic repartitioning from a synthesis
step and retains unresolved mechanistic evidence. Three further cached reviews
are underway: ACADVL, ACAT1 and ACOX1. Review-service retries resumed only after
newer workflows demonstrated successful runs; the previous session-limit report
is historical evidence rather than a blanket block until its quoted reset time.
2026-09-26 UTC: Renewed reviewer feedback identified one remaining ABCB4
YAML-alias mismatch. Published 85a037efcedc539102539973d26fbd2ae328133f
with both biliary secretion rows consistently replaced by phospholipid efflux;
all 68 source assertions and other decisions are unchanged. ABCC9, ACADSB and
ACADS received source-scope or wording feedback and are being revised. ACAD8
and ACAD9 received approval on their latest published heads; protected auto-merge
is enabled while required tests finish. They are not yet counted complete.
2026-09-26 UTC: Project review caught links to four unmerged gene pages in the
generated inventory. Regenerated using only gene-review paths present in the
project branch's target tree (c58fcd6b4589c722799c6e39da663420db32dd7c),
so ABCB4, ABCC8, ABCC9 and ABCD1 remain unlinked until merge. The previous
conflict-resolution note confused locally available reviews with published
pages; this entry corrects that decision. ABCA3, ABCA4 and ABCC6 remain linked
because their pages have merged. Future tracking renders must use the target
Git tree's gene availability rather than the coordinator's working directories.
2026-09-26 22:38 UTC: Confirmed ACAD8 #3151 merged at 22:20:09 UTC as
9e34ed516aaa199a283a5042f6adf3b60ae45d54 and ACAD9 #3152 merged at 22:25:09
as e0d565bddebefd72f7c1c32562b588f75a893f69; both final heads were approved
and required CI succeeded. The local working tally at that point was nine complete;
it was not a separately published project snapshot.
Published ACADVL #3157 (42 source rows, clean validation) and ACAT1 draft #3158
(49 source rows, eight missing donor caches listed). Published quote-scope
corrections for ABCC9, matching binding judgments for ACADS, chemistry/provenance
corrections for ACADSB, and the glycogen participation correction for ACADM.
The unpublished local queue then recorded thirteen latest revisions and 84 hashes, verified against
their recorded Git trees. ABCC9's latest four-file follow-up replaces its earlier
34-file initial manifest; historical commits retain the earlier snapshot.
Receipts identify published trees, not necessarily each PR's live head.
2026-09-26 22:38 UTC: Reviewed feedback on drafts as well as ready PRs. ABCC8
needs a cache-availability flag correction; ABCD1 needs its elongation actions
reconciled with source evidence; ACAN needs binding, localization, qualifier and
narrative corrections. All retain required missing-cache gates. ACADM's biological
blocker is resolved; its missing PMID:18459129 cache still blocks readiness.
ACOX2 research is preserved while ACAN feedback takes priority. Targeted fetches
and a read-only search of available local checkouts and current main found none
of the missing publication caches; no replacements were invented.
2026-09-26 23:10 UTC: Confirmed four further completed genes: ABCB4 #3135 at 22:41:04,
ACADS #3155 at 22:46:40, ABCC9 #3148 at 22:49:06, and ACADSB #3154 at
22:57:39 UTC. Each final head was approved and required CI passed. The inventory
now checks 13 of 2,876 genes complete. Project progress PR #3137 also merged;
this is a separate tracking snapshot. Published ACOX1 #3159, ACSL4 #3160 and
ACOX2 #3161 as drafts with explicit required-cache gates. Updated ACAN, ABCC8,
ABCD1, ACAT1 and ACADVL follow-ups. The queue records 16 latest revisions and
64 file hashes, all checked against their immutable published Git trees.
Earlier larger manifests remain recoverable through their publication receipts.
2026-09-26 23:10 UTC: ACAN and ABCC8 re-reviews accepted the biological fixes
and clarified that citation verification and cache availability are separate.
Missing caches still prevent undrafting. ABCC8 identifier-spacing feedback is
corrected in current prose and a new history record; published history remains
unchanged. Notes-inclusive citation checks also identified ACSL4 PMID:23766516
and six ACOX2 missing records, now explicit in their draft gates. Ordinary PubMed
endpoint and fetch retries still failed DNS despite an unrelated institutional
PDF download succeeding. No source or research report was fabricated.
2026-09-26 23:20 UTC: Corrected the provenance explanation for two generated
history filenames carrying the scaffolder's default claude-code actor token.
Codex performed the work; --agent-tool codex had been supplied but the separate
--actor-name codex option was omitted. Actor metadata was corrected afterward;
filenames and session ids were generated by the helper, not hand-written. New
correctly scaffolded records document this explanation for ABCC8 and this project.
Previously published history remains unchanged under the append-only rule.
Future commands supply both actor name and agent tool explicitly. No inventory,
biological judgment or publication receipt changes in this clarification.
2026-09-26 23:35 UTC: Confirmed ACADVL #3157 merged at 23:27:27 UTC as
af7a6ea1c9a6dd7ceecc8b04b120577d1a4070cb after final-head approval and successful CI.
The campaign total is 14 of 2,876 genes. ACTA1 #3163 and ACTA2 #3164 are published;
ACTA1 review feedback is being addressed. ABCC8 and ACSL4 current heads are approved
but remain drafts for their missing source caches. ACOX2 reaction and preparation-scope
follow-up is published as a draft. ACTB and ADA are under substantive review. ADA is
the next cached Definitive gene after unseeded ACTC1, ACTG1, ACTN1, ACTN2, ACVR1,
ACVRL1 and ACY1 under the documented temporary source-access scheduling exception.
2026-09-26 23:35 UTC: Recovered the omitted ABCC8 31c9d66b publication receipt.
Its original four-file manifest matches the immutable published blobs and the
independently preserved local commit 7f3bf739 has the exact published tree.
Historical receipts now include deterministic changed-file manifests and hashes,
ordered newest first with parent continuity checked. The current queue records
18 latest revisions and 71 file hashes; 8 of these PRs are merged, while the first
6 campaign completions predate this queue. Correctly scaffolded history links this
snapshot to PR #3162 and preserves the earlier generated actor-token record.