[NoopDog]

Closes #720.

What changed

I regenerated the reports from run 20261019_225524 as reconciled on main after #719 (fa826ed), comparing with 20261019_084533 as before:

make all-reports RUN_DIR=output/anvil/20261019_225524 PREVIOUS=output/anvil/20261019_084533

Three reports change. Each change comes from #719's SGDP donors:

• Reconcile report (reconcile-report.md, reconcile-dashboard.html):
  • it is now reconciled by fa826ed;
  • the translation table has a new sha256 (SGDP's M/F rows), so the comparison line now says "a different translation table";
  • no file-level number moves, since all 16 file artifacts are identical.
• Rules report (rules-report.md, rules-dashboard.html):
  • 190 rules (+2) and 18 sample and donor rules (+2), counted over 3,668 donor records (was 3,416);
  • the new rows: sgdp_id 104, hgdp_cell_line 130, phenotypic_sex.sgdp_m 174 and phenotypic_sex.sgdp_f 101. 174 + 101 = 275, one sex claim on each SGDP donor record: 252 new (130 hgdp, 104 sgdp, 18 coriell) and the 23 existing 1000 Genomes donors, measured on the reconciled run;
  • coriell_cell_line_or_dna +18, which is the 18 new Coriell NA… donors.
• Review queue (review-queue-report.md, review-queue.html): SGDP's 129 populations are listed as unreviewed external values. That section said "No unreviewed values" before.

The coverage, consistency, unprocessable and validation reports regenerated byte-identical.

Why

#719 imported SGDP's donors and reconciled the shared run again on main. The run's file records didn't change, but its donor and sample records did, and the code gained two identifier rules. The committed reports still showed the run as #716 left it.

Assumptions I made

• Comparison run. I compare with 20261019_084533, as Read each NIA_CARD fast5's sample from its first 64 KiB (#605) #716 did, so the reconcile report's change table spans Read each NIA_CARD fast5's sample from its HDF5 attributes (cross-check with #604's folder join when it lands) #605 and SGDP donors for ANVIL_T2T_CHRY, with transitive synonyms (#559) #719 together.
• Classified by. "Classified by" still names 5841265, Read each NIA_CARD fast5's sample from its HDF5 attributes (cross-check with #604's folder join when it lands) #605's branch commit, because Regenerate the reports from run 20261019_225524 as reconciled on main after #719 #720 only reconciles. That commit isn't on main, but GitHub keeps it under refs/pull/716/head, and its readings were measured equal to main's. Classifying the corpus again on main is a separate decision that wasn't taken.
• Decided at review (all declined for this PR):
  • The rules report has no row for the same_donor_as join. It keys records rather than claiming values, and each synonym cites an identifier rule the report already lists.
  • SGDP's review-queue rows have an empty dataset cell, because SGDP's file has no dataset level. The source and table columns still identify it.
  • SGDP's populations stay unauthored per Capture the donors and biosamples AnVIL leaves out of T2T_CHRY, MAGE and nhp_dGTEx, and link them across datasets #559's decision 3A. Twelve candidate HANCESTRO matches are noted on Capture the donors and biosamples AnVIL leaves out of T2T_CHRY, MAGE and nhp_dGTEx, and link them across datasets #559.
  • The reconcile report has no donor section. That's Show sample-ID coverage per dataset and outcome in the published dashboards #676, which now has SGDP donors for ANVIL_T2T_CHRY, with transitive synonyms (#559) #719's numbers.

How to verify

#720's definition of done:

• Every generated report under docs/ that reads an AnVIL run names 20261019_225524, and any report reading the run's reconcile provenance shows fa826ed.
  grep -l 20261019_225524 docs/*.md lists the reconcile, rules, consistency and unprocessable reports. The coverage report names the run by its start time, "Classification run: 2026-10-19 22:55:24". docs/reconcile-report.md reads "Reconciled by: commit fa826ed".
• The rules report lists the sgdp_id and hgdp_cell_line identifier rules, and shows the new donor total where it counts donor records.
  In docs/rules-report.md, the identifier-rule table has sgdp_id (104) and hgdp_cell_line (130), and the header reads "3,668 donor records".
• Any report whose numbers change is explained, with where the change comes from. See "What changed" above. To reproduce, rerun the make all-reports command above on main after SGDP donors for ANVIL_T2T_CHRY, with transitive synonyms (#559) #719: git status should then show no changes.

🤖 Generated with Claude Code

https://claude.ai/code/session_01VF7tbFa5ZTmYWZ6JWKiBBS
🐦 Built with cc-claude-tools

[Copilot]

🟢 Approval recommended

The supplied report changes match the stated scope, and the alleged missing HTML populations are present.

No open issues 🎉

What changed in this PR

Refreshes generated reports for run 20261019_225524 after #719’s SGDP donor import, addressing #720.

Changes:

• Updates reconciliation provenance and translation-table identity.
• Reports 3,668 donors and the new SGDP-related rules.
• Adds 129 unreviewed SGDP populations to both queue formats.

File	Description
docs/​rules-report.md	Updates donor totals, rules, and usage counts.
docs/​review-queue.html	Adds SGDP populations and sex mappings.
docs/​review-queue-report.md	Reflects the same queue additions in Markdown.
docs/​reconcile-report.md	Updates reconciliation commit and comparison metadata.

_{Review effort: Balanced}