[hyphaltip]

v0.8.1 results: tables, manifests, site data and the v081 release entry

Run: 110 jobs, all COMPLETED; freeze commit 50bc950 (main after #272), --preset v0.8.1. This PR adds the results of that run and makes v081 the site's default release. No change to src/ or db/.

Counts (baseline: v0.8.0 reports)

Campaign	Listed	Called	Uncalled	No report	Loci (high / medium / low)
Ascomycota	19,415	17,618	1,725	72	20,843 (14,623 / 6,120 / 100)
Basidiomycota	3,275	2,995	268	12	5,993 (3,822 / 2,161 / 10)
Mucoromycotina	974	868	105	1	914 (identical to v0.8.0)

Two Ascomycota genomes (GCA_016165855.1 P. ulei IBUN, GCA_023522825.1 DampeI2T2) hit the 3,600 s per-genome timeout on an Opteron node and were rerun with GENOME_TIMEOUT=14400, same frozen code and preset, on an EPYC node (3,063 s and 1,685 s); noted in the Ascomycota MANIFEST.txt. Cause of the slowness: tblastn -seg no on long low-complexity reference proteins (#276; nothing changed in this release).

Determinism

60 genomes (seed 20261106) run twice on the frozen worktree: 0 differences in family, idiomorph, confidence, contig, start, end between the two runs or against the campaign. The only difference is the order of gene names in genes_found in one genome.

Attribution of every difference against v0.8.0

23,570 genomes have a report in both runs: 23,214 identical, 356 differ. No calls lost, no idiomorph flips, nothing unexplained.

• S. pombe silent pair split (S. pombe silent pair: split the merged mat1 call into mat2 P and mat3 M by arrangement (#254) #274): 37 genomes.
• Hyphopichia MTLA1 record (db: add the Hyphopichia burtonii MTLA1 record (#148) #273): 302 genomes (a control run without the record reproduces v0.8.0 exactly for all 302); 190 change a call field: 4 new calls, 160 span changes, 36 loci up and 14 down in tier.
• Fallback after withheld hits (Ogataea routing: a lineage-routed MATog family with all hits withheld leaves no fallback to MATsc/MTL #228, preset only): 17 genomes newly called (7 Ascomycota, 10 Basidiomycota); --preset v0.8 reproduces v0.8.0 for the 7 Ascomycota genomes.
• Allowed rust genomes (Allow list for suppressed genomes: search the nine rust genomes (#165) #269): 9 genomes, 8 called (14 rustHD loci), 1 uncalled.
• Report-only: resolution_flags on 15,582 loci (Report-only resolution_flags for fragmentary or low-confidence loci #257), two_idiomorphs in 9 Basidiomycota genomes (Report two_idiomorphs for Basidiomycota MAT and redPR (#169) #271); the cassette labels (Cross-family cassette indication with a shared-flank check (#189) #264) changed no call.
• One interplay: Cephaloascus GCA_030575015.1 is called MTL A through the new record; without it the Ogataea routing: a lineage-routed MATog family with all hits withheld leaves no fallback to MATsc/MTL #228 fallback would call MATsc.

Site and version

scripts/build_site_data.py (v081 in RELEASES, DEFAULT_RELEASE = "v081") and its test; regenerated web/intro/data/*; README, web/intro/about.md, landing.html; wheel file names 0.8.0 to 0.8.1 under web/spike/. Site headline: 22,979 NCBI genomes, 20,871 called, 27,103 loci. Related tests pass.

Not done here

• The CHANGELOG header date is still TBD, and CITATION.cff date-released is still 2026-10-31: both are the tag date, to be set by the curator.
• The full pages.yml build (WASM, npm ci, GmG fetch, check_site.py) could not run locally; a build-only dispatch on this branch is being run.
• web/intro/data/idiomorph_labels.json was untracked on main before this work and is not included.

Merge before tagging v0.8.1 so the site built at the tag carries these data.

🤖 Generated with Claude Code

https://claude.ai/code/session_01DDHWqaMRqf5cjFgDKX9CoK