Repository navigation
Curate the 80 yeast genomes with a secondary-system MAT locus; Pichia has no reference record #189
Description
Problem
In the v0.7.0 clean run, 80 Saccharomycotina genomes have a locus of a second MAT family at a position different from the main family's locus (suggested_action curate or curate+reroute in results/2026-10-28_idiomorph_crosscall_flags/crosscall_genomes.tsv). They could be a second copy, a silent mating-type cassette (known for Ogataea and Pichia, which switch mating type), or a second cross-call. 29 Ogataea and 27 Pichia genomes carry loci from two systems; the secondary loci are nearly all idiomorph_gene_only hits to MTLalpha1 / MTLalpha2 / MTLA2.
Proposed
- Curate one Ogataea and one Pichia genome in detail (positions, synteny with the MATog and MATkp records, which copy is expressed).
- Pichia has no reference record at all (157 of 158 genomes phylum-fallback): add a curated candidate record (for example Pichia kudriavzevii).
- Decide whether the secondary loci are labelled as a silent cassette, a second system call, or dropped.
Suggested milestone v0.8.0 (not created). Depends on the routing issue (re-run); the duplicate calls are handled by the output-level dedupe rule.
Activity
hyphaltip commentedon Oct 29, 2026
Added after review (2026-10-28)
Protocol: for each of the 80 genomes (
results/2026-10-28_idiomorph_crosscall_flags/crosscall_genomes.tsv) decide which family's locus is real (or both, or neither) and record it in a committed TSV with the evidence (identity, gene content, synteny).Pichia reference: a record is needed from a Pichia species with a known MAT locus (Pichia kudriavzevii is a candidate; curator to name the source and sign off before it enters the DB).
Acceptance: a decision per genome in the TSV; no genome left with two high-confidence loci at one position; the Pichia record passes the DB checks and does not change calls on the v0.7.0 regression panels. Depends on the routing ruling in #188; milestone v0.8.0.
hyphaltip commentedon Oct 29, 2026
From the routing experiment (PR #213): 35 genomes still show two MATsc loci under v1 (Ogataea 12, Pichia 8, Ambrosiozyma 6, others); per-genome table is in results/2026-10-29_yeast_routing/per_genome_calls.tsv. The Pichia reference candidate (P. kudriavzevii) and the Ogataea MATog layout are still to be checked.
yeast routing: Ogataea cassette pairing and Pichia kudriavzevii refer…
hyphaltip commentedon Oct 29, 2026
Reference check (PR #213, ogataea_pichia_reference_check.md): in 71 of 134 Ogataea genomes the two MATog loci are one MATa and one MATalpha locus on different contigs, which reads as the two cassettes of the reference arrangement (69 of the 71 have a locus within 500 bp of a contig start, so assembly breaks are likely; gene content checked by hand in 4 pairs only). Pichia: Douglass et al. 2018 (PMID 30024981) report P. kudriavzevii as heterothallic with no silent loci, so second Pichia loci are more likely the other allele of a heterozygous assembly. Candidate reference: CBS573 MATalpha allele MH260578 (annotated, SUI1-SLA2-MATalpha1-MATalpha2-TGL1-SEC3) and the MATa side on CP028774.1 / NC_042507.1 near 307-313 kb, identified by position and size only. No DB record added; needs a ruling.
detect: indicate MATa and MATalpha cassettes (MATog cassette_report) …
detect: cross-call tie decided by best core identity (near tie keeps …
Merge pull request #213 from stajichlab/yeast-routing
Merge pull request #214 from stajichlab/family-unconfirmed
hyphaltip commentedon Oct 30, 2026
Status for v0.8.0 (2026-10-30).
Pichia (#216, open):
MATpk(scope: P. kudriavzevii, taxid 4909; heterothallic, no silent loci) with two records of CBS 573, the MATa allele (CP028774.1, RefSeq gene models, MATa2 splice checked with SRR6966973) and the MATalpha allele (MH260578.1). Replay of the 158 Pichia genomes: 54 P. kudriavzevii genomes gain calls (42 -> 52 called of 56 in the clean count); the genus was not put in scope because the other 101 Pichia species were calledhighfrom a one-species reference.SLA2-type genera (#220, open, stacked on #216): candidate records, all tier 3 (our own gene models;
needs_review):MATss(Saccharomycopsis schoenii, species scope, homothallic, genetic code 12): two chromosome IV records (the authors' GTF is not public, so models were rebuilt from PRJNA1364152 RNA-seq).MATpf(Petasospora fabianii, species scope): MATa (JOY008) and MATalpha (TBRC 4498), AUGUSTUS and miniprot; no RNA-seq exists.MATbz(Barnettozyma, genus scope): one combined PB4207 record (MATalpha2 a partial fragment).New MATa1/MATalpha2/MATa2 profiles (
results/2026-10-30_mat_hmms) support the gene names (supporting evidence only); they corrected the first S. schoenii drafts.Not done: genera with no record (Ambrosiozyma, Citeromyces, Hansenula, Kuraishia, Pachysolen, Wickerhamomyces); the 80 secondary-system genomes still need the v0.8.0 replay; no record is
acceptedyet (the five new ones await review).Merge pull request #216 from stajichlab/pichia-family
#189: correct W. edaphicus (allelic pair), A. philentoma (second cass…
#189: tier 1 relabelled as suppressed by policy; nucleotide check of …
#189: proposed decisions for the 80 secondary-system yeast genomes (#263
Cross-family cassette indication with a shared-flank check (#189) (#264)
hyphaltip commentedon Nov 5, 2026
Moved to v0.9.0 (curator, 2026-11-05). Merged: the two-tier decisions table for the 80 genomes (#263), the cross-family cassette indication (#264) and the S. pombe-unrelated report flags. Remaining: review a sample of the 70 tier-1 genomes (suppressed by policy, not independently resolved); the nucleotide check of the tier-1 pairs that share sequence (A. oregonensis, S. javanensis, Xingzhongia, Ruyongia) and read depth for W. edaphicus; candidate records for the Ambrosiozyma genus (MATam, deferred) and W. edaphicus (MATwe, proposed in #265).
🤖 Generated with Claude Code