Batrachochytrium read depth, heterozygosity and parent-of-origin analysis - #352

What

Exploratory read-depth, heterozygosity, parent-of-origin, idiomorph and chromosome-depth analysis of public Batrachochytrium dendrobatidis WGS (69 isolates: 52 Brazil runs, 17 other countries) against two references (JEL423, UCR_BdCLFT044). No detector behaviour change.

• scripts/bd_coverage_isolates.slurm, bd_summarize.py: per-run alignment, depth windows, SNP allele balance.
• scripts/bd_parent_of_origin.py: JEL423-vs-CLFT044 divergent sites scored per isolate in 50 kb windows.
• scripts/bd_idiomorph_scan.py: cross-isolate depth scan for a mating-type idiomorph.
• scripts/bd_snp_pca.py: isolate-level SNP ordination.
• scripts/bd_site_genotype.slurm, bd_site_allele_counts.py, bd_site_summary.py: depth-controlled read counts at the divergent sites.
• scripts/bd_hetrich_check.py: hybrid-versus-aneuploid check for chosen isolates.
• scripts/bd_aneuploidy.py, bd_aneuploidy_af.py: isolate-specific chromosome depth changes and the allele balance on the flagged chromosomes.
• scripts/bd_report.py and the report: datasets/reads/sketch_store/reports/batrachochytrium_dendrobatidis/coverage_heterozygosity/ (report.md, idiomorph_scan.md, aneuploidy.md, ancestry_pca.md, site_genotypes.md, tables, figure).
• tests/test_bd_coverage.py: 19 tests pass. Full suite on the head commit 93ae72e (job 32094324): 2030 passed, 8 skipped.

Results (see the report pages for limits)

• Each isolate gives about 1,400-1,900 SNPs/Mb against its own reference, mostly het-like (diploid; repeats not separated). Het fraction alone cannot call a hybrid.
• Parent-of-origin classes: Brazil 13 CLFT-type, 37 JEL-type, 2 HET-rich (CLFT160, CLFT039); other countries 15 JEL-type, 2 CLFT-type (Switzerland, South Korea; low recovery, possibly a third lineage; not checked). HET-rich is not a hybrid call; the 0.5 cutoff comes from two control isolates and is not validated.
• Idiomorph depth scan: negative (71 and 70 candidate regions; most are divergence artefacts; 3 and 5 pass a low-het filter; no mating-related gene; the rotation null is uninformative).
• Chromosome depth (20 JEL423 chromosomes): CP161924.1, CP161925.1 and CP161938.1 deviate in every isolate (reference-specific); isolate-specific flags: 48 in 28 isolates. Of 24 gains, 15 show the allele-balance shift expected for an extra chromosome copy (1/2 -> 1/3), clearest in L2203, KRBOOR_331 and RTP5 (depth 71-175x); 9 do not (CLFT065 CP161936.1 shows the opposite). 4 of 7 losses show lower het density. Baseline ploidy is not called: the genome-wide allele-balance label is confounded with depth.
• The two HET-rich isolates (CLFT160, CLFT039): heterozygous windows are the majority on almost every large scaffold (median 0.70 and 0.72), JELHOM windows almost absent, depth normal apart from two gain flags in CLFT160 that sit on scaffolds with ordinary HET shares. The heterozygosity is genome-wide, not an aneuploidy effect; two CLFT160 scaffolds look CLFT-homozygous (possible loss of heterozygosity). A third lineage sharing alleles with both references would give the same pattern.
• SNP ordination (PCA, ancestry_pca.md): PC1 (47%) is the CLFT-type/JEL-type axis; PC4 separates KRBOOR_331 and the Swiss isolate (a third-lineage-like signal, high depth). CLFT160 (26x) lies midway on PC1 and off the PC4 axis, beyond JEL-type isolates of similar depth: it fits a CLFT/JEL mixture and not the third lineage. CLFT039 (15x) cannot be read from the ordination (depth and PC1 correlate, Spearman -0.78 in JEL-type); the depth-controlled counts below place it.
• Depth-controlled read counts (site_genotypes.md): nine isolates re-mapped and downsampled to 12x, reads counted at the 172,706 divergent sites. CLFT160 and CLFT039 are 41% CLFT-only, 44 to 47% heterozygous and 12 to 15% JEL-only; the JEL-type controls are 19 to 21%, 15 to 19% and 60 to 66%; CLFT044 is 89%, 10.5% and 0.5%. CLFT039 (15x) is heterozygous on every one of the 14 large scaffolds (0.36 to 0.56) while no control exceeds 0.30 on any scaffold, so its position is not a low-depth effect. CLFT160 is patchy (0.08 to 0.73). CLFT165 is a partial mixture (median 0.24).
• A first version counted sites without a SNP row as CLFT044 allele; the JEL423 control then scored only 0.57, so it was dropped (reason in the report).

Not done

• Three runs excluded: SRR6375442 (SRA Lite only), SRR13660529 and ERR14883802 (read-count mismatch with ENA).
• No tree or admixture model. No third-lineage reference assembly: NCBI has 9 B. dendrobatidis assemblies; the only possible third-lineage candidate, India T46-F (GCA_982397245.1), is contig-level (N50 0.13 Mb), its lineage is not stated, and it was not used (see site_genotypes.md).
• The depth-controlled site counts cover 9 isolates only (the two HET-rich isolates, CLFT165, CLFT044 and five JEL-type controls).
• The HPCC result folder results/2026-11-13_bd_coverage/ is not committed. Jobs 32070127/28, 32070630/31, 32089404, 32093759 (site counts).

The curator merges.

🤖 Generated with Claude Code