Skip to content

v1.6.5 — Faster UCE recruitment, rescue and assembly

Choose a tag to compare

@GUIBA-EX GUIBA-EX released this 11 Oct 15:52
· 20 commits to master since this release

Highlights

  • Prunes exact-match (MEM) evidence in UCEFilter: a read start stops extending past run-k once its longest possible match can no longer reach the current best seed, and unique run-k strands come from a per-row table instead of a sampled suffix-array walk.
  • Rewrites rescue internal-gap QC on packed 15-mer keys with one pass over reads for both the before and after contigs, and lists rescue read directories once per round instead of once per restored locus.
  • Counts assembler k-mers with 64-bit keys when k ≤ 32, including the k=21 rescue assembly.
  • Leaves scheduling unchanged: stage thread budgets, stage priorities and sample order still follow the GeneMiner2 dispatcher.

Performance

575,435-pair UCE benchmark (2,211 bivalve loci; on-target pairs simulated from real DK40 UCE contigs, off-target pairs from a real WGS library; Apple M2, -p 8, default filter assemble --assembly-mode uce):

Stage v1.6.4 v1.6.5
Total 82.6 s 49.1–49.6 s
First-pass filter 8.3 s 5.2–5.4 s
Rescue round 70.7 s 39.5–40.5 s

Verification

  • All result files were byte-identical to v1.6.4 on the benchmark above and on real phyluce Gallus UCE reads; only run logs and manifests differ.
  • New randomized equivalence tests compare each optimized path with the previous implementation (MEM seeds, counts and strands; per-base gap-support profiles; u64 vs u128 k-mer counts) and were mutation-checked.
  • 283 workspace tests passed; cargo fmt --check and Clippy with warnings denied passed; release build completed.