Issues

• The benchmark harness has no tests, and has twice silently changed what it measures

  wdecoster/STRdust#38 · Filter by author wdecoster opened on Sep 13, 2026

• A second, unidentified source of run-to-run non-determinism

  wdecoster/STRdust#37 · Filter by author wdecoster opened on Sep 13, 2026

• remove_outliers trims symmetrically on an asymmetric distribution, and is untested at length-variable loci

  wdecoster/STRdust#36 · Filter by author wdecoster opened on Sep 13, 2026

• POA consensus: path selection is not support-aware (and report the endpoint bug upstream)

  wdecoster/STRdust#35 · Filter by author wdecoster opened on Sep 13, 2026

• Synthetic expansion ladder: at what size does sizing break, and does it fail loudly?

  wdecoster/STRdust#34 · Filter by author wdecoster opened on Sep 13, 2026

• DBSCAN unphased clustering is mistuned for ordinary heterozygous loci

  wdecoster/STRdust#32 · Filter by author wdecoster opened on Sep 13, 2026

• STRdust has no tolerance for read noise, and over-calls alternate alleles at short repeats

  wdecoster/STRdust#31 · Filter by author wdecoster opened on Aug 29, 2026
  1