BioFastq-A v2.3.1

[DilaDeniz]

v2.3.1 — 2026-06-24

Summary

v2.3.1 is a correctness and stability release. It fixes a long-read accuracy
regression and a paired-end pairing bug introduced by v2.3.0's performance
work, fixes a packaging issue that could crash already-published Bioconda
binaries on some CPUs, ships a WebAssembly build for running BioFastq-A in
the browser, and removes the last hardcoded version strings from
report/TUI output.

New Features

WebAssembly Browser Build

BioFastq-A now compiles to a wasm/ target with the same analysis pipeline
as the native CLI — adapter detection, full metrics, HTML report generation —
running entirely client-side via a drag-and-drop browser UI. No install,
no upload: FASTQ files never leave the browser.

Performance

Change	Effect
zlib-rs backend for flate2 gzip I/O	Faster .fastq.gz decompression on the read path
parse_illumina_tile iterator rewrite	Removes a per-read Vec allocation during per-tile quality tracking

Bug Fixes

• Critical (packaging): the build was pinned to target-cpu=native, baking
  in the exact CPU instruction set of whatever machine compiled the binary.
  Bioconda builds once per platform and ships that same binary to everyone, so
  any user with a CPU that didn't match the build server's feature set got an
  instant crash (SIGILL) on startup with no explanation. The flag served no
  purpose — BioFastq-A's SIMD (via memchr) already does runtime
  CPU-feature detection — so it's removed with zero performance cost,
  confirmed via benchmark (554K reads/s on 200K reads before vs. after).
• Critical: N50/N90 were wrong for any read ≥2000bp (long-read/ONT/PacBio
  data) — v2.3.0's flat-array length histogram collapsed all such reads into
  one bucket keyed by 2000, corrupting the cumulative-length calculation.
  Now tracked exactly via an overflow map, with no performance cost for
  ordinary short-read data.
• Major: paired-end mode (--compare/R1+R2) could permanently shift every
  read pair by one position after a single malformed R2 record, silently
  cross-pairing the rest of the file. R2 retries are now isolated from the R1
  cursor.
• FASTA input (no quality scores) showed a misleading Pass/Fail for
  "Per-base quality" and "Per-sequence quality" instead of omitting modules
  that don't apply.
• Hardcoded version strings in the HTML report and TUI title bar (v2.0)
  could drift from Cargo.toml; both now derive from it at compile time.

Installation

Bioconda

conda install -c bioconda biofastq-a

Build from source

cargo install --path

Have fun using!

by Dila Deniz

---

This discussion was created from the release BioFastq-A v2.3.1.