2.16.0 — three named modes, and SHM that stops at Cys104

[mikessh]

⛔ Two breaking changes

1. arda rnaseq run is removed; the regime is the command name.

It was the only pipeline entry point and was used for amplicon libraries as well, with the regime spelled out as four loose flags that do not compose — and --two-pass, the one tuning flag it exposed for four releases, is a loss in both regimes (0.762× bulk, 0.87× IGH amplicon). The Nextflow module already worked around this by naming the regime and building the flag string itself; that workaround is now the CLI.

was	now
arda rnaseq run … --prefilter	arda rnaseq …
arda rnaseq run … --two-pass --fast-segments --v-only-on-segment	arda amplicon …
arda rnaseq run …	arda rnaseq … --exact
arda rnaseq map/correct/assemble	arda map/correct/assemble
arda rnaseq split/reduce/slurm	arda cluster split/reduce/submit
arda split/merge/slurm	arda cluster split-fasta/merge/submit-fasta

arda singlecell is reserved and exits non-zero with a not-implemented message (design in ROADMAP).

⚠ Mode defaults change behaviour: arda rnaseq enables --prefilter (~0.15 % of mapped reads, 122 bulk datasets) and each mode's --ec-mode defaults to its own regime. --exact --ec-mode fast reproduces the pre-2.16.0 output.

2. v_identity / v_mutations / j_mutations are scoped to the FRAMEWORK.

This retracts a guarantee docs/shm.rst printed until 2.14.0 and moves every SHM number arda has published. Segment scoping is not junction exclusion: both germlines' templated tails lie inside the junction, so chew-back was emitted as substitution against a germline that does not template it. Measured on a TRA amplicon — TCRs cannot hypermutate, so every entry was spurious by construction — 1.046 V and 1.658 J entries per read, 86.2 % of the J ones at germline position ≤ 10. On the committed example TRBV28*02's v_identity was 0.8723 and is now 1.0000.

New --shm framework|both|off, and a standalone arda shm that rescopes an existing TSV with no reference and no re-map (it raises on a pre-2.14.0 file rather than copying it through with a success message).

Added

--call-level {allele,gene} — collapses allele-level call splits (Jurkat: TRGJ1*01 64 reads vs TRGJ1*02 140 on one junction, invisible to every error model) and allele-only tie lists. --no-isotype.

Install path, fixed by actually running setup.sh

• arda.__version__ and pyproject.toml's version are two independent literals and they disagreed — a release could ship reporting the previous version. Now pinned by a test.
• --tests ran a pytest that was never installed and swallowed the failure under || true.
• Nothing verified the C++ extensions built (arda silently falls back to pure Python).
• A stale build/ from a deleted venv broke every on-import rebuild.
• pybind11 was unbounded; docs/installation.rst still described a conda setup.sh.

Verification

• pytest tests/unit tests/synthetic tests/realworld — 791 passed, 9 skipped
• ruff==0.15.9 (CI's pin) clean on src/
• make html SPHINXOPTS="-W --keep-going" — zero warnings
• setup.sh --tests re-run end to end from a clean tree
• The SHM regression tests were verified to fail without the fix (7 of 8; the eighth is a no-anchor case that is correct either way)

🤖 Generated with Claude Code