feat(protein): add the spectra mode for mass spectrometry QC

rustqc protein spectra reads mzML and reports per-level spectrum and peak
counts, total ion current, the retention time range, the fragmentation ratio
and the precursor charge distribution.

mzdata does the reading, behind a `proteomics` cargo feature that is on by
default. Building with --no-default-features drops the mode from the help
entirely rather than offering it and then failing, and drops mzdata's
transitive packages with it. Both configurations are tested.

Validated against pyteomics 5.0.1 on mzdata's own test file: two independent
readers agree on all 48 spectra, 305213 peaks, the per-level minima and
maxima, the retention time bounds and the precursor m/z range. Total ion
current is compared with a relative tolerance rather than exactly, because
the intensities are 32-bit in the file and the two readers accumulate them at
different precision.

Two reporting decisions worth noting. Total ion current is summed from the
peaks actually present rather than read from the spectrum header, so a
profile spectrum that has since been centroided reports what the file now
holds. And a run with no survey scans reports its fragmentation ratio as NA
rather than 0, since 0 would read as "no fragmentation" when the truth is
"nothing to divide by".

The fixture annotates no precursor charge states, which the run warns about
and the report shows as `without_charge`, rather than quietly reporting an
empty charge distribution.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

Author: BenjaminDEMAILLE

AGENTS.md

@@ -95,6 +95,10 @@ src/
       stats.rs      — Length statistics reproducing seqkit stats -a
       defects.rs    — Internal stops, non-standard residues, duplicates
       output.rs     — The seqkit-compatible table and the RustQC report
+    spectra/        — behind the `proteomics` cargo feature
+      mod.rs        — mzML reading via mzdata
+      metrics.rs    — Per-run and per-level spectrum metrics
+      output.rs     — The run report
   rna/
     mod.rs          — Re-exports the RNA submodules (dupradar, featurecounts, rseqc, qualimap)
                       and re-exports the shared ones from `common` for compatibility

CHANGELOG.md

@@ -18,6 +18,10 @@
   length statistics reproducing `seqkit stats -a` exactly, plus amino acid
   composition and defect detection (internal stop codons, non-standard
   residues, duplicate sequences and identifiers) that seqkit does not report.
+- New `rustqc protein spectra` mode: mass spectrometry run QC from mzML, with
+  per-level spectrum and peak counts, total ion current, retention time range
+  and precursor charge distribution. Behind the `proteomics` cargo feature,
+  which is on by default; `--no-default-features` drops it along with mzdata.
 
 ### Changed
 

Cargo.lock

@@ -21,7 +21,17 @@ path = "src/lib.rs"
 name = "rustqc"
 path = "src/main.rs"
 
+[features]
+default = ["proteomics"]
+# Mass spectrometry support for `rustqc protein spectra`. Enabled by default;
+# disable with `--no-default-features` to drop mzdata and its transitive
+# packages, which is worth doing when only the sequencing pipelines are needed.
+proteomics = ["dep:mzdata"]
+
 [dependencies]
+# mzML reading for the protein spectra mode (optional, see the `proteomics` feature)
+mzdata = { version = "0.66", default-features = false, features = ["mzml"], optional = true }
+
 # CLI argument parsing
 clap = { version = "4", features = ["derive", "env"] }