Using pmultiqc with DIA-NN Output¶
The diann plugin processes results from DIA-NN for data-independent acquisition (DIA) proteomics experiments.
DIA analysis uses quantmsdiann, not quantms
DIA experiments are processed by the dedicated quantmsdiann pipeline — not quantms. The quantms pipeline handles DDA workflows only (LFQ, TMT/iTRAQ). If you ran DIA-NN through quantmsdiann, your results will contain report.tsv/report.parquet files — not mzTab.
Supported Input Files¶
DIA-NN output does not use mzTab format. The input files are DIA-NN native reports:
| File Pattern | Description | Source |
|---|---|---|
report.tsv or report.parquet |
DIA-NN main report with precursor-level results | DIA-NN / quantmsdiann |
report.log.txt or diannsummary.log |
DIA-NN log file with software version and parameters | DIA-NN / quantmsdiann |
*sdrf.tsv |
SDRF-Proteomics metadata file (optional) | User-provided |
*ms_info.parquet |
Precomputed mzML statistics (optional) | quantms-utils |
No mzTab for DIA
Unlike the quantms DDA pipeline which produces mzTab output, DIA-NN and quantmsdiann produce report.tsv / report.parquet as the primary output. pmultiqc reads these directly — there is no mzTab conversion step for DIA experiments.
Running the Report¶
From quantmsdiann pipeline output¶
# Point to the quantmsdiann results directory
multiqc --diann-plugin /path/to/quantmsdiann/results -o ./report
From standalone DIA-NN¶
# Point to the directory containing report.tsv
multiqc --diann-plugin /path/to/diann/output -o ./report
Common options¶
# Disable hover tooltips for static/publication-ready reports
multiqc --diann-plugin /path/to/results -o ./report --disable-hoverinfo
# Disable large peptide/protein tables (faster for large experiments)
multiqc --diann-plugin /path/to/results -o ./report --disable-table
# Custom contaminant prefix
multiqc --diann-plugin /path/to/results -o ./report --contaminant-affix CONT
DIA-Specific Metrics¶
DIA-NN report files contain precursor-level data with DIA-specific confidence metrics (Q.Value, PG.Q.Value). The following QC sections are generated:
DIA-NN Metadata Table¶
Software version, library type (in-silico predicted vs. empirical), and key search parameters extracted from the log file:
- DIA-NN version
- Spectral library source
- Mass accuracy settings (MS1 and MS2)
- Protein inference method (genes vs. protein groups)
Experimental Design¶
When an SDRF file is present, an experimental design table links DIA raw files to samples, conditions, biological replicates, and fractions.
Identification Statistics¶
- Identification Summary Table — per-sample counts of precursors, peptides, and protein groups passing FDR
- MS/MS Identification Rate — fraction of MS2 events matched to library entries, by run
- Number of Peptides per Protein — unique peptide distribution per protein group
Precursor-Level Metrics¶
- Charge State Distribution — precursor charge frequencies (+1 through +6+)
- Peptide Length Distribution — sequence length histogram
- Peak Intensity Distribution — precursor intensities (log scale)
- Peaks per MS2 Spectrum — fragment ion peak count histogram
MS1 Information (when ms_info.parquet is available)¶
- Total Ion Chromatograms (TIC) — full-gradient signal per sample
- Base Peak Chromatograms (BPC) — most abundant ion at each RT point
- MS1 Peak Count — number of survey scans per run
Long-Trend Analysis¶
Cumulative identification plots across the retention time gradient — useful for detecting chromatographic issues.
DIA vs DDA Output Comparison¶
| Feature | DIA (quantmsdiann) | DDA (quantms) |
|---|---|---|
| Primary output | report.tsv / report.parquet |
mzTab |
| pmultiqc plugin | --diann-plugin |
--quantms-plugin |
| Quantification level | Precursor (MS2) | Feature (MS1) |
| FDR columns | Q.Value, PG.Q.Value | PSM-level PEP |
| Library | In-silico or empirical | Not applicable |
| Pipeline | quantmsdiann | quantms |
File Format Notes¶
- DIA-NN produces both
.tsvand.parquetreport formats. The.parquetformat is preferred for large experiments (>50 samples) as it loads significantly faster. - When running through quantmsdiann, the report is typically named
diann_report.tsvordiann_report.parquet. - Fragment-level filtering uses Q.Value and PG.Q.Value thresholds from the DIA-NN log.
- Bruker
.draw formats are detected automatically; chromatographic metrics may differ slightly from Thermo.rawdata.