DIA-NN advanced features: phosphoproteomics, plexDIA, immunopeptidomics, DDA, and infrastructure improvements

[ypriverol]

Summary

quantms currently supports label-free DIA analysis with DIA-NN v1.8.1. DIA-NN has evolved significantly through v2.3.2 with features that the community needs: QuantUMS quantification, plexDIA multiplexing, phosphoproteomics with site localization, InfinDIA for immunopeptidomics, and beta DDA support.

This epic tracks the full refactoring across 6 phases, ordered by complexity. The workflow will support all DIA-NN versions (v1.8.1 through v2.3.2) with version-aware feature gating.

Phases

Phase	Category	Feature	Risk	Min DIA-NN	Issue
1	Technical fixes	Container flexibility + config parsing fix + QPX output + test fix	Low	Any	#TBD_1
2	Small additions	QuantUMS + timsTOF + Phospho params	Low	v1.9	#TBD_2
3	New feature	plexDIA (Multiplexed DIA)	Medium	v1.9	#TBD_3
4	New feature	Immunopeptidomics / InfinDIA + Model fine-tuning	Higher	v2.0	#TBD_4
5	Big refactoring	DDA Support via DIA-NN	Medium	v2.3.2	#TBD_5
6	Big refactoring	Parameter Architecture (config gen migration)	Medium	Any	#TBD_6

DIA-NN Version Compatibility Matrix

Feature	Min Version	Key Flags
Basic LFQ DIA	v1.8.1	(current baseline)
QuantUMS	v1.9	--quant-train-runs, --quant-sel-runs
plexDIA	v1.9	--channels, --fixed-mod ... label, --lib-fixed-mod
Synchro-PASEF	v1.9	--quant-tims-sum, --im-window
Parquet output	v1.9.1	(automatic)
Phospho improvements	v1.9.2	--monitor-mod, peptidoform scoring
InfinDIA	v2.0	--pre-select, --pre-select-force
Model fine-tuning	v2.0	--tune-rt, --tune-im
Site report	v2.0	--site-ms1-quant
DDA support (beta)	v2.3.2	--dda

Dependencies

Phase 1 (technical fixes)
  |
  |-->  Phase 2 (small additions — can test immediately)
  |-->  Phase 3 (plexDIA — needs sdrf-pipelines release from Phase 1)
  |-->  Phase 4 (InfinDIA — needs v2.0+ container from Phase 1)
  |-->  Phase 5 (DDA — needs v2.3.2 container from Phase 1)
  |-->  Phase 6 (param refactoring — best done after Phase 3)

Repos Involved

• quantms (Nextflow) — workflow, modules, config
• quantms-utils (Python) — dianncfg.py, diann2mztab.py (being deprecated)
• qpx (Python) — QPX output format, DIA-NN converter
• sdrf-pipelines (Python) — DIA-NN SDRF converter (plexDIA, per-file tolerances)
• quantms-containers (Docker) — DIA-NN v2.3.2 container needed
• pride-ontology (OBO) — mTRAQ/Dimethyl terms (committed, needs release)

Related Existing Issues

• DIANN parameters are generated in multiple places #2 — DIANN parameters are generated in multiple places (Phase 6)
• test_dia runs into THERMORAWFILEPARSER error quantms#661 — test_dia THERMORAWFILEPARSER error (Phase 1)
• Passing external arguments to DIA-NN workflow #3 — Passing external arguments to DIA-NN workflow (Phase 2)
• DIANN phospho data configuration #17 — DIANN phospho data configuration --monitor-mod (Phase 2)
• Support for timsTOF data quantms#440 — Support for timsTOF data (Phase 2)
• Support for Phosphoproteomics datasets. quantms#334 — Support for Phosphoproteomics datasets (Phase 2)

References

• DIA-NN README
• plexDIA paper
• QuantUMS preprint
• Synchro-PASEF paper

[jpfeuffer]

@ypriverol I think existing bugs and new tests with vendor file formats should be highest priority. Let's add bigbio/quantms#672 as first sub issue of this epic.

[ypriverol]

Agree, Im working in that bug already.

[jpfeuffer]

For the rest of the phases we maybe we should do some analysis on how often certain features are used. Do you have some insights from a PRIDE perspective? Maybe we ask at DiaNN?

I also think phase 6 should have higher prio! Right after fixing all existing bugs.
Then there should be:

• docs for how to use own Diann version. Which versions are supported, etc.
• new Diann params for the basic DIA mode

Then at the very end the new Diann modes. I think QuantUMS should be a low-hanging fruit. The rest could be more involved.