Research & applications
Papers describing DIA-NN’s algorithms and research workflows that use the software.
Core algorithms
The original DIA-NN paper introduces neural-network scoring to distinguish peptide signals from interference, together with signal correction strategies for quantification.
Establishes the computational foundations for identifying and quantifying peptides in complex DIA data, including high-throughput experiments.
QuantUMS
QuantUMS learns from signals within an experiment to correct and combine quantitative measurements. It improves precision, reduces interference-related ratio compression and estimates the quality of peptide and protein quantities.
Describes DIA-NN’s quantification module, its precision and accuracy modes, and quality-based filtering for downstream statistical analysis.
PTMs
A DIA ubiquitinomics workflow combines optimised sample preparation with modified-peptide scoring in DIA-NN. The study follows changes in ubiquitination and protein abundance after USP7 inhibition.
Introduces additional scoring for confident identification of modified peptides, including diglycine-modified peptides in ubiquitinomics.
Scanning SWATH
Scanning SWATH adds a precursor-mass dimension by continuously moving the quadrupole isolation window. The paper couples this acquisition strategy to DIA-NN analysis for short-gradient proteomics.
DIA-NN scores how fragment signals vary with quadrupole position, helping assign fragments to the correct precursor in complex data.
dia-PASEF
Describes ion-mobility-aware processing for dia-PASEF, together with a FragPipe workflow for building compatible spectral libraries. The study evaluates these methods with low-input and fast-gradient data.
DIA-NN combines two-dimensional peak picking, ion-mobility-aware extraction and neural-network scoring. Ion mobility also helps exclude interfering signals during quantification.
Slice-PASEF
Introduces an acquisition strategy that uses trapped ion mobility separation to increase the fraction of ions selected for fragmentation, targeting sensitive proteomics at low sample amounts.
Describes the dedicated Slice-PASEF processing module implemented in DIA-NN and demonstrates its use with the acquisition method.
plexDIA
plexDIA combines non-isobaric sample labelling with DIA to analyse multiple samples in one acquisition, including low-input and single-cell experiments.
Describes DIA-NN’s multiplexed analysis: transferring identification evidence between channels, estimating channel-specific confidence and using fragment ratios for relative quantification.
CysQuant
CysQuant labels reduced and reversibly oxidised cysteines with light and heavy iodoacetamide isotopologues, quantifying cysteine oxidation degrees and protein abundance in the same experiment, with DDA or DIA acquisition.
The study uses DIA-NN’s plexDIA module with in silico predicted spectral libraries to quantify the light and heavy cysteine labels together.
SynchroSep-MS
SynchroSep-MS introduces several LC columns, each carrying an independent sample, into a single mass spectrometer inlet. A controlled retention-time offset between injections gives each sample a distinct elution profile, for label-free multiplexed proteome analysis.
The study uses a modified DIA-NN workflow to process the parallelised data, accounting for the retention-time offsets between samples.
quantms
quantms is an open-source, Nextflow-based pipeline for massively parallel proteomics data analysis in the cloud and on distributed architectures. It was used to reanalyse 83 public ProteomeXchange datasets comprising 29,354 instrument files.
The quantms DIA workflow runs DIA-NN across compute nodes, distributing its analysis steps for large-scale processing. This workflow is now developed as quantmsdiann.
Explore DIA-NN’s capabilities and analysis workflows.
Explore the software