lehtiolab/nf-diann

A Data-Independent Analysis pipeline based on DIA-NN

This Nextflow workflow uses DIA-NN to analyze mass spectrometry proteomics data acquired using data-independent analysis (DIA). It runs on multiple compute infrastructures in a portable manner, and scales horizontally over e.g. HPC nodes by parallelizing jobs.

How to run

• install Nextflow
• install Docker or Singularity
• run pipeline, e.g.:

nextflow run lehtiolab/nf-diann -profile docker -resume \
    --input /path/to/input_definition.txt \
    --tdb /path/to/proteins.fa \
    --ms1acc 10 --ms2acc 10

One can pass a library using --library libfile.speclib or --libfile.parquet, where the pipeline assumes the former to be a predicted in-silico library (i.e. only used fasta as input data), which will be used to generate an empirical library using raw spectra data input files. The second invocation using a parquet file is assumed to be an empirical library, which will go directly to the next step (search raw files).

flowchart TD
    V@{shape: database, label: "db.speclib"}
    V@{shape: database, label: "db.parquet"}

    A[(db.fasta)] --> B@{shape: rect, label: "Predict in-silico
    library"}
    T@{shape: docs, label: "spectra.raw"}
    U@{shape: docs, label: "spectra.quant"}

    B --> C[(db.speclib)]
    C -->D@{shape: procs, label: "Create empirical
            library"}
    D --> G@{shape: rect, label: "Combine empirical
             libraries"}
    G --> E[(db.parquet)]
    E --> F@{shape: procs, label: "Search empirical
    library"}
    H@{shape: docs, label: "spectra.raw"} --> D
    H --> F
    F --> I@{shape: docs, label: "spectra.quant"}
    I --> J@{shape: rect, label: "Train quantUMS"}
    I --> K@{shape: rect, label: "Combine output"}
    J --> |quant parameters| K
    K --> L@{shape: docs, label: "Report files"}

    style L fill: #62D97E
    style E fill: #62D97E
    style I fill: #62D97E
    style C fill: #62D97E
    style T stroke-dasharray: 8
    style U stroke-dasharray: 8
    style V stroke-dasharray: 8

Loading