Accelerating proteomics in nf-core: from modular building blocks to community workflows

[julianu]

Title

Accelerating proteomics in nf-core: from modular building blocks to community workflows

Abstract

nf-core is the reference community and repository for curated Nextflow bioinformatics pipelines. While widely adopted in genomics and transcriptomics, mass spectrometry-based proteomics remains still underrepresented. Established pipelines (quantms for quantitative proteomics, mhcquant for immunopeptidomics) are now complemented by the new in-development pipelines msproteomics (integrating DIA-NN and FragPipe) and mspepid (DDA identification and rescoring). A dedicated nf-core Mass Spec Proteomics SIG coordinates these efforts. However, key gaps like e.g. dedicated quality control and metaproteomics workflows remain, and only a handful of proteomics tools exist as reusable nf-core modules, constraining new pipeline development.
This hackathon advances proteomics integration into nf-core on several fronts. We will implement additional tool modules spanning preprocessing, identification, and quantification, and promote interchangeability across pipelines through community-standard exchange formats (mzML, mzTab, SDRF). If time and participants allow, we will work on the implementation of a quality control workflow that exports standardized mzQC and leverages MultiQC for visualization as well as a metaproteomics pipeline. The project is ideally suited as a community effort: it is inherently modular, lets participants of mixed experience contribute in parallel on independent building blocks, builds directly on widely used open-source tools and standards, and feeds back into a living, maintained ecosystem that lives on well beyond the meeting.

Project Plan

The project is broken into independent building blocks that can be tackled in parallel by sub-teams of 2-3 people, then integrated in a plenum. This makes it well suited to a ~4-day meeting with up to 10 participants of mixed experience.

Kickoff and onboarding (early)

• Short, hands-on training on Nextflow and the nf-core module system, the development workflow, and nf-test-based module testing, so participants new to Nextflow can contribute.
• Brief introduction to proteomics data and the relevant community standards (SDRF, mzML, mzTab, mzQC) by domain experts on the team.
• Collectively agree on conventions for harmonized parameters, inputs, and outputs across modules, and on clear interface definitions so the parallel sub-units integrate cleanly.

Parallel work packages

• Modules. Implement and extend nf-core modules across preprocessing, identification, and quantification, including additional search-engine and quantification-algorithm wrappers, a pmultiqc module for QC reporting, or other tools of interest.
• Standardization / interchangeability. Adapt modules to consume and emit community-standard formats (mzML, mzTab) and adopt SDRF as a unified experimental-design input, demonstrating that search engines or quantification algorithms can be swapped within a pipeline without major adaptation.
• Metaproteomics workflow (stretch goal, if capacity allows). Begin an initial metaproteomics workflow reusing the identification and quantification modules already available. This is not fixed yet and the topic can still be discussed by the group.

Integration and wrap-up (late)

• Align interfaces across work packages, add tests and documentation, and open pull requests against the official nf-core repositories.
• Summarize outcomes and define remaining tasks.

Follow-up strategy

• Completing and polishing the initiated workflow continues after the meeting. Sustained development is ensured through the team's ongoing nf-core pipeline maintenance and coordination via the nf-core Mass Spec Proteomics SIG, with natural continuation at recurring nf-core community hackathons. All code is contributed upstream under a permissive open-source license, in line with the EuBIC-MS FAIR principles.

Technical Details

Programming languages / frameworks

• Nextflow (DSL2 / Groovy) for workflows and modules; Python and bash for tooling and tool wrappers.
• nf-core tooling (nf-core/tools), nf-test for module/pipeline testing, and containerization via Docker / Singularity / Conda for reproducibility.

Existing software featured / built upon

• nf-core proteomics pipelines: quantms, mhcquant, as well as the in-development msproteomics, and mspepid.
• QC starting points: McQuaC (mpc-bioinformatics/McQuaC) and QCloud2.
• Underlying tools and reporting: OpenMS, Comet, DIA-NN, FragPipe, ThermoRawFileParser, and MultiQC.

Standards

• HUPO-PSI formats mzML, mzTab, and mzQC; SDRF-Proteomics as the unified experimental-design input.

Datasets

• Small test datasets will be assembled ahead of the meeting for module continuous-integration testing and distributed to participants.
• Publicly available, SDRF-annotated DDA and DIA datasets from PRIDE / ProteomeXchange can be used for larger testing and benchmarking.

Contact information

• Julian Uszkoreit, julian.uszkoreit@rub.de, Ruhr University Bochum
• Henry Webel, heweb@dtu.dk, Technical University of Denmark
• Jonas Scheid, jonas.scheid@uni-tuebingen.de, University and University Hospital Tübingen