nf-core pipeline · active-modules, complex-diseases, disease-mechanisms
nf-core/diseasemodulediscovery
A Nextflow pipeline for identifying and evaluating network-based disease modules from molecular interaction data.
What nf-core/diseasemodulediscovery is
nf-core/diseasemodulediscovery is a bioinformatics pipeline designed for hypothesis generation in network medicine. Developed and maintained by the RePo4EU consortium, it identifies active disease modules by investigating the local molecular neighbourhoods of known disease-associated seed genes or proteins within molecular interaction networks.
Understanding the network context of disease-associated molecules helps elucidate underlying pathological mechanisms and identify candidate therapeutic targets. The pipeline runs multiple module detection algorithms and provides a unified comparative framework to evaluate and benchmark the resulting modules.
In addition to module inference and comparative evaluation, the workflow annotates identified modules with biological annotations such as targeting drugs, adverse effects, and associated disorders from NeDRexDB, exports networks to the Drugst.One platform, and carries out drug prioritisation.
What the pipeline does
The workflow takes user-supplied seed lists and a molecular interaction network to discover disease modules. By default, it executes six module identification algorithms: DOMINO, DIAMOnD, ROBUST, ROBUST (bias-aware), 1st Neighbors, and Random Walk with Restart (RWR). Network visualisations are rendered using graph-tool and pyvis, and modules are annotated with drug targets, side effects, disorders, and cellular localisations retrieved from NeDRexDB before conversion into BioPAX format.
Following module inference, the pipeline benchmarks the results through multiple evaluation modules. It performs over-representation analysis via g:Profiler, functional coherence testing via DIGEST, network topological analysis via graph-tool, and pairwise module overlap calculations. Optional flags enable seed-set perturbation analysis (to test robustness and seed rediscovery) and network perturbation analysis. Module data can also be sent to the Drugst.One API for drug prioritisation, with overall pipeline performance and diagnostics summarised in a MultiQC report.
Key facts
- Workflow framework
- Nextflow (nf-core)
- Core inputs
- Seed genes or proteins (--seeds) and interaction network (--network)
- Module detection algorithms
- DOMINO, DIAMOnD, ROBUST, ROBUST (bias-aware), 1st Neighbors, Random Walk with Restart (RWR)
- Evaluation tools
- g:Profiler, DIGEST, graph-tool
- Annotation source
- NeDRexDB
- External integrations
- Drugst.One API, BioPAX export
- Summary reporting
- MultiQC
What you can use it for
- Disease module identification
- Discover functional sub-networks and disease modules around known disease-associated seed genes across a molecular interaction network.
- Cross-algorithm benchmark and comparison
- Compare modules generated by DOMINO, DIAMOnD, ROBUST, 1st Neighbors, and Random Walk with Restart using topological analysis and module overlap metrics.
- Drug repurposing and prioritization
- Identify candidate therapeutic compounds for a disease module using NeDRexDB annotations and the Drugst.One API.
- Functional coherence and pathway enrichment
- Assess the biological consistency of discovered modules using DIGEST functional coherence testing and g:Profiler over-representation analysis.
- Module robustness evaluation
- Test module stability and seed recovery rates under seed set and network perturbation experiments.
Run it here
The research agent runs nf-core/diseasemodulediscovery with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
Strengths and limits
Good at
- Executes multiple disease module discovery methods simultaneously under a standardised configuration.
- Incorporates rigorous evaluation layers, including functional coherence, topological characteristics, and perturbation testing.
- Automates downstream functional enrichment, NeDRexDB biological annotation, and Drugst.One interactive export.
- Delivers an aggregated quality and metric report using MultiQC.
Watch out for
- Certain Docker container images are natively compiled for amd64 architecture and do not support arm architectures natively.
- Optimal execution requires a Linux environment, necessitating compatibility tools such as OrbStack on macOS or WSL on Windows.
Questions
Can I run nf-core/diseasemodulediscovery without my own cluster?
The research agent runs nf-core/diseasemodulediscovery with Nextflow on its own machine — up to 32 cores, 256 GiB of memory and 24 hours per run, which covers the test profile and real single-sample runs. It builds the samplesheet, checks it against the pipeline's schema and reads the results back. When a run is too big for one machine, it gives you the exact command for your own cluster instead.
What licence is nf-core/diseasemodulediscovery under?
MIT, like every nf-core pipeline.
What inputs are mandatory to run the pipeline?
The workflow requires a set of disease-associated seeds using the --seeds parameter and a molecular interaction network provided through the --network parameter.
Which module detection algorithms are executed by default?
By default, the pipeline runs DOMINO, DIAMOnD, ROBUST, ROBUST (bias-aware), 1st Neighbors, and Random Walk with Restart (RWR).
How are perturbation-based analyses enabled?
Seed set perturbation is activated with the --run_seed_perturbation flag, while network perturbation is enabled using --run_network_perturbation.
What formats and external platforms are supported for module visualisation?
The pipeline produces visualisations with graph-tool and pyvis, exports biological data to BioPAX format, and can export networks directly to the Drugst.One web visualisation platform.
Can failing or non-essential pipeline steps be omitted?
Yes, most analytical steps are non-essential and can be skipped if a specific process fails during execution.
How to cite it
Johannes Kersting, Chloé Bucheron, Lisa M Spindler, Joaquim Aguirre-Plans, Quirin Manz, Tanja Pock, Mo Tan, Fernando M Delgado-Chaves, Cristian Nogales, Harald H H W Schmidt, Jörg Menche, Andreas Maier, Jan Baumbach, Emre Guney, Markus List, Inferring and evaluating network medicine-based disease modules with nextflow, Bioinformatics, Volume 42, Issue Supplement_1, July 2026, btag223, doi: 10.1093/bioinformatics/btag223
Written from nf-core/diseasemodulediscovery's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.