nf-core pipeline ·
nf-core/metapep
Predicts peptide MHC and HLA binding affinity directly from metagenomic sequences, taxa, and proteins.
What nf-core/metapep is
nf-core/metapep is a bioinformatics pipeline designed to predict epitopes and peptide binding affinity specifically from metagenomic datasets. Developed by Léon Kuchenbecker at the Kohlbacher lab alongside Sabrina Krakau and Till Englert at the Quantitative Biology Center (QBiC), it enables immunological analysis of complex microbial communities.
The pipeline links metagenomic sequencing data with immunoinformatics by translating microbial sequences into candidate peptides and evaluating their potential recognition by the host immune system. It accepts various metagenomic inputs, including assembled contigs, metagenome-assembled genome bins, taxonomic identifiers, or direct protein sequences, bridging the gap between microbiome profiling and antigen discovery.
What the pipeline does
The workflow begins by obtaining protein sequences depending on the provided input format. For taxonomic inputs, it retrieves corresponding protein sequences from NCBI Entrez, whilst for assemblies or genomic bins, it predicts coding sequences and proteins using Prodigal. Peptides are then generated from these proteins.
To enable parallel computation across high-throughput datasets, the generated peptide files are split into smaller chunks while tracking summary statistics. The pipeline then evaluates epitope binding for specified alleles using one of three integrated prediction tools: SYFPEITHI, MHCflurry, or MHCnuggets. Finally, it compiles the results and produces summary plots.
Key facts
- Pipeline engine
- Nextflow DSL2
- Latest release
- 1.0.0 (2024-11-22)
- Input formats
- Taxa, assemblies, bins, proteins
- Gene prediction tool
- Prodigal
- Epitope prediction engines
- SYFPEITHI, MHCflurry, MHCnuggets
- Sequence retrieval source
- NCBI Entrez
- Software containerisation
- Docker, Singularity
What you can use it for
- Metagenomic antigen discovery
- Identify potential antigenic peptides from de novo metagenomic assemblies or binned microbial genomes.
- Taxon-based epitope profiling
- Retrieve reference proteins for specified microbial taxa from Entrez to predict host MHC-binding epitopes.
- Microbiome immunogenicity assessment
- Screen vast sets of microbial protein sequences against specific human leukocyte antigen (HLA) or MHC alleles using MHCflurry, MHCnuggets, or SYFPEITHI.
Run it here
The research agent runs nf-core/metapep 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
- Accepts four distinct stages of metagenomic input: taxa, assembled contigs, metagenomic bins, or direct protein sequences.
- Splits peptide files into chunks to allow parallelised epitope prediction across compute clusters.
- Provides multiple prediction engine options, including SYFPEITHI, MHCflurry, and MHCnuggets.
- Implements process-level containerisation using Docker or Singularity for reproducibility.
Watch out for
- Raw sequencing reads are not directly accepted; inputs must be assembled contigs, bins, taxa, or proteins.
- Epitope prediction is restricted to alleles and models supported by SYFPEITHI, MHCflurry, or MHCnuggets.
- Taxon-based processing relies on external internet connectivity to download protein records from NCBI Entrez.
Questions
Can I run nf-core/metapep without my own cluster?
The research agent runs nf-core/metapep 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/metapep under?
MIT, like every nf-core pipeline.
What input data does the pipeline accept?
The pipeline accepts four primary input types: microbial taxa, assembled nucleotide sequences, metagenomic bins, and raw protein sequences.
Which tools are available for epitope prediction?
Users can predict epitope binding affinities using SYFPEITHI, MHCflurry, or MHCnuggets.
How does the pipeline handle gene prediction for unannotated metagenomic assemblies?
For inputs supplied as assemblies or bins, the pipeline runs Prodigal to predict coding genes and produce protein sequences.
Can the pipeline process large peptide datasets efficiently?
Yes, peptide files are automatically broken into smaller chunks for parallel prediction across compute infrastructure, and summary statistics are tracked.
How to cite it
10.5281/zenodo.14202996
Written from nf-core/metapep's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.