nf-core pipeline · classification, illumina, long-reads
nf-core/taxprofiler
Classifies and profiles shotgun metagenomic reads across multiple tools and reference databases simultaneously.
What nf-core/taxprofiler is
nf-core/taxprofiler is a bioinformatics pipeline for taxonomic classification and taxonomic profiling of shotgun metagenomic data. Originally developed by James A. Fellows Yates, Sofia Stamouli, Moritz E. Beber, Lili Andersson-Li, and the nf-core/taxprofiler team, it accommodates both short-read and long-read sequencing technologies.
Metagenomic analyses often suffer from discrepancies between classification algorithms and reference databases. nf-core/taxprofiler addresses this by enabling parallel execution of multiple taxonomic profiling and classification tools across diverse reference databases, producing standardised output tables to streamline cross-tool and cross-database comparisons.
What the pipeline does
The workflow begins with raw read quality inspection via FastQC or falco, followed by optional pre-processing. Short reads can undergo adapter clipping and merging with fastp or AdapterRemoval2, low-complexity and quality filtering with bbduk or PRINSEQ++, and host-read removal via BowTie2. Long reads can be processed with Porechop or Porechop_ABI, filtered with Filtlong or Nanoq, and depleted of host sequences using Minimap2. Metagenomic coverage can be evaluated using Nonpareil, and host-depletion statistics are generated via Samtools.
Pre-processed reads are then passed to one or more taxonomic profilers and classifiers, including Kraken2, MetaPhlAn, MALT, DIAMOND, Centrifuge, Kaiju, mOTUs, KrakenUniq, KMCP, ganon, sylph, Melon, and MetaCache, with optional abundance re-estimation via Bracken. Results are harmonised into standard tabular outputs using Taxpasta, visualisations are generated for selected tools using Krona, and summary quality metrics are collated with MultiQC.
Key facts
- Pipeline framework
- Nextflow (nf-core)
- Latest release
- 2.0.1 (2026-06-15)
- Supported sequencing data
- Shotgun metagenomics (short reads and long reads)
- Input formats
- FASTQ (single-end or paired-end), FASTA, CSV samplesheet and database sheet
- Classification tools
- Kraken2, MetaPhlAn, MALT, DIAMOND, Centrifuge, Kaiju, mOTUs, KrakenUniq, KMCP, ganon, sylph, Melon, MetaCache
- Table standardisation
- Taxpasta
- QC reporting
- FastQC, falco, MultiQC
- Visualisation tools
- Krona
What you can use it for
- Cross-tool taxonomic benchmarking
- Evaluate concordance between multiple profiling algorithms, such as Kraken2, MetaPhlAn, and Centrifuge, on the same metagenomic sequencing library.
- Multi-database screening
- Query metagenomic reads against distinct reference databases in parallel to detect discordant taxonomic assignments across databases.
- Long-read microbiome profiling
- Filter, decontaminate, and profile long-read data such as Oxford Nanopore reads using specialised tools including Filtlong, Nanoq, and Minimap2.
- Host-depleted metagenome assessment
- Remove host contamination with BowTie2 or Minimap2, track removal rates via Samtools, and assess dataset completeness using Nonpareil.
Run it here
The research agent runs nf-core/taxprofiler 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
- Runs multiple taxonomic classifiers and databases in parallel from a single execution.
- Standardises disparate classifier output tables into uniform matrices using Taxpasta.
- Includes dedicated pre-processing and host-depletion pathways tailored separately for short and long reads.
- Generates interactive Krona visualisation charts for Kraken2, Centrifuge, Kaiju, and MALT results.
Watch out for
- Requires users to supply pre-built reference databases or paths formatted for each selected tool.
- Pipeline parameters cannot be supplied via custom Nextflow configuration files and must be passed using the command line or parameter files.
- Krona interactive visualisations are restricted to a subset of the supported profilers.
Questions
Can I run nf-core/taxprofiler without my own cluster?
The research agent runs nf-core/taxprofiler 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/taxprofiler under?
MIT, like every nf-core pipeline.
What sequence input formats does the pipeline support?
The pipeline accepts single-end and paired-end short reads in FASTQ format, as well as long reads in FASTQ or FASTA formats, configured via a CSV samplesheet.
How are reference databases supplied to the tools?
Databases are specified in a dedicated CSV databases sheet that defines the target tool, database name, optional tool arguments, and directory paths or tar.gz archive locations.
Can host reads be removed prior to taxonomic classification?
Yes, host contamination can be removed during pre-processing using BowTie2 for short-read data or Minimap2 for long-read data, with alignment statistics gathered via Samtools.
How does the pipeline allow comparisons across different profiling tools?
It integrates Taxpasta, which standardises and aggregates the diverse output tables produced by various profiling tools into uniform formats.
How to cite it
Stamouli, S., Beber, M. E., Normark, T., Christensen II, T. A., Andersson-Li, L., Borry, M., Jamy, M., nf-core community, & Fellows Yates, J. A. (2023). nf-core/taxprofiler: Highly parallelised and flexible pipeline for metagenomic taxonomic classification and profiling. In bioRxiv (p. 2023.10.20.563221). https://doi.org/10.1101/2023.10.20.563221
Written from nf-core/taxprofiler's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.