nf-core pipeline · database, database-builder, metagenomic-profiling
nf-core/createtaxdb
Automated and parallelised generation of reference databases for multiple taxonomic profilers and classifiers.
What nf-core/createtaxdb is
nf-core/createtaxdb is a bioinformatics pipeline developed by James A. Fellows Yates, Sam Wilkinson, Alexander Ramos Díaz, Lili Andersson-Li, and the nf-core community, with extensions from Moritz Beber and Haris Spyridis. It automates and parallelises the construction of custom reference databases for multiple metagenomic classifiers and profilers starting from a single, shared collection of reference genomes.
Metagenomic classification tools often require custom reference databases built according to specific file formats and taxonomic indexing standards. Building these indices independently across multiple tools can be time-consuming and prone to inconsistencies. nf-core/createtaxdb solves this by standardising inputs across both nucleotide- and protein-based profilers, allowing researchers to generate comparable and reproducible databases concurrently.
The pipeline was designed as a companion tool to nf-core/taxprofiler, providing ready-to-use databases and compatibility files for downstream metagenomic profiling analyses, but it can also be run as a standalone utility for general database creation.
What the pipeline does
The pipeline takes a comma-separated values (CSV) sample sheet containing organism identifiers, NCBI taxonomic IDs (taxids), and paths to FASTA files (either nucleotide, amino acid, or both, with optional gzip compression), alongside relevant taxonomy mapping files such as accession2taxid, nodes.dmp, and names.dmp. It first validates and prepares the input FASTA files for database building.
Following preparation, the workflow invokes the native indexing and database build tools selected by the user in parallel. It supports 13 profiling and classification tools: Bracken, Centrifuge, Centrifuger, DIAMOND, ganon, Kaiju, KMCP, Kraken2, KrakenUniq, MALT, sourmash, sylph, and MetaCache. Each tool can receive custom build arguments, and database files are placed into designated subdirectories within the specified output directory. Optionally, the pipeline generates downstream sample sheets formatted for direct use in pipelines such as nf-core/taxprofiler.
Key facts
- Latest release
- 3.1.0 (2026-07-03)
- Primary inputs
- CSV sample sheet (id, taxid, fasta_dna, fasta_aa) and taxonomy mapping files (e.g. nodes.dmp, names.dmp, accession2taxid)
- Supported profilers
- Bracken, Centrifuge, Centrifuger, DIAMOND, ganon, Kaiju, KMCP, Kraken2, KrakenUniq, MALT, MetaCache, sourmash, sylph
- Pipeline outputs
- Database directories per selected profiler and optional downstream sample sheets
- Original authors
- James A. Fellows Yates, Sam Wilkinson, Alexander Ramos Díaz, Lili Andersson-Li
What you can use it for
- Multi-tool metagenomic database preparation
- Building indices for several metagenomic profiling algorithms (such as Kraken2, ganon, and DIAMOND) from identical reference FASTA genomes in a single automated run.
- Protein-based taxonomic classifier indexing
- Generating search databases from amino acid FASTA sequences for protein classifiers like DIAMOND and Kaiju.
- Downstream workflow integration
- Creating tailored reference databases and automated sample sheets intended directly for use with nf-core/taxprofiler.
Run it here
The research agent runs nf-core/createtaxdb 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
- Builds databases for up to 13 different taxonomic profilers simultaneously from a single genome list.
- Handles both nucleotide (DNA) and protein (amino acid) reference FASTA inputs, including gzipped archives.
- Supports custom build flags passed directly to individual underlying database creation tools.
- Generates downstream configuration sample sheets automatically for companion pipelines like nf-core/taxprofiler.
Watch out for
- Requires structured input taxonomy files (such as nodes.dmp, names.dmp, or accession2taxid tables) corresponding to the input genomes.
- Pipeline options and parameters cannot be supplied via custom configuration files using Nextflow's -c option and must be provided via the command line or -params-file.
Questions
Can I run nf-core/createtaxdb without my own cluster?
The research agent runs nf-core/createtaxdb 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/createtaxdb under?
MIT, like every nf-core pipeline.
Can I supply both nucleotide and amino acid FASTA sequences?
Yes. The input CSV accepts paths for both fasta_dna and fasta_aa on the same row, or either individually depending on what sequence types are available.
Which classifiers and profilers are supported by nf-core/createtaxdb?
The pipeline can build databases for Bracken, Centrifuge, Centrifuger, DIAMOND, ganon, Kaiju, KMCP, Kraken2, KrakenUniq, MALT, MetaCache, sourmash, and sylph.
Can I pass custom flags to individual database building tools?
Yes. Dedicated parameters exist for tool-specific build options, such as --kraken2_build_options, --ganon_build_options, and --diamond_build_options.
How does the pipeline integrate with nf-core/taxprofiler?
Using the --generate_downstream_samplesheets option, the pipeline creates pre-prepared database specification sheets formatted specifically for nf-core/taxprofiler.
How to cite it
10.5281/zenodo.15696114
Written from nf-core/createtaxdb's own README. Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.