autodatabase

Automated build of a Kraken2 database using Nextflow DSL2. Requires Nextflow version>= 20.01.0 and either Docker or Singularity.

Quick Start

There are seven stages to the workflow:

1. Downloading the NCBI taxonomy (autoDatabase_getTaxonomy)
2. Adding the taxonomic ID to the sequence IDs and the filenames (autoDatabase_addTaxon)
3. Creating a mash matrix for each taxon (autoDatabase_mash)
4. Using the mash matrix to select high quality assemblies (autoDatabase_qc)
5. Creating a channel containing the high quality assemblies (autoDatabase_cleanFasta)
6. Building the Kraken2 database (autoDatabase_kraken2Build)
7. Creating a Krona chart showing the composition of the Kraken2 database (autoDatabase_krona)

The expected input for autodatabase are fasta files. They should sorted into directories for each taxon where the directory name is the taxon name with spaces replaced with underscores. E.g.

Mycobacterium_avium    Mycobacterium_bovis    Mycobacterium_tuberculosis_complex

The workflow will scrape the name of the taxon from the directory name and use it look up the taxonomic ID.

There are five global parameters needed to run autodatabase which can be set in nextflow.config:

• params.addFasta The directory containing the assemblies to be added to the database. Should consist of sub-directories where the fastas are sorted into named directories for each taxon (see above)
• params.newDatabase The directory where the results of the workflow will be saved. The default is ${baseDir}/results. On completion of the pipeline, the output kraken database .k2d files can be found in ${params.newDatabase}/krakenBuild_autoDatabase_kraken2Build
• params.previousDatabase Autodatabase can build on top of the fasta files from a previous database build. The fasta files from a completed build are found in ${params.newDatabase}/selectFasta_autoDatabase_cleanFasta/assemblies If there is no previous database build then set to null
• params.ncbiDate Date stamp of the NCBI taxonomy you wish to use to build the database. Takes the form YYYY-MM-DD
• params.modeRange For each taxon, autodatabase builds a mash distance matrix, finds the average mash distance for each assembly, and then finds the mode of the average mash distances to 2 s.f. The assemblies that have an average distance that is within params.modeRange of the mode will be used to build the database. Default value is 0.1, i.e. accepts mash distances within the range: mode±10%

The pipeline requires either Docker or Singularity to run. Scripts for building the containers needed to run the pipeline can be found in docker and singularity.

To run the pipeline:

nextflow run main.nf -profile [docker, singularity]

The workflow for the pipeline can be found below, for more information consult the wiki

Workflow