dib-lab / sourmash-oddify

Tools and workflows for evaluating bacterial and archaeal genome contamination with sourmash and GTDB.

sourmash-oddify

sourmash-oddify contains tools and workflows for examining bacterial and archaeal genome contamination with k-mers using sourmash.

Briefly, sourmash-oddify will:

• classify genomes with GTDB-Tk
• build sourmash LCA databases based on the resulting taxonomy
• identify, align, report, and ~remove taxonomically discordant sub-sequences

The primary value-add of sourmash-oddify is that it uses sourmash to quickly identify genomes that share many taxonomically discordant k-mers. These genomes can then be further investigated. (See the Kraken section in this blog post for more info on how we identify taxonomically discordant k-mers.)

sourmash-oddify is in the early stages of development but we are happy to support your using it! Please post questions and problems to our issue tracker.

Quickstart

You can install all of the software dependencies with bioconda by running:

conda env create -f environment.yml

The primary way to use sourmash-oddify is to run the snakemake workflow in the Snakefile. Do so by:

1. copying conf/default.yml to a new file and editing it
2. running snakemake --use-conda --configfile=newconf.yml

The output will be placed in the outputs_dir specified in your config file. More on what this output is ...later. :)

Configuring prefixes

For now, your genome identifiers must have a common prefix of at least one character. Put this prefix in the filter_prefixes in your config file.

GTDB-Tk reference data

Since sourmash-oddify runs gtdbtk, you'll need the GTDB-Tk reference data. Download it, unpack it somewhere, and then set the value for gtdbtk_data in your config file to the unpacked directory path. Note, this should be set to the top level directory containing the subdirectories fastani/ and taxonomy/, e.g. release89/.

Acknowledgements

sourmash-oddify relies on the GTDB taxonomy and the GTDB-Tk toolkit. Thanks, nice people!