This is an extension to the Sunbeam pipeline to run the kraken2 read classifier. It works on qc'd reads, and requires that you have a kraken2 database available. Check out the instructions here for help with making a kraken2 database.
Clone the repo into your sunbeam extensions/ folder and add the new options to your existing configuration file. Make sure you've installed Sunbeam first!
source activate sunbeam
cd $SUNBEAM_DIR
git clone https://github.com/louiejtaylor/sbx_kraken2/ extensions/sbx_kraken2
Add the options to your config file (replace "sunbeam_config.yml" with the name of your config file).
cat extensions/sbx_kraken2/config.yml >> sunbeam_config.yml
In your configuration file, make sure you choose how many threads you'd like kraken2 to use, as well as providing the path to your database:
sbx_kraken2:
threads: 4
group_file: '/path/to/your/kraken2/database_dir'
Finally, run Sunbeam as usual (with the --use-conda flag) with your extension's target rule specified:
sunbeam run --use-conda --configfile=sunbeam_config.yml all_kraken2
The --use-conda flag is crucial as it enables running kraken2 in an isolated environment.
This rule generates output in the following location: sunbeam_output/classify/kraken2/
sbx_kraken2_env.ymlspecifies the extension's dependencies and provides the environment that Sunbeam usesconfig.ymlcontains configuration options that can be specified by the user when running an extensionsbx_kraken2.rulescontains the rules (logic/commands run) of the extension