snakemake_workflows

Useful Snakemake workflows related to bioinformatics work in the Magwene lab.

Basic setup

• Create a directory for your analysis ~/my_analysis

• Create a subdirectory for config files: ~/my_analysis/config

  • Copy an appropriate config.yaml for the workflow you want to run to ~/my_analysis/config; edit any parameters as necessary

  • Note that for multi-tool workflows you simply add appropriate YAML blocks for each tool

  • Each workflow also typically takes a CSV formatted table named as toolname_table.csv (e.g. fastp_table.csv) that lists the samples analyzed and where the files used in the analysis can be found. The required format for these tables is specified in the workflow README documents (e.g. README_fastp.md). For small inputs these can be created by hand, but when dealing with hundreds of samples these are best created using bash (or other) scripts.

    • In multitool workflows, the output of one workflow may be the input into another workflow

• Copy one of the example "driver" snakefiles from snakesfiles/ to ~/my_analysis/Snakefile (the Snakefile found in each tool subdirectory is local to that tool and should not be modified)

• Run the snakefile (e.g. snakemake -c 24 --use-conda) for the top-level directory (~/my_analysis).

• Results will be written to ~/my_analysis/results, with an appropriate sub-directory per tool (e.g. ~/my_analysis/fastp) and each samples results typically written to a sample specific directory (e.g. ~/my_analysis/fastp/PMY1234/)

• When appropriate, a directory with soft-links to the key results files is also typically created for each tool (e.g. ~/my_analysis/fastp/softlinked). This facilitates access all the key result files from one directory, and is useful when the output of one tool is the input into the next tool