Joint k-mer and variant phasing

This is a pipeline designed to phase variants and PacBio Hifi data using a combination of k-mers and variants.

However, by running it in different modes you can also use it to phase using only variants and not k-mers (hiphase), or to just generate variant calls without any phasing of reads or the vcf (DeepVariant).

Install

Please start by installing pixi which handles the environment of this Snakemake workflow. You also must have apptainer available in your path/system.

You can then install the pixi environment by cloning this repository and running:

pixi install

pixi then handles the execution of the Snakemake workflows.

We then recommend you run a partial test of your installation with:

pixi run test

Usage

First set up a configuration file. See config/config.yaml for a commented example. Then run snakemake with the following command pointing to your configuration file.

pixi run snakemake --profile profiles/slurm-executor --configfile config/your_config.yaml

Running without parental data

You can run the pipeline the same way, it will only use hiphase instead of a combination of hiphase and k-mers to phase. See config/no_parental.yaml for an example input file.

Running only DeepVariant

First set up a configuration file. See config/deepvariant.yaml for a commented example. Then run snakemake with the following command pointing to your configuration file.

pixi run snakemake deepvariant --profile profiles/slurm-executor --configfile config/your_config.yaml

Running only alignment (pbmm2)

First set up a configuration file. See config/align.yaml for a commented example. Then run snakemake with the following command pointing to your configuration file.

pixi run snakemake pbmm2 --profile profiles/slurm-executor --configfile config/your_config.yaml

Running from another directory

If you want to run this Snakemake from another directory you can do so with:

pixi run --manifest-path /path/to/snakemake/pixi.toml snakemake ...

where you update /path/to/snakemake/pixi.toml to the path of the pixi.toml you cloned.

Dependencies

Requirements for executing the pipeline are, apptainer/singulariy or docker and pixi.

If you wish to distribute jobs across a cluster you will need to install the appropriate snakemake executor plugin. For example, to use SLURM you can install the snakemake-executor-slurm plugin using pip:

pip install snakemake-executor-plugin-slurm

Additional dependencies are managed automatically by snakemake using mamba. For Stergachis lab members be sure to include the following in your .bashrc if you want to use a pre-computed conda env.

export SNAKEMAKE_CONDA_PREFIX=/mmfs1/gscratch/stergachislab/snakemake-conda-envs