pyfraglib 0.6.2

Software suite to calculate fragmentomics features from cfDNA and perform downstream analyses.

pyfraglib

Overview

pyfraglib is a Python library to analyze high-throughput sequencing data of cell-free DNA (cfDNA). More specifically, it facilitates the investigation of fragmentomics features of which a list can be found below.

Because fragmentation of cfDNA is non-random, tissue- and disease-specific patterns emerge in e.g. fragment length or end motif distributions. To date, no comprehensive tool exists to implement a "fragmentomics workflow". pyfraglib aims at being such at tool.

Installation

It is recommended that pyfraglib is installed into a dedicated conda environment using pip, the Python package manager:

git clone [e-mail]:schwarzlab/pyfraglib.git
cd pyfraglib
conda env create -f pyfraglib.yml
conda activate pyfraglib
python3 -m pip install .

During development, all code in pyfraglib is thoroughly type-checked using mypy. After an initial installation as described above, do:

./tools/dev_install.sh # typing & linting errors are reported

Usage

pyfraglib comes with a command line utility. After successful installation, it can be used as follows:

pyfrag.py version # show currently installed version
pyfrag.py --help # show available subcommands and flags

tools/workflow.sh gives an example of a commonly used sequence of commands.

Pipeline

pyfraglib offers a simple batch mode for most operations, i.e. it can take a directory of BAM or FRAG files and perform analyses on them. Some operations (i.e. the extract subcommand) are even parallelized. We explicitly do not recommend using this functionality for the analysis of large cohorts. Even 10 BAM files cannot be extracted at once without exceeding most workstation's memory. Thus, we include a convenient Nextflow pipeline in tools/pipeline.nf. It reads the input data from a TSV file and Nextflow parallelizes pyfraglibs operations as much as possible. See tools/nextflow.config for all paths and variables that must be set by the user. Currently, only slurm is supported as a scheduler (via the ramses profile).

Available fragmentomics features

Please refer to the project's documentation for a list of currently supported features (see below for instructions on how to build documentation using Sphinx).

Documentation

Sphinx is used to create documentation. The easiest way to create repository-wide documentation in HTML or PDF format, run:

./tools/build_docs.sh pdf # or:
./tools/build_docs.sh html

Citation

As long as pyfraglib is not published, please cite this repository if you use it in your scientific work.

License

pyfraglib is licensed under the GPL-v3 as indicated in the source files. Please direct requests to [e-mail].