Adapting the Signature Estimation R package for Python. Currently, this package only implements the quadratic programming (QP) algorithm. See [1] for details.

SignatureEstimationPy requires Python 3. We recommend using Anaconda to install dependencies.

Once you have Anaconda installed, you can create an environment with all the dependencies installed, and activate it, with the following commands:

conda env create -f environment.yml
source activate signature-estimation-py-env

The input to the signature_estimation.py script is a mutation count table and a mutation signatures table, and an (optional) list of signatures. To get the full list of parameters, run:

python signature_estimation.py -h

We provide scripts in example/ to run SignatureEstimation on the COSMIC mutation signatures and mutation data from the Alexandrov, et al. (Nature 2013) paper.

Within the example/ directory, you can run a full example on the 397 melanoma samples from the Alexandrov, et al. data using the provided Snakefile with the command:

snakemake all

This will download and process the COSMIC signatures and melanoma mutation data, and compute their exposures using the SignatureEstimation quadratic program. The pipeline will produce a plot of exposures per signature, as automatically produced via continuous-integration below.

We gratefully acknowledge Xiaoqing Huang for her help developing and testing this code.

Xiaoqing Huang, Damian Wojttowicz, and Teresa M Przytycka. Detecting presence of mutational signatures in cancer with confidence. Bioinformatics(Oxford, England), 2017. doi: 10.1093/bioinformatics/btx604.