This repository contains all the scripts used to provide the results for the CloneSig article.
Abécassis, Judith, Fabien Reyal, and Jean-Philippe Vert. "CloneSig: Joint inference of intra-tumor heterogeneity and signature deconvolution in tumor bulk sequencing data." BioRxiv (2019): 825778.
The folder signature_code contains most of the code, with the master script in the bash script run_all.sh. It is not advised to execute it all at once, considering the run time, and the paths to change (python virtual environments, paths to executables etc).
The folder notebooks contains notebooks and scripts in which the final analyses are generated. The tables used to generate those results are in the folder result_tables.
The folder external_data contains data from external sources that were useful to carry out the project. A specific readme is included to provide the origin of each file.
It should be noted that more complete and up-to-date instructions of the CloneSig python package can be found here. Feel free to get in touch or to open an issue if you have any question or suggestion.
this project uses many packages. Here is a list of the most important ones. Requirement (python27_requirements.txt and python36_requirements.txt) files are provided to use with pip and setup two virtual environnments for Python 2.7 and 3.6 (or conda for some specific packages, in particular mkl and PyClone). Here is a list of the main ones, including some that can be installed with pip (specified). Code was run with R version 3.3.2, and Python 3.6.8 (Anaconda installation), except to run PyClone (Python 2.7.9). All computation was performed under a Centos distribution, with torque as scheduler.
- segment_liftover 0.951 (python >= 3.6 - installed by pip)
- PyClone 0.13.0 (Python 2.7)
- Sciclone 1.1
- Tracksig
- Ccube 205f5e7b89
- palimpsest
- CloneSig
- deconstructSigs 9bbaf15387
- TrackSigFreq 23b2f3f
- DPClust 75f5d7e
- PhylogicNDT c229cec
All scripts should be run from the root of the folder. Total run time is several weeks on a cluster with 60 CPUs.