This repository consist of code relevant for additive variance analysis performed on PCAWG mutations.
Following dependencies are required to run this workflow.
FunSeq2 (http://funseq2.gersteinlab.org/)
Python
Matlab
GCTA (http://cnsgenomics.com/software/gcta/#Overview)
This workflow consist of two components.
-
pre-processing step In this step we generate summary file for each cancer cohort.
Two input files are needed for this step. a) PCAWG driver mutation list and b)FunSeq2 output file(in BED format)
Usage: generateSummaryInfo.py -d -I -O generateSummaryInfo.py (-h | --help)
-
post-processing steps
Pipeline is run by calling additive_variance_demo.m
Full pipeline requires the following inputs:
-
in bedFiles folder: cohortName.null.bed cohortName.obs.bed
-
in summaryFiles folder: cohortName.null.summary.txt cohortName.obs.summary.txt
and generates the output cohortName.txt in the results folder.
-
A gcta executable is required (Linux version is included, Windows and Mac versions are available from cnsgenomics.com/software/gcta), which should be placed in the gctaFiles folder.
Current pipeline is set up to call only final stage, which summarizes the results from precomputed intermediate outputs. Outputs are computed from the Breast-AdenoCa cohort with randomized samples used for both null and obs conditions. Results text file shows calculated additive variance for each funseq threshold, which is ~0 (1e-6), along with associated p-values (0.5 indicates that no significant genetic variance was found). Values of -1 in the results file for funseq thresholds 5 and 6 indicate that insufficient data was found at these thresholds.