This repository houses scripts for classifying post-zygotic indel and SNV variants as either true or false positives. The classification is based on a random forest model trained using molecularly validated variants to distinguish real post-zygotic mutations from others.
The script takes as input a VCF file generated by
Mutect2,
and inputs should contain the same attributes in the INFO
and FORMAT
fields as Mutect2 output.
This directory
contains example VCF files and their corresponding index.
This script is available as a Python package, Docker container, and WDL workflow. You can run it locally, via command line, in Docker, or on the cloud with Cromwell or Terra. Please refer to this README for details.