/dpclust3p

Dirichlet Process Clustering Pre-Processing Pipeline

Primary LanguageRGNU Affero General Public License v3.0AGPL-3.0

DPClust pre-processing

This R package contains various functions to produce input data for DPClust using SNV variant calls and copy number data from Battenberg. Most importantly, it contains the runGetDirichletProcessInfo function that produces the input data for SNV based clustering.

Installation instructions

dpclust3p is an R package and can be installed with the commands right below. It also requires the alleleCounter tool to be in $PATH.

source("http://bioconductor.org/biocLite.R"); biocLite(c("optparse","VariantAnnotation","GenomicRanges","Rsamtools","ggplot2","IRanges","S4Vectors","reshape2"))'
devtools::install_github("Wedge-Oxford/dpclust3p")

Running pre-processing

The typical usage is to create the DPClust input data. See inst/example for a few example pipelines. A pipeline typically consists of three steps:

  • Transform loci from a VCF file into a loci file
  • Obtain allele counts for all mutations, either by invoking alleleCount or by dumping counts from the VCF file
  • Convert allele counts and copy number information into DPClust input

The R package contains many functions from which one can build their own pipeline

File Description
preprocessing.R Main preprocessing functions to create DPClust input, perform mutation phasing, filter by mutational signature
allelecount.R Functions to count alleles in a BAM file, or dump counts from a range of VCF formats
kataegis.R Functions to identify kataegis events (requires fastPCF.R)
copynumber.R Various functions related to copy number
qualitycontrol.R Create plots that can be used for QCing
interconvertMutationBurdens.R Basic functions for data transformations
util.R Various utility functions

Docker

This package has been Dockerised, build as follows:

docker build -t dpclust3p:1.0.8 .