eventTimingOverList: eventTiming for multiple samples and regions
In epurdom/cancerTiming: Estimation of Temporal Ordering of Cancer Abnormalities

Description Usage Arguments Details Value Author(s) See Also Examples

eventTimingOverList is a wrapper to eventTiming that allows for timing of common events over several regions of a sample and/or multiple samples. getPi0Summary gets summary information about pi[0] (the first event) from the output of that function and returns a simple dataframe of the estimate of pi[0] for every region and sample combination.

1 2	eventTimingOverList(dfList, normCont, eventArgs) getPi0Summary(eventList, CI = TRUE)

`dfList`	a list of mutation data per sample. Each element of the list corresponds to a sample, and consists of a data frame of mutation information for all of the regions that are to be timed. The data frame must have a certain format, see description below.
`normCont`	a vector of equal length as dfList giving the normal contamination of each sample
`eventArgs`	list of arguments passed to eventTiming via ‘do.call’. Should NOT contain the arguments ‘x’,‘m’,‘history’,‘totalCopy’,‘type’,‘mutationId’ or ‘normCont’
`eventList`	Output of eventTimingOverList (see below)
`CI`	logical, whether to grab CI from the output.

The data frame of mutation data for each sample must have the following columns: ‘segId’, ‘type’, ‘nMutAllele’, ‘nReads’, ‘mutationId’. The rows of the data frame correspond to mutated locations within the sample. ‘segId’ is an identifier of the segmented region that the mutation is in; eventTiming will be run using the mutation data for each ‘segId’ in the sample. ‘type’ identifies the type of segment, one of c("Other","CNLOH","SingleGain","Diploid","DoubleGain"). ‘nMutAllele’ gives for every row (mutated location) the number of reads with the mutation and corresponds to ‘x’ imputed into eventTiming. ‘nReads’ gives the total number of reads covering the location and corresponds to ‘m’ imputed into eventTiming. ‘mutationId’ is a unique identifier for the mutation.

For arguments passed to eventTiming, it is recommended to pass verbose=FALSE if you want to avoid many warnings about mutations that do not meet the necessary criteria of coverage, etc.

eventTimingOverList returns a list equal to the number of samples. Each element of that list (i.e. per sample) is itself a list of length three corresponding to the three types of regions that can be timed, “CNLOH",“SingleGain", and “DoubleGain". Each of these gives the output of eventTiming per region of that type.

Elizabeth Purdom

eventTiming

if(require(GenomicRanges)){
	#fix up mutation data to right format
	data(mutData)
	colnames(mutData)[1]<-c("chr")
	colnames(mutData)[grep("t_alt_count",colnames(mutData))]<-"nMutAllele"
	colnames(mutData)[grep("t_ref_count",colnames(mutData))]<-"nRefAllele"
	mutData$nReads<-mutData$nMutAllele+mutData$nRefAllele
	mutData$mutationId<-1:nrow(mutData)
	
	#add segmentation annotation -- second region is bogus, only for illustration
	segs<-data.frame(chr=c(17,17),start=c(1,1.8e7+100),end=c(1.8e7,81195210),
		normCont=0.22,segId=c("Seg1","Seg2"),type=c("CNLOH","SingleGain"))
	##Create Trivial segmentation annotation for example
	mutId<-mut2Seg(mutData,segs)
	eventOut<-eventTimingOverList(dfList=list(Sample1=mutId),normCont=0.22)
	getPi0Summary(eventOut)
}