callVariants: callVariants using beta binomial distribution

In umiAnalyzer: Tools for Analyzing Sequencing Data with Unique Molecular Identifiers

Description

Calculate variant p-values using permutation-based testing. A prior is fitted to model the background error using maximum likelihood estimation of a beta distribution. The maximum likelihood estimate of the beta distribution is then used to define the shape of a beta-binomial distribution used to estimate variant P-Values. This can be interpreted as a probability for a variant to not have arisen by chance.

Usage

callVariants(object, minDepth = 3, minCoverage = 100, computePrior = FALSE)

Arguments

object	A UMIErrorCorrect object.
minDepth	Minimum consensus depth required default is 3
minCoverage	Minimum Coverage to use, default is 100 reads.
computePrior	Should a new distribution be derived from data? Default is FALSE.

Value

Object containing raw and FDR-adjusted P-Values

See Also

filterVariants on how to filter variants.

Examples

library(umiAnalyzer)
main <- system.file("extdata", package = "umiAnalyzer")

simsen <- createUmiExperiment(main)
simsen <- filterUmiObject(simsen)
simsen <- callVariants(simsen, computePrior = FALSE)

umiAnalyzer documentation built on Nov. 25, 2021, 9:07 a.m.