bestAICsearch: Searches for clustering with the best AIC

In junseonghwan/ScRNACloneEvaluation: ScRNA clone evaluation.

Description

Determines the AIC for different numbers of clusters.

Usage

bestAICsearch(dataVec, minK = 2, maxK, coverage, startseed = 100, nIterations = 40, breakOnIncrease=FALSE, verbose=FALSE)

Arguments

dataVec	A vector of mutation frequencies, or a matrix or dataframe with each row corresponding to a bulk sample.
minK	Minimal number of clusters
maxK	Maximal number of clusters
coverage	Average coverage of the sequencing data
startseed	Offset for the seed for the random number generator
nIterations	Number of different iterations for each parameter pair
breakOnIncrease	Whether to exit the search over the number of clusters whenever the AIC increases
verbose	Whether to report on results for each number of clusters

Value

List of different clustering results, with AIC values, for minK:maxK.

Author(s)

Jack Kuipers

See Also

VAFclusterEM

Examples

coverage <- 1e4
kclust <- 3
freqs<-c(0.25,0.32,0.36)
nos <- c(5,3,1)
datas <- c()
for(ii in 1:length(nos)){
  datas<-c(datas,rnorm(nos[ii],mean=freqs[ii],sd=VAFclusterEM:::sdfromcoverage(freqs[ii],coverage)))
}
maxK <- 5
aicMatrix <- bestAICsearch(datas, minK = 2, maxK = maxK, coverage = coverage)

junseonghwan/ScRNACloneEvaluation documentation built on Aug. 18, 2020, 8:53 p.m.