freq.p.test: Frequency hot spot detection Obtains significance cutoff for...

In ccbiolab/svpluscnv: svpluscnv: analysis and visualization of complex structural variation data

Description

Frequency hot spot detection

Obtains significance cutoff for the frequency of binary events encoded in a matrix such as that generated by shattered.regions and shattered.regions.cnv algorithms

Usage

freq.p.test(
  mat,
  method = "fdr",
  p.cut = 0.05,
  iter = 100,
  zerofreq = TRUE,
  plot = TRUE,
  verbose = FALSE
)

Arguments

mat	(numeric matrix) a binary matrix where columns will be tested for their sum value compared to a permutated matrix
method	(character) the method to pass to p.adjust function
p.cut	(numeric) the cutoff for multiple hypothesis corrected p.value
iter	(numeric) Number of iterations to produce null distribution (note that null size will be iter*ncol(mat))
zerofreq	(logical) whether to remove bins with observed frequency = 0; It is recommended to set to TRUE when the bins span genomic regions of low coverage
plot	(logical) whether to generate a histogram comparing observed and null frequency distributions
verbose	(logical) whether to return messages

Value

an instance of the class 'freq.cut'

Examples

## validate input data.frames
cnv <- validate.cnv(segdat_lung_ccle)

## obtain a matrix of genomic bins vs samples indicating high density of breaks
shatt.regions <- shattered.regions.cnv(cnv)
mat <- [e-mail]

freq.p.test(mat)

ccbiolab/svpluscnv documentation built on Sept. 9, 2020, 4:52 a.m.