BMfinder: Discover bimodal distrubition features

In yuhuihui2011/vasp_test: Quantification and Visulization of Variations of Splicing in Population

Description

Find bimodal distrubition features and divide the samples into 2 groups by k-means clustering.

Usage

BMfinder(
  x,
  p.value = 0.01,
  maf = 0.05,
  miss = 0.05,
  fold = 2,
  log = FALSE,
  cores = detectCores() - 1
)

Arguments

x	a numeric matrix with feature rows and sample columns, e.g., splicing score matrix from spliceGenome or spliceGene function.
p.value	p.value threshold for bimodal distrubition test
maf	minor allele frequency threshold in k-means clustering
miss	missing grouping rate threshold in k-means clustering
fold	fold change threshold between the two groups
log	whether the scores are to be logarithmic. If TRUE, all the scores are log2 tranformed before k-means clustering: x = log2(x+1).
cores	threads to be used. This value is passed to mclapply in parallel package

Details

The matrix contains 1, 2 and NA, and values of 'x' in group 2 are larger than group 1.

Value

a matrix with feature rows and sample columns.

Examples

data(rice.bg)
score <- spliceGene(rice.bg, "MSTRG.183", junc.type = "score")
score <- round(score, 2)
as <- BMfinder(score, cores = 1)  # 4 bimodal distrubition features found

## compare
as
score[rownames(score) %in% rownames(as), ]

yuhuihui2011/vasp_test documentation built on March 5, 2020, 12:50 a.m.