R/MSET.R

In SuperExactTest: Exact Test and Visualization of Multi-Set Intersections

#Statistical Test of Multi-Set Intersections
#Author: Minghui Wang, minghui.wang@mssm.edu
#Date: 20 July, 2014

#a wrapper to call cpsets
MSET <- function(x, n, lower.tail=TRUE, log.p=FALSE){
#Input:
#x           a list of sets.
#n           background size.
#lower.tail  logical; if TRUE (default), probability is P[overlap < m], otherwise, P[overlap >= m], where m is the number of elements shared by all sets.
#log.p       logical; if TRUE, probabilities p are given as log(p).
	L=sapply(x,length)
	if(n<1 | any(L>n) | any(L==0)) stop('Invalid input\n')
	intersects = intersect.list(x)
	Obs = length(intersects)
	Exp = n*do.call(prod, as.list(L/n))
	if(Obs == 0 ){
		p = ifelse(lower.tail, 0.0, 1.0)
	}else{
		p = cpsets(Obs-1, L, n, lower.tail, log.p)
	}
	list(intersects=intersects, FE=Obs/Exp, p.value=p)
}
#Sample usage:
##not run###
#sample data
#x=list(S1=c(letters[1:20]),S2=c(letters[10:26]),S3=c(sample(letters[1:26],10,FALSE)),S4=c(sample(letters[1:26],10,FALSE)))
#MSET(x,26,FALSE)