findDiff: findDiff

In slzhao/MultiRankSeq: MultiRankSeq

Description

The function will perform DESeq, edgeR, and baySeq analysis.

Usage

findDiff(rawCounts, group = c(rep(0, ncol(rawCounts)/2), rep(1,
  ncol(rawCounts)/2)), seed = NULL, paired = NULL)

Arguments

rawCounts	the count data used in analysis, which should be a matrix or a data.frame.
group	the group of count data, which should be 0 or 1. The samples labeled as 1 will be compared with samples labeled as 0.
seed	the random seed for bayseq. The user can set it so that the bayseq result can be reproduced.
paired	the pairs of count data. Will be NULL if no pair. For example, if sample 1 and 2 are one pair, sample 3 and 4 are another pair, the paired should be c(1,1,2,2).

Value

A matrix with log2 fold change, p value, adjusted p value, and rank of DESeq, edgeR, and baySeq analysis.

Examples

set.seed(123456789) #generate data
ngenes <- 500 #1000
q0 <- rexp(ngenes, rate = 1/250)
is_DE <- runif(ngenes) < 0.3
lfc <- rnorm(ngenes, sd = 2)
q0A <- ifelse(is_DE, q0 * 2^(lfc/2), q0)
q0B <- ifelse(is_DE, q0 * 2^(-lfc/2), q0)
true_sf <- c(1, 1.3, 0.7, 0.9, 1.6,1.5)
conds <- c("A", "A", "A","B", "B", "B")
exampleCounts <- t(sapply(seq_len(ngenes), function(i) sapply(1:6, function(j) rnbinom(1,mu = true_sf[j] * ifelse(conds[j] == "A", q0A[i], q0B[i]),	size = 1/0.2))))
colnames(exampleCounts)<-c(paste("Control",1:3,sep=""),paste("Disease",1:3,sep=""))
row.names(exampleCounts)<-paste("Gene",1:ngenes,sep="")
#this code may take two minutes, because performing baySeq, DESeq2, and edgeR may be slow. 
exampleCountsDiff<-findDiff(exampleCounts) #find diff result

slzhao/MultiRankSeq documentation built on April 19, 2020, 6:14 p.m.