Molecule Level Statistics

Share:

Description

Takes a vector of statistics with each element corresponds to a treatment vs control comparison, and calculates a combined statistics accross multiple treatments.

Usage

1
geneStats(T, method="OSP")

Arguments

T

a vector of statistics (z-scores converted) with each element correspond to a treatment vs control comparison.

method

the p-value integration method for combining accross multiple treatments. Available methods are Stouffer, OSP, Fisher, and maxP. The default method is OSP.

Value

a p-value after integration across treatments.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
# load the example data
data(PM)

# convert statistics into z-scores
PM.zscores <- apply(PM, 2, function(x){qnorm(rank(x)/(nrow(PM)+1))})

# Rotate the matrix by contrast 1, -1, -1 (i.e. up-regulation, down-regulation, dow-regulation).
PM.rotated <- rotate3d(PM.zscores, contrast = c(1, -1, -1))

# combine rotated statistics across treatments
gene.pvalues <- apply(PM.rotated, 1, geneStats)