CTGSA: CTGSA

Description Usage Arguments Details Value Examples

View source: R/CTGSA.R

Description

CTGSA

CTGSA : Covariance Thresholding for Gene Set Analysis of Genetic Pathway Data

Usage

1
2
3
CTGSA(x, y, out.type = c("stat", "pvalue"),
  percentile.type = c("case-control", "all", "user-defined"),
  percentile.user = NULL, nperm = NULL)

Arguments

x

an input matrix of dimension nobs x nvars.

y

a response variable where y=1 is the case and y=0 is control.

out.type

character of output type of statistic('stat') and p-value('pvalue').

percentile.type

three threshold calculating types are supported as 'case-control', 'all', and 'user-defined'. To choose the threshold, 'case-control' computes the percentile with the first singular value for case and control. Percentile.type 'all' calculates the percentile with whole samples. The percentile is given by user in 'user-defined', and the threshold is then calculated for case and control. The default is case-control.

percentile.user

a numeric value of percentile in [0, 1]. To specify the percentile, give the percentile.type 'user-defined', and the integer value taken here will be used.

nperm

number of permuation - default is 1000.

Details

See the reference below for more information.

Value

A list of a statistic, an optional pvalue and permuted statistics, and arguments used.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
x <- replicate(5, rnorm(100, 0, 1))
y <- rep(0:1, each=50)
CTGSA( x = x, y = y, out.type="stat", percentile.type = "all")


require(mnormt)
x <- rbind(rmnorm(50, rep(0, 5), diag(5) ),
           rmnorm(50, rep(0, 5), outer(1:5, 1:5, function(i, j) 0.6^abs(i-j) ) ))
y <- rep(0:1, each=50)
CTGSA(x=x, y=y, out.type="pvalue", percentile.type = "case-control", nperm=1000)

statpng/CTGSA documentation built on Jan. 17, 2020, 2:03 p.m.