sample.dep.2: Simulate a sample of groups of dependent normal observations

In warnes/exp.ssize: Estimate Effectivenes of Microarry Sample Size Estimation via Simulation

Description

Given a list of covariance matrices, this function simulate normal observation based on the covariance matrix from first until the last one in the list of cov.matrix

Usage

sample.dep.2(ngenes.dep, n, var.ratio, cov.matrix, delta, distrn = "normal")

Arguments

ngenes.dep	integer number of genes to sample the dependent expression
n	positive integer number of replicates for each gene each group (control/treatment)
var.ratio	a positive number for the ratio of variance between treatment and control
cov.matrix	a list of covariance matrices based on which normal samples are generated
delta	effect size for sample size calculation the absolute distance from null value for expression values of genes coming from the true alternative hypothesis group
distrn	distribution of expression data, set to be normal for here

Value

returns a matrix with ngenes.dep rows and 2*n columns with the first n columns being samples from control group and last n columns being from treatment group for each gene (row). each group of genes consists of subgroups of genes that are dependent

Author(s)

Peng Liu [e-mail]

Examples

library(MASS)

n.dep <- 9; n <- 100; delta <- 5; var.ratio <- 1;
n.matrix <- 3; cov.matrix <- list(n.matrix);

for ( i in c(1:n.matrix)) {
    cov.matrix[[i]] <- diag(rep(i*0.1, 3))
}

cov.matrix[[3]][1,1] <- cov.matrix[[2]][1,1] <- 1
cov.matrix[[3]][2,1] <- cov.matrix[[3]][1,2] <- 0.15 

test.sample.dep <- sample.dep.2( n.dep , n, var.ratio, cov.matrix, delta, distrn = "normal")

warnes/exp.ssize documentation built on May 4, 2019, 12:59 a.m.