# simu.sample: Simulate a sample with normal distribution In warnes/exp.ssize: Estimate Effectivenes of Microarry Sample Size Estimation via Simulation

## Description

Simulate samples from normal distributions with specified proportion of observations from true alternative hypothesis, dependent groups, variance ratio for treatment/control group

## Usage

 `1` ```simu.sample(ngenes, n, fractn.alt, fractn.dep, var.ratio, cov.matrix, ngenes.matrix, delta) ```

## Arguments

 `ngenes` number of genes to simulate `n` number of replicates for each group of control/treatment `fractn.alt` the fraction of genes coming from true alternative hypothesis `fractn.dep` the fraction of genes coming from dependent group `var.ratio` ratio of variance between treatment/control `cov.matrix` a list of covariance matrices to sample the covariance from `ngenes.matrix` the number of covariance matrices in the list of cov.matrix `delta` effect size for sample size calculation the absolute distance from zero for expression values of genes coming from the true alternative hypothesis group

## Value

returns a matrix with ngenes 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 and/or coming from true alternative hypothesis. the row order is: null.ind -> null.dep -> alt.ind -> alt.dep

## Author(s)

Peng Liu peng\[email protected]

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18``` ```library(MASS) cov.matrix <- list(7); for ( i in c(1:7)) { cov.matrix[[i]] <- diag(rep(i*0.1, i+1)) # cov.matrix[[i]] <- diag(rep(i*0.1,4)) } cov.matrix[[4]][1,1] <- cov.matrix[[5]][1,1] <- cov.matrix[[6]][1,1] <- 1 cov.matrix[[4]][2,1] <- cov.matrix[[4]][1,2] <- 0.15 # corr = 0.15 / (1*0.2) = 0.75 cov.matrix[[5]][2,1] <- cov.matrix[[5]][1,2] <- 0.5 # corr = 0.15 /(1*sqrt(0.5)) = 0.59 cov.matrix[[7]][1,1] <- 10 cov.matrix[[7]][2,1] <- cov.matrix[[7]][1,2] <- 0.6 # corr = 0.67 cov.matrix[[7]][3,1] <- cov.matrix[[7]][1,3] <- 0.5 cov.matrix[[7]][3,2] <- cov.matrix[[7]][2,3] <- 0.4 test.sample <- simu.sample(ngenes = 100, n = 6, fractn.alt = 0.8, fractn.dep = 0.2, var.ratio=1, cov.matrix, ngenes.matrix = 7, delta = 10) ```

warnes/exp.ssize documentation built on May 28, 2017, 12:58 a.m.