Average Power and True FDR Based on limma/voom RNAseq Analysis Pipeline

Share:

Description

For the limma/voom RNAseq analysis pipeline, when we control false discovery rate by using the Benjamini and Hochberg step-up procedure (1995) and/or Storey and Tibshirani's q-value procedure (Storey et al, 2004), check.power calculates average power and true FDR for given sample size, user-specified proportions of non-differentially expressed genes, number of iterations, FDR level to control, mean counts in control group, dispersion, and fold change.

Usage

1
2
check.power(nGenes = 10000, pi0 = 0.8, m, mu, disp, fc, up = 0.5,
  replace = TRUE, fdr = 0.05, sims = 100)

Arguments

nGenes

total number of genes, the default value is 10000.

pi0

proportion of non-differentially expressed genes, the default value is 0.8.

m

sample size per treatment group.

mu

a vector (or scalar) of mean counts in control group from which to simulate.

disp

a vector (or scalar) of dispersion parameter from which to simulate.

fc

a vector (or scalar, or a function that takes an integer n and generates a vector of length n) of fold change for differentially expressed (DE) genes.

up

proportion of up-regulated genes among all DE genes, the default value is 0.5.

replace

sample with or without replacement from given parameters. See Details for more information.

fdr

the false discovery rate to be controlled.

sims

number of simulations to run when computing power and FDR.

Value

pow_bh_ave

average power when controlling FDR by Benjamini and Hochberg (1995) method.

fdr_bh_ave

true false discovery rate when controlling FDR by Benjamini and Hochberg (1995) method.

pow_bh_ave

average power when controlling FDR by q-value procedure (Storey et al., 2004).

fdr_bh_ave

true false discovery rate when controlling FDR by q-value procedure (Storey et al., 2004).

Author(s)

Ran Bi biran@iastate.edu, Peng Liu pliu@iastate.edu

References

Benjamini, Y. and Hochberg, Y. (1995) Controlling the false discovery rate: a practical and powerful approach to multiple testing. J. R. Stat. Soc. B, 57, 289-300.

Storey, J. D., Taylor, J. E. and Siegmund, D. (2004) Strong control, conservative point estimation and simultaneous rates: a unified approach. J. R. Stat. Soc. B, 66, 187- 205.

Examples

1
2
3
4
5
6
7
8
library(limma)
library(qvalue)
m <- 14                      ## sample size per treatment group
mu <- 10                     ## mean read counts in control group
disp <- 0.1                  ## dispersion for all genes
fc <- 2                      ## 2-fold change for DE genes

check.power(m = m, mu = mu, disp = disp, fc = fc, sims = 2)