power.de: Power calculation for a DE gene

In heiniglab/scPower: Experimental design framework for scRNAseq population studies (eQTL and DE)

Power calculation for a DE gene

Description

This function calculates the power to detect an DE gene (comparsion of two groups 0 and 1) by using the function power.nb.test of the package MKmisc. The power for a gene with mean of 0 is defined as 0 (the gene will have an expression probability of 0, so the overall detection power is also 0).

Usage

power.de(
  nSamples,
  mu.group0,
  RR,
  theta,
  sig.level,
  approach = 3,
  ssize.ratio = 1
)

Arguments

nSamples	Total samples size (n0 + n1), group balancing defined by size.ratio
RR	effect size of group 1 vs group 0 (fold change mu1/mu0)
theta	1/dispersion parameter of the negative binomial fit
sig.level	Significance threshold
approach	Choose between three different methods implemented in the package for the power calculation (1,2,3)
ssize.ratio	Sample size ratio between group 1 and 0 (n1/n0) (Remark: there is a mistake in the package documentation that the ratio is n0/n1, but I checked the code and the associated package)
mu.grou0	Mean value of group 0

Value

Power to detect the DE gene

heiniglab/scPower documentation built on Jan. 9, 2025, 12:13 p.m.