Please see the mouthwash
function in the R package vicar.
Let
Y = XB + ZA + E,
for
Not accounting for the hidden covariates, Z, can reduce power and result in poor control of false discovery rate.
succotashr
fits this model under a two-step empirical Bayesian approach. It places a non-parametric prior on B and jointly estimates B and ZA. The main function is succotash
.
To install, run the following code in R
:
install.packages(c("devtools", "SQUAREM"))
devtools::install_github("stephens999/ashr")
devtools::install_github("dcgerard/vicar")
devtools::install_github("dcgerard/succotashr")
The following packages are suggested:
install.packages("cate")
source("https://bioconductor.org/biocLite.R")
biocLite(c("limma", "sva"))