Adjust for cell composition in 450K data.
Much of the adjust.beta code was originally written by Andres Houseman. I have only put it into a simple to use function and packaged it.
> library(devtools)
> install_github("brentp/celltypes450")
See this thread for details.
Compared to minfi, which normalizes the sorted data with the new data:
To get your beta (0, 1) values adjusted for cell-type composition, use:
library(devtools)
install_github("brentp/celltypes450")
library(celltypes450)
adjusted = adjust.beta(beta)To get the cell-composition estimates such as those from the image above:
cell.types = adjust.beta(beta, est.only=TRUE)