Entering edit mode
Could you explain in simple terms what is the purpose of DESeq2 unmix
function?
What would be a scenario where it would be useful?
What is meant by the pure
components?
Could you explain in simple terms what is the purpose of DESeq2 unmix
function?
What would be a scenario where it would be useful?
What is meant by the pure
components?
The description and examples in ?unmix
may help also get an idea for what it is doing.
Briefly, unmix
estimates contributions for input samples in terms of purified reference expression profiles (estimates, per sample, a vector that sums to 1) so it's similar to other reference-based deconvolution tools. It was developed around 2017 to help a collaborator, and is a basic NNLS-type approach but using the DESeq2 VST for the loss function.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
I understand the general scenario to be the following : I should have expression from a mixture of cell sub-types A,B,C. Also, I should have "pure" expression in all cell-types A,B,C separately (coming from another experiment). Then according to the pure expression from another experiment,
unmix
estimates the proportions of cells A,B,C in the mixture. Is that correct?Yes, that's right.