I want to conduct a differential expression analysis with MAST between two groups. In the analysis, I want to correct for a continuous confounder in a manner similar to this article: https://science.sciencemag.org/content/sci/suppl/2019/05/15/364.6441.685.DC1/aav8130VelmeshevSM.pdf . They say that "To identify genes differentially expressed due to the disease effect, likelihood ratio test (LRT) was performed by comparing the model with and without the diagnosis factor". Analogously, I would constract the following two models:
zlmsignal <- zlm(~ iscancer + biasscore, sca) zlmbackground <- zlm(~ bias_score, sca)
Thus, comparing both models I could avoid finding genes that vary due to "bias_score", and only detect those that are important in cancer. Could I use a likelihood test for that? If so, how?
Thanks a lot in advance.