Hi there

I'm not sure how it's called what I'm trying to do, but I would like to extract (residuals of) normalized expression/abundance values that are "corrected/adjusted" for certain design factors.

For example, using the regular lm(), I can correct for spatial variation (characterized by a bunch of variables, everything on the right hand side of the tilde) like this:

resid(lm(myVar ~ R+H+I(RR)+I(HH)+I(R*H), data = forTest, na.action = na.exclude))

and then do further analyses or plotting on the residuals.

Is there a way to do something similar directly with DESeq2? I.e., not via using normalized and log2(x+1)-transformed in a separate lm().

Best regards,

Marc

There isn’t really a residual in a GLM that is similar to the residual in a linear model. I’d recommend working with the VST data if you want to assess the variation in the data before and after controlling for variables.