1

I'm interested in finding the interaction impact of 3 genes together on a quantitative phenotype with adjustment of 3 covariates . I want to study the interaction effect but not the main effect. I'm fitting this model:

model<-lm(quantitative trait~ covariate 1+ covariate 2+ covariate 3+gene1:gene2:gene3,data)

So my questions are: Is it correct to look at the interaction effect without looking at the main effect? Is the model correct? Your help is very appreciated.

Marwah Al-kaabi
  • 309
  • 1
    Please look at this page and see if you still think it’s reasonable to omit “main effects” or the two-way interactions. If so, please edit your question to clarify what you hope to gain by doing so. I suspect that the scientific question will require including all those lower-level terms. – EdM Aug 17 '22 at 02:58

0 Answers0