Can anyone please tell me whether I should go for multivariable analysis with my data in this condition?

Question

[deep771992chanda]

Hi, community- Firstly, sorry, my question may be silly to you. Still, I am asking just to be certain if my thinking is correct or not. I am doing metagenomic data analysis and trying to find out significantly associated bacteria with control and disease A groups. Now, this is an established fact that biologically disease A leads to another comorbid phenotype called disease B. Thus, many (but not all) disease A individuals show the disease B. Control samples never show disease B.

In this condition, my understanding is there is no need to adjust the association between disease A and control, considering disease B as a covariate.

Please suggest to me whether my perception is correct or not.

Thanks in advance

Answer 1

[Arsenal]

I think you should indeed consider diseaseB, and verify taxa associated in control, diseaseA_only, and disease_A+B.