Hello, I am currently using the mixOmics package to run DIABLO analysis. My dataset contains both male and female individuals which is known to impact the metabolites I am including in the analysis. So I wanted to ask if there was a way to control for sex as a confounding variable in my analysis? I thought of adding an extra X dataset with the sex information encoded using binary code but I am not sure if that would cause problems when building the model?
Is there any solutions apart from building two separate models for each sex?
Thank you very much in advance for your input.
hi @kabaakil,
Unfortunately at this stage we are not able to include a covariate in the model to account for it - although we are thinking about developing a method towards this (it will take year).
If you add Sex as an extra dataset, DIABLO will try to look for variables that explain Sex, this is not what you want. If Sex effect is strong, you could consider correcting for Sex effect first (if the sex effect is not strong, or not totally confounded* then it might be fine). We have a list of methods you could try (regardless of whether it is microbiome data or not:
*if the treatment is totally confounded with Sex then there is not much you can do.
Kim-Anh
Thank you very much Kim-Anh for your detailed answer. I will have a go at the methods you listed.