I’m new to MixOmics and I’m having some trouble understanding how I should integrate my biological replicates into my analysis. I have 16 strains and for each, 4 or 5 biological replicates. For each replicate I have RNA Seq and metabolite data (no missing data). My strains belong to two main categories. I want to compare these two broad categories and identify RNA and metabolites present/over-expressed in one of the two groups versus the other, but taking into account the replicates. I was thinking of using Diablo to do this. I consulted the forum and found these topics which seemed to me the most similar to my case: withinVariation() on part of dataset and DIABLO analysis with time points
My questions are the following:
Can I use withinVariation on each of the 16 strains and then concatenate the 16 submatrices with rbind?
Can I do that on both datasets (RNA and metabolites)?
Is the use of diablo appropriate with my data and with the withinVariation function?
Do I necessarily have to use a linear mixed model to account for variation within strains?
I’ve looked in the tutorials (multilevel and multilevel sPLS pages) and in the forum, as well as in the related publications, but I feel like each time it doesn’t quite fit (for example, repeated measurements could have been appropriate, but I don’t have the same strains in my two main categories), so I’d rather ask the question here. i’m sorry if I missed the explanation somewhere.
Thanks a lot