Or is it possible to construct a diablo model based on co-expression/co-variance networks that are from different types of data?
Thank you very much!
Hi @martina,
Your co-expression matrix would be of size p x q, where p, q are the variables from your two data sets X (n x p) and Y (n x q). Somehow that defeats the purpose given that DIABLO is trying to extract a similarity matrix between X and Y form the start!
From a theoretical perspective: we do need the input data sets to be matching on the same number of samples and your co-expression matrices wont allow that.
I understand from your question that you may not have the same matching n, and so the straight answer is: no 
Kim-Anh