Integration of SNVs, CNAs and SVs (binary, continuous and discrete data)


I would like to know if it is okay to implement DIABLO with the following matrices:

SNVs matrix with binary data: 0,1 depending on if a gene is mutated or not in each sample
CNVs matrix with continuous data: from 0 to 20, representing the copy number of each gene for each sample
SVs matrix with discrete data: from 0 to 50, representing the number of events of each type that each sample harbours (the columns would be: nb of translocations, nb of insertions, nb of inversions, nb of complex rearrangements). So this particular matrix does not harbour genes in its columns, but structural variant types.

Do you think this data can be integrated by using DIABLO?


I can’t say I’ve seen DIABLO used in this way, but my gut feeling is that it should be fine - given all your blocks are measured across the SAME SAMPLES (and these are in the same order). I’d also recommend exploring some normalisation procedures to apply to your datasets as the varying scales and types may cause some issues with balancing the influence of each block.

Please let me know how this goes as I’m curious

Thanks for your answer! Yes, the data are measured across the same samples. Do you know where can I find information regarding normalization methods? I am not very familiar with those, specially taking into account that I am using binomial, discrete and continuous data.

Thanks a lot!