This repository contains the code for running all the experiments on TCGA data. In particular, for each multi-omics cohort:

• we reduce individual views by using a tailored pipeline guided by block-analysis.
• all the reduced views are integrated via SNF, MOFA+, or uMKL
• the integrated data is input to RF classifiers for survival prediction

The crucial code is the one computing the blocking analysis. It is available in file 'blocking_ID.R' (block-ID analysis) and 'estimate_ID_twoNN.R' (unbiased twonn ID estimate via undersampling)

For any help and questions please contact elena.casiraghi@unimi.it