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 [email protected]
