Entering edit mode
11 weeks ago
cmorris2945 • 0
How do I use TCGA website or database to perform Non-negative matrix factorization? Do I use NMF on the TCGA website data or use NMF on the data from the cancer patients to see which genes and cells are contributing mostly to the cancer?
Why do you want to perform NMF? What exactly is the matrix you want to factorize?
As a side note, NMF (nor any method) will tell you which genes contribute most to cancer. It can separate groups and potentially identify subgroups but it neither measures interestingness nor pathophysiology and pathorelevance.
I would download the data via the likes of TCGABiolinks and run NMF myself, for example via RcppML or other implementations.