In BRCA, clustering on all variables yields no significant separation ( P = 0.28), while SPLS and the CCA‐based methods achieve modest significance only when using all 5 components (SPLS: P = ; PMDCCA: P = ; RGCCA: P = ) but fail when restricted to their selected features.
← all excerpts
An integrative multiomics random forest framework for robust biomarker discovery.
1
—
—