Consensus Clustering (kmeans) of LUSC RNAseq k.eq.4


Created By Peter Waltman peter.waltman

Disease: LUSC
species: Human
platform: RNAseq
Description: PanCancer clusters generated via Consensus Clustering, using k-means as underlying clustering method. The choice of k used is based on survival analysis (p-values from log rank tests). Data was 1) median normalized prior to clustering, by normals (median of the normals for a given tissue, subtracted from the tumor samples); 2) Z-transformed (mean==0, std==1). Features (genes) fed into consensus clustering were selected by selecting the 1500 with the greatest Median Absoluted Deviation (aka MADs). For better coverage, we took the union of the union of the 1500 features with the greatest MADs, centered on their median, as well as those centered on the mean, yeilding 2152 total features.
Number_of_Samples: 221

Consensus Clustering (kmeans) of LUSC RNAseq k.eq.4 page is loading…