Skip to main content

Table 2 Best combination of methods for each step: Summary of the best combination of algorithms for each view used to obtain the best grouping of patients that identifies significant sub-classes

From: MVDA: a multi-view genomic data integration methodology

   (a) (b) (c) (d)
Dataset Views Feature Feature Patients Late
   clustering selection clustering integration
TCGA.BRCA RNASeq Pam CAT-score Kmeans MF
  miRNASeq Pam CAT-score Pam  
TCGA.OV Gene Expression Pam Random Forest DM MF
  Protein Expression Pam - DM  
  miRNA Expression Pam - DM  
TCGA.GBM Gene Expressions Spectral CAT-score Kmeans MF
  miRNA Expression Ward - Kmeans  
OXF.BRCA.1 Gene Expressions Pam Random Forest Ward GLI
  miRNA Expression Pam Random Forest Kmeans  
OXF.BRCA.2 Gene Expressions Pvcluster CAT-score Kmeans MF
  miRNA Expressions Pam Random Forest Kmeans  
MSKCC Gene Expressions Pam CAT-score Kmeans MF
  miRNA Expressions Pam - Pam  
  CNV Spectral CAT-score Kmeans  
  Clinical - - Pam  
  1. In the feature selection column the symbol (-) means that feature selection was not executed because the number of features was small. Symbol (DM) in Patient clustering column means that same classification error was obtained with all the algorithms used