Fig. 2

Overview of integrative clustering in integrative phenotyping framework (iPF): (a) Vertically combined multiple omics data sets (b) A distance matrix between any two features within and across omic data sets (c) Multidimensional scaling (MDS) mapping to a two-dimensional Euclidean space (d) Smoothed feature intensities in the reduced 2D space for each patient (e) Unsupervised clustering to identify potential disease subtypes and averaged feature intensities for representative plots of each cluster