Identifying and Assessing Interesting Subgroups in a Heterogeneous Population

Figure 1

(a) Step I is clustering of genes. Genes (A) (shaded region) are grouped and will be used as a subtype identifier in the downstream analysis. (b) Step II is clustering of patients using Genes (A) (i.e., a gene-set obtained from Step I). Here, MS1 (a set of patients, shaded regions) is obtained as a subtype. (c) Step III shows how is obtained from the FDR curve. We count the number of genes having FDR < 0.1. We repeat implementing (a), (b), and (c) across different clustering results extensively. Thus, no shaded columns in (b) will be covered subsequently.

(a) Step I

(b) Step II

(c) Step III