Figure 3.

The integrating clustering and selection procedure. The left panel depicts a pseudo-code description of the algorithmic procedure. The right panel illustrates an example of the process. At iteration 1, the algorithm selects a 'clusterable' subset of genes including (x, z, t, v) that results in two clusters (x, z, v) and (t). The remained AM consists of corresponding rows and columns of genes (y, u, w) from the original AM. At iteration 2, the procedure selects (y, u) and the remained AM now contains only one gene (w); at that point, the process terminates.