Figure 2. 14 probe sets optimally separate patients into good and poor survival groups.

(A) Experimental strategy to identify an optimal signature to separate patients with BLBC into high and low risk groups. (B) Comparison of relative risk between leave-one-out cross-validation predicted high and low risk groups for n length signatures (n = 1,2,3…,50). 14 probe sets produces maximal risk separation between high and low risk groups (blue arrow).