Fig. 4From: Predicting the recurrence of noncoding regulatory mutations in cancerNumber of mutations whose recurrence was revealed by sample addition. The top-voted mutations from the random forest trained with the significant recurrence model (100 kbp or 10 kbp as the background window) were compared to the same number of the most highly recurrent mutations based on the raw or significance model. a-b Recurrence recalculated in the window of (a) 10Â bp and (b) 100Â bp after the addition of new breast cancer samples. c-d Recurrence recalculated in the window of (c) 10Â bp and (d) 100Â bp after the addition of lung cancer samplesBack to article page