Fig. 2.

RECAP recalibrates peak callers' P-values to a near-uniform distribution. (A) Log-log plot of the empirical CDF of recalibrated P-values for MACS, SICER and diffReps, on the simulated, 10% foreground, null-hypothesis data. (B) Similar plot for a representative ChIP-seq replicate pair from ENCODE. (C) Reductions in deviation statistic, which measures difference from the uniform distribution, for the RECAP-recalibrated P-values for several types of simulated data (10 datasets each) and 50 matched pairs of ENCODE replicate ChIP-seq data (10 pairs for each of five cell types)