Fig. 1

A signal-processing view of ChIP-seq data. Here, we show two ChIP-seq peaks (potential regions of enrichment) with different signal levels. Let us assume that after accounting for mappability and accessibility, what is left in the overall signal corresponds to true binding. Peak 1 is unlikely to be a genuine peak because only a small portion of the overall signal may correspond to binding. Peak 2, on the other hand, is much more likely to indicate true binding, as a relatively smaller share of the overall signal is attributable to mappability and chromatin accessibility, leaving a much larger portion for true binding