Figure 1.
Promoter–enhancer pairs with overlapping windows in training and validation data sets. (A) Schematic illustration showing how information could be shared between training and validation data sets because of overlapping EP windows. (B) Distribution of distances between boundaries of overlapping EP windows. For each EP pair, we found the window of another EP pair, so that the distance between boundaries of their windows (d = D1 + D2) is minimal. Histogram shows distribution of the obtained values of d.
