Fig. 1

Overview of the OnTAD pipeline. a OnTAD uses a sliding diamond-shaped window to calculate the average contact frequency within the window at each locus on the genome. The five loci marked by letters “a”–“e” are examples being evaluated as potential TAD boundaries, with “d” being a clear false positive. b Identification of candidate TAD boundaries in OnTAD. Blue curve—the average contact frequency of the diamond-shaped windows, calculated at different window sizes (W) and different loci. Red arrows—the location of significant local minimums of the average contact frequency, i.e., candidate TAD boundaries. c OnTAD assembles candidate boundary pairs using a dynamic programming algorithm (see the “Methods” section) d Visualization of the final output from OnTAD. In the genome browser, the identified hierarchical TAD is displayed as a series of horizontal bars, where each (sub) TAD is represented as a horizontal bar colored according to its TAD level