Fig. 3
From: In silico prediction of high-resolution Hi-C interaction matrices
HiC-Reg cross-chromosome performance. a The distance-stratified Pearson’s correlation plot when training on one chromosome and testing on a different chromosome from the same cell line. Shown are the distance-stratified correlations for models trained on chromosome 9 and 11 (row groups) and tested on chromosome 9, 11, 14, 17, 19 (columns) at 5 kb resolution in different cell lines. Each pair of rows corresponds to a particular cell line and results are shown for five cell lines: Gm12878, K562, Huvec, Hmec, Nhek. The blue line refers to training and testing on the same chromosome in cross-validation mode. b Heatmap of AUC for all pairs of tested cross-chromosome experiments. Each off-diagonal entry in the heatmap denotes the AUC when trained on the row chromosome and tested on the column chromosome. The more red an entry the better the performance in that chromosome pair combination. The diagonal entries are the AUC values when training and testing on the same chromosome in cross-validation mode. Source data are provided as a Source Data file.
