Fig. 2: C.Origami accurately predicts 3D chromatin organization. | Nature Biotechnology

Fig. 2: C.Origami accurately predicts 3D chromatin organization.

From: Cell-type-specific prediction of 3D chromatin organization enables high-throughput in silico genetic screening

Fig. 2

a, Validation loss of models trained from different combinations of input features. Lower validation loss indicates better model performance. b,c, Experimental (b) and C.Origami-predicted (c) Hi-C matrices of IMR-90 on training (chr2), validation (chr10) and test (chr15) chromosomes. d, Input CTCF-binding and chromatin accessibility profiles. e, Insulation scores calculated from experimental (solid line) and C.Origami-predicted (dotted line) Hi-C matrices. Pearson correlation coefficients between prediction and target insulation scores are presented. f, Insulation score correlation between predicted and experimental Hi-C matrices across all windows in both validation and test chromosomes with Pearson (r) and Spearman (ρ) correlation coefficients. g, Chromosome-wide distance-stratified interaction correlation (Pearson) between prediction and experimental data. h, Comparison of model performance across Akita, DeepC, Orca and C.Origami using genome-wide insulation score correlation between prediction and experimental data from IMR-90 cells. Error bars in the violin plots indicate minimum, mean and maximum values. μ, average insulation correlation.

Back to article page