Global Landscape Of 3d Architectural Anchors From Genome-Wide Analysis Of Specific Chromatin Interactions

Chromatin folding and interactions play essential roles in fundamental processes such as regulation of gene expression and cellular specialization. While Hi-C measurements often detect numerous interactions (∼108), our previous results showed that only 2-9% of Hi-C interactions, which we call specific interactions, are required to generate ensembles of 3D single-cell chromatin conformations that collectively exhibit the observed Hi-C contact patterns. In this study, we carried out a genome-wide comprehensive analysis of these specific interactions. We use a fractal Monte Carlo sampling method to construct an ensemble of random self-avoiding chromatin polymers under nuclear confinement, which are used as our physical null model. This randomly sampled polymer ensemble is then used to standardize the measured Hi-C frequencies in a novel fashion. We estimate the statistical significance (p-values) of each Hi-C interaction by comparison with bootstrapped random polymer chains. Hi-C interactions with FDR-adjusted p-values below a certain threshold are then identified as specific interactions. Our results show there are certain small genomic regions that are enriched for these specific interactions, which we term as architectural anchors. In the Hi-C heatmaps, these appear as horizontal and vertical stripes. We report the results of analysis on the specific interactions at these architectural anchors. We further discuss how they are related to epigenetic and other biological markers.