High changing curvature regions detect chromatin enrichment in single cell data

Chromatin organization is nowadays accessible from population and single cell Hi-C data in the form of large contact matrices. Yet reconstructing the three-dimensional chromatin structure remains challenging and relies on polymer models and numerical simulations to account for these Hi-C data. Here we propose a novel optimization algorithm to identify cross-linker ensemble reproducing the experimental contact matrix. Furthermore, based on the polymer configurations extracted from the optimization procedure, we identify high changing curvature (HCC) regions in the chromatin, detected both in population and in single cell Hi-C, that we also compare to Topological Associated Domains (TADs). We report here that the HCC regions correlate with gene expression and CTCF high density distribution. Interestingly, the HCC region locations are heterogeneous across one cell repeats, revealing cell-to-cell variability. To conclude, HCC regions appear both in single and population Hi-C polymer reconstruction and can provide a possible unit for gene regulation.