Fig. 2

DCA-scapes of N proteins reveal known sites of association in vivo. a The structure of P22 N (PDB 1A4T [https://www.rcsb.org/structure/1A4T]) bound to box B motif is shaded according to direct information (DI) score. Yellow bases have the highest score while darker colors denote lower scores. A cartoon of the box B structure is shown to indicate the position of shaded sites in the 3D coordinates. The DI landscape indicates coupled relationships in the box B motif. b The structure of λ N (PDB 1QFQ [https://www.rcsb.org/structure/1QFQ]) bound to box B, a cartoon indicating specific coupling interactions observed on the DI landscape. c Discovery of in vivo-binding sites for P22 N based on Hamiltonian scores. Hamiltonian scores (Y-axis) across the genomic region encompassing the operator region are calculated based on estimated global probability parameters using DCA and provide a quantitative evaluation of binding in different genomic regions. Negative values indicate a higher probability of association and are shaded according to the inset color bar. d Predicted binding sites for λ N following the same procedure as in c