Graph-Based Integration of Histone Modification Profiles

In this work, we introduce a similarity-network-based approach to explore the role of interacting single-cell histone modification signals in haematopoiesis—the process differentiation blood cells. Histones are proteins that provide structural support chromosomes. They subject chemical modifications—acetylation or methylation—that affect degree accessibility genes and, turn, formation different phenotypes. The concentration modifications can be modelled as continuous signal, which used build profiles. present profiles cell types involved haematopoiesis built based on all major (i.e., H3K27ac, H3K27me3, H3K36me3, H3K4me1, H3K4me3, H3K9me3) by counting number peaks signals; then, compute modification-specific similarity networks among considered As come signals, applied network fusion technique integrate these unique graph, with aim studying simultaneous effect for determination permit defining graph-cut-based separation score evaluating homogeneity subgroups corresponding myeloid and lymphoid phenotypes classical representation haematopoietic tree. Resulting scores show into reflects actual haematopoiesis.