Modelling molecular mechanisms at cellular and multicellular scales with multilayer networks

Multicellular organisms require precise coordination across many cell types. A cell’s identity arises from the interplay of multiple molecular layers that collectively determine its state and function. Recent advances in single-cell sequencing enable the profiling of these “omics” modalities at the single-cell resolution. Yet transforming these multimodal datasets into mechanistic insight remains a challenge. Even integrative approaches that combine several modalities still ignore the full spectrum of available data. We introduce two flexible frameworks - HuMMuS and ReCoN - to study cellular and multicellular systems, respectively, using heterogeneous multilayer networks (HMLNs). This structure preserves modality-specific information in distinct layers, connected through inter-layer edges. It enables the reconstruction of comprehensive regulatory maps spanning intracellular regulation and cell communication, and their exploration via random walks to generate different regulatory hypotheses. It can notably identify regulators of specific genes or drivers coordinating different cell types in complex responses.