We address two major challenges for computational gene regulatory network (GRN) inference.

• The first challenge is that the topological structures of gene regulations are context-dependent. Different interactions of gene activities will be active in different experimental conditions (e.g., culture media, temperatures, pH), leading to a different topological structure of the inferred GRNs.
• The second one is that aggregating gene expressions with their regulatory networks to infer new relationships in a supervised learning paradigm.