Orateur
Sophie LEBRE
Description
The growing availability of diverse omics datasets motivates integrative approaches for gene regulatory network inference.
Regression-based methods for gene regulatory network inference (Inferelator, GENIE3, DynGENIE3, IRafNet) identify key regulators and can incorporate prior knowledge to guide variable selection.
We propose a method to tune the strength of prior knowledge integration in regression models such as Lasso and Random Forests, using null hypothesis simulations to balance prior information with data-driven inference.