Gene regulation is at the core of modern predictive medicine. As such, gene regulatory networks constitute a crucial foundation for the generation of upstream causative predictions for RNA-Seq data based on transcription factor (TF) – promoter regulations. Yet, the inference of such networks is challenging due to the highly nonlinear and cell-type-specific machinery. Thus, linear and non-linear approaches to reverse-engineer gene regulatory networks probably capture different aspects of the system.This project aims to address the space between systems biology modeling and machine learning by formulating a biologically constrained machine learning setup to identify gene regulatory interactions. The end goal is a framework for drug re-purposing and disease mechanism identification.