Adapt modern causal learning models to high-dimensional perturbation data as occurring in single-cell genomics and combine this with methods from computational biology to model gene regulatory interactions.