Structure learning for biomolecular pathways containing cycles

Bayesian network structure learning is a useful tool for elucidation of regulatory structures of biomolecular pathways. The approach however is limited by its acyclicity constraint, a problematic one in the cycle-containing biological domain. Here, we introduce a novel method for modeling cyclic pathways in biology, by employing our newly introduced Generalized Bayesian Networks (GBNs) and proposing a structure learning algorithm suitable for the biological domain. This algorithm relies on data and perturbations which are feasible for collection in an experimental setting, such as perturbations affecting either the abundance or activity of a molecule. We present theoretical arguments as well as structure learning results from simulated data. We also present results from a small real world dataset, involving genes from the galactose system in S. cerevisiae.