On Construction of Sparse Probabilistic Boolean Networks from a Prescribed Transition Probability Matrix

Probabilistic Boolean Networks (PBNs) are useful models for modeling genetic regulatory networks. In this paper, we propose efficient algorithms for constructing a sparse probabilistic Boolean network when its transition probability matrix and a set of possible Boolean networks are given. This is an interesting inverse problem in network inference and it is important in the sense that most microarray data sets are assumed to be obtained from sampling the steady-state.