Inferring Genetic Regulatory Networks with an Hierarchical Bayesian Model and a Parallel Sampling Algorithm

Bayesian Networks (BNs) are used in a wide range of applications, being the representation of regulatory networks a recurrent one. Nowadays great interest is dedicated to the problem of inferring the network´s structure solely from the data. Aiming more precise results, the inclusion of extra knowledge in the inference process has been already suggested, as well as a Bayesian coupling scheme for learning genetic regulatory networks from a combination of related data sets which were obtained under different experimental conditions and are therefore potentially associated with different active sub-pathways. Furthermore, this approach has been combined to a MCMC sampling scheme and it has been verified that due to the complexity of the model, the MCMC suffered from poor convergence. We now propose the use of a Metropolis Coupled Markov Chain Monte Carlo (MC)³ algorithm in order to improve the mixing and convergence of the inference process.