Abstract
We review recent Bayesian network inference methodologies we developed to infer genetic and metabolic pathways associated to oncological drug chemoresistance. Bayesian inference is supported by a rigorous and widely accepted mathematical formalization of predictive analytics. It is an inherently integrative approach allowing the incorporation of prior knowledge and constraints. Moreover, it is recommended to treat noisy data, and large amount of data whose dynamics laws are mostly unknown. We focus on variational Bayesian methods for the inference of stochastic reaction processes and we present a compendium of the recent results of inference of gene and metabolic networks presiding at the development of pancreas cancer resistance to gemcitabine.