Learning Yeast Gene Functions from Heterogeneous Sources of Data Using Hybrid Weighted Bayesian Networks