A bayesian network classification methodology for gene expression data

We present new techniques for the application of the Bayesian network learning framework to the problem of classifying gene expression data. Our techniques address the complexities of learning Bayesian nets in several ways. First, we focus on classification and demonstrate how this reduces the Bayesian net learning problem to the problem of learning subnetworks consisting of a class label node and its set of parent genes. We then consider two different approaches to identifying parent sets which are supported by current evidence; one approach employs a simple greedy algorithm to search the universe of all genes, and a second approach develops and applies a gene selection algorithm whose results are incorporated as a prior to enable an exhaustive search for parent sets over a restricted universe of genes. Two other significant contributions are the construction of classifiers from multiple, competing Bayesian network hypotheses and algorithmic methods for normalizing and binning gene expression data in the absence of prior expert knowledge. Our classifiers are first developed under a cross validation regimen against two publicly available data sets and then validated on corresponding out-of-sample test sets. The classifiers attain a classification rate in excess of 90 % on each of these out-of-sample test sets