Bayesian non-linear matching of pairwise microarray gene expressions

Abstract

In this paper, we present a Bayesian non-linear model to analyze matching pairs of microarray expression data. This model generalizes, in terms of neural networks, standard linear matching models. As a practical application, we analyze data of patients with Acute Lymphoblastic Leukemia and we find out the best neural net model that relates the expression levels of two types of cytogenetically different samples from them.