Bivariate models for identifying differentially expressed genes in microarray experiments Deepak Mav and N. Rao Chaganty Department of Mathematics and Statistics Old Dominion University Norfolk, VA 23529-0077 In this paper we present two hierarchical bivariate probability models for the joint distribution of the red and green intensities in cDNA microarrays. The first is a bivariate distribution, with gamma marginals. We assume gamma priors for the scale parameters. Here we use the Bayes estimates of the ratio of the mean intensities to select differentially expressed genes. The second model incorporates in the first model a latent Bernoulli variable, which indicates the presence or absence of differential expression. Here we use the EM algorithm to calculate the posterior probabilities of gene expressions. Using Escherichia coli microarrays data, we show that both models, which account for the correlation between the intensities, are improvements of the models suggested in a seminal paper by Newton et. al. (2001). |
Graybill Conference |