Bayesian dynamic linear models for predicting temporal gene expression profiles

Abstract

Prediction of gene and protein dynamic behavior is a challenging and important problem in genomic research while estimating the temporal correlations and non-stationarity are the key research challenges in this process. Unfortunately, most existing techniques used for the inclusion of the temporal correlations treat the time course as evenly distributed time intervals and use stationary models with time-invariant settings. This assumption is often violated in microarray time course data since the time course expression data are at unequal time points, where the difference in sampling times varies from minutes to days. Furthermore, the unevenly spaced time courses with sudden changes make the prediction of genetic dynamics more difficult. In this paper, we develop two types of dynamic Bayesian models to tackle this challenge for predicting the gene expression profiles associated with diseases. In the multiple univariate time-varying dynamic linear model, we treat both the stochastic transition matrix and the observation matrix time-variant with linear setting. In the multivariate model we include temporal correlation structures in the covariance matrix estimations. We apply our models to multiple tissue polygenetic affymetrix data sets. Preliminary results show that the predictions of the genomic dynamic behavior can be well captured by the proposed models.