Improved Methods for Inferring Regulatory Networks from Temporal Expression Data

Over the past few years, the advent of microarray technology has enabled the simultaneous measurement of the expression levels of thousands of genes. When the expression levels of these genes are measured at multiple time points during an experiment, the result is a temporal expression profile. These expression profiles may be processed to extract the underlying gene regulatory network relationships. This paper broadly reviews the methods available for exploring time-series expression data. It then focuses on some of the inherent problems of using correlation-based methods (in particular), after which two recent methods are described that attempt to overcome these problems. Finally, preliminary results are presented for a new similarity measure and optimizing technique intended to overcome these same problems. The yeast cell cycle data of Spellman et. al. [26] is used to evaluate these new methods.