An efficient method for fuzzy identification of regulatory events in gene expression time series data

The increase in the number of time-series gene expression experiments drives an increase in the development of analysis methods that aim to reverse engineer genetic regulatory networks. Methods have previously been developed that test all combinations of gene interactions in a gene expression time-series dataset for those that closely fit a model of gene regulation. We describe an efficient method for reducing the computational cost of this approach by pruning the search space of candidates least likely to fit a regulatory model. Our approach represents changes in expression levels as directional slope information and discards combinations whose slope relationships are most likely outside the scope of the given model of regulatory relationships. We demonstrate a 70% reduction in the computational time required to process three S. cerevisiae microarray time-series datasets.