Information theoretic inference of direct gene regulation from time series data

This paper considers the problem of inferring the oriented connectivity of genetic regulatory networks based on the time series gene expression profiles produced by DNA microarrays. The inference algorithm is implemented based on the minimum description length (MDL) principle. Simulation results show that for synthetic networks the proposed algorithm achieves a good performance, while in the presence of realistic time series data it is capable of providing novel insight into the inner structure of various regulatory circuitries.