Algorithms for identification of clusters and gene regulatory networks

Tarek Mathew

This talk will be in two parts (time permitting). The first part will focus on an algorithm for identifying clusters based on a Gaussian mixture model for the similarity data (as in the CLICK algorithm). The resulting algorithm has a complexity similar to the k-means algorithm. The second part of the talk will focus on ongoing work on a probabilistic model for the evolution of mRNA populations within a cell. The mRNA populations are modelled as queues (Markov chains) whose rates of transcription are regulated by probabilistic generalizations of Boolean functions. A statistical inference procedure based on the EM algorithm will be described. As this work is in progress only sample numerical results will be presented.