Mini-Symposium MLCB 2009

Thursday, December 10, 2009, 13:30–16:15

13:30-14:25 Daphne Koller

Understanding Gene Regulatory Networks and Their Variations

A key biological question is to uncover the regulatory networks in a cellular system and to understand how this network varies across individuals, cell types, and environmental conditions. In this talk I will describe work that uses machine learning techniques to reconstruct gene regulatory networks from gene expression data. Specifically, we exploit novel forms of Bayesian regularized regression to enable transfer between multiple related learning problems, such as between different individuals or between different cell types. We demonstrate applications in two domains: understanding the effect of individual genetic variation on gene regulation and its effect on phenotypes including human disease; and understanding the regulatory mechanisms underling immune system cell differentiation.

14:25-15:20 Eric Xing

Time Varying Graphical Models: reverse engineering and analyzing rewiring networks

A plausible representation of the relational information among entities in dynamic systems such as a social community or a living cell is a stochastic network that is topologically rewiring and semantically evolving over time. While there is a rich literature in modeling static or temporally invariant networks, until recently, little has been done toward modeling the dynamic processes underlying rewiring networks, and on recovering such networks when they are not observable. In this talk, I will present a new formalism for modeling network evolution over time based on time-evolving probabilistic graphical models, such as TV-GGM, TV-MRF, and TV-DBN, and several new algorithms for estimating the structure of such models underlying nonstationary time-series of nodal attributes. I will show some promising results on recovering the latent sequence of evolving social networks in the US Senate based it voting history, and the gene networks over more than 4000 genes during the life cycle of Drosophila melanogaster from microarray time course, at a time resolution only limited by sample frequency. I will also sketch some theoretical results on the asymptotic sparsistency of the proposed methods.

15:20-16:15 Elizabeth Halloran

Novel Applications of Computational Biology in Infectious Disease Interventions

Interventions in infectious diseases are increasingly relying on computational biology and genomic methods. Estimating changes in viral genetic diversity in a population could be a new potential method to evaluate vaccination strategies in populations. Transgenic mosquitoes immune to a pathogen are being developed to replace the native mosquito vector of a number of vector-borne diseases. High throughput methods are being used to elucidate mechanisms of immune memory with the promise of developing better vaccines. Large-scale computer simulation models are useful for exploring interventions and could benefit from input from network and graph theory. In this talk, we discuss a few novel applications of computational biology in understanding infectious diseases and interventions.