Seminer - Dr. Mehmet Tan - Scalability in Modeling and Control of Gene Regulatory Networks

 
Scalability in Modeling and Control of Gene Regulatory Networks
 
Mehmet Tan
 
Tarih: 25 Ocak 2010  Pazartesi  10.30
Yer: 2. Kat Toplantı Salonu
 
 
Abstract:
Gene regulatory networks model the interactions within the cell and thus it is essential to understand their structure and to develop some control mechanisms that could effectively deal with them. Scalability is the main issue as the number of genes in an organism is very large compared to the number of factors that can be handled in a graphical model. In this talk I will present two different algorithms that deal with scalability in modeling and control of gene regulatory networks. Constraint-based structure learning algorithms generally perform well on sparse graphs and it is true that sparsity is not uncommon. But gene regulatory networks also include some dense regions that raise some difficulties for the current modeling algorithms. Proposed modeling algorithm is capable of dealing with these dense regions and is based on a well-known modeling algorithm called the PC algorithm. Once a model exists, we can address the second problem, namely the control of gene regulatory networks for various applications. Again, as the number of genes that regulate a biological activity can be large, the curse of dimensionality is the main issue here. The algorithm we proposed for Factored Markov Decision Problems (FMDPs) increases the number of genes that can be handled in control of Probabilistic Boolean Networks (PBNs), where FMDPs avoid enumerating whole state space by representing the state transitions for control problems by compact models like dynamic Bayesian networks, and PBNs are one of the oldest and widely investigated methods for modeling gene networks. The effectiveness and applicability of the results are demonstrated on both synthetic and real biological data sets.
 
Biography:
Mehmet Tan graduated from Middle East Technical University Computer Engineering Department in 2000 with a high honour degree. He received his Ph.D. degree in Computer Engineering from Middle East Technical University in 2009. His Ph.D. dissertation focused on modeling, analysis and control of gene regulatory networks. During his Ph.D. he worked in Department of Computer Science at University of Calgary, AB, CA as a visiting scholar and instructor in 2007 and 2008. His research interests include machine learning and bioinformatics with a special focus on biological networks, Markov decision problems and drug discovery.