Reconstruction of Signaling Network Topology from Steady State and Dynamic Perturbation Data

The first talk of the Bioinfonet Series will be given by Assoc. Prof. Dr. Tolga Can.

Dr. Can has a bioinformatics research group at one of the Turkey's top notch universities, Middle East Technical University, Computer Engineering Department.

As the ISCBSC RSG Turkey crew, we are proud to invite the bioinformatics research community to this inspirational talk.

"Given a set of proteins of interest in a signaling pathway, a common problem is to elucidate the signaling interactions and the dynamics of these interactions between these proteins. A first step towards an accurate solution is to figure out the underlying network topology; before working out the details of the dynamics of these interactions using biochemistry or structural biology. Several experimental techniques, such as RNAi screens, are designed to infer the structure of signaling networks. Existing solutions, either manual or automated, usually suffer from the scalability problem: they are not practical for signaling networks containing more than 20-30 proteins. In our research group, we target this problem and have developed a couple of algorithms in the past to construct larger scale signaling networks. In my talk, I will first show that reconstructing signaling networks from perturbation data is indeed a difficult problem. I will then propose two heuristics to solve the problem approximately and show that the approximate solutions are biologically acceptable. Finally, I will discuss our recent formulation which utilizes available time series data in reconstruction of the network topology. Our results show that, time series data is very valuable in scaling the solution easily to networks of larger sizes."