Statistical Methods for Exploring Metabolic Networks
September 26, 2011
11:00 a.m.
Erkki Somersalo
Abstract
One of the great challenges in medical research is to understand the functioning of metabolic systems and how the genes express themselves on the phenotype level. In this talk we take a look at complex metabolic networks in various organs such as brain, liver and skeletal muscle, and some of the mathematical models for analyzing them. Characteristic to the models is a large number of parameters that need to be inferred on based on scarce data that is often averaged over a population. In addition, the literature contains a wealth of complementary information that is often given in an indirect and qualitative form. The Bayesian statistical framework provides a natural basis to implement the information into the models, and the computational challenge is to explore the resulting posterior densities using Monte Carlo methods such as MCMC. The presentation reviews some of our work in this field and points out some findings in which the statistical framework has been of crucial help.