Inferring an underlying reaction network from the data
Tuesday, May 13, 2008 - 2:45pm - 3:45pm
Consider the system of reaction rate equations (RRE) describing a chemical network with the reaction rate constants considered to be unknown parameters. The talk shall describe a statistical approach to identifying the most likely network from a given set of RRE coefficient estimates. The idea relies on mapping the estimated reaction constants into an appropriate convex region of a network stochiometric space in order to identify the reactions which are most likely to span that region. This approach reduces the original problem to inferring parameters of a certain multinomial distribution which may be solved using the general methods of algebraic statistics.