Statistical and deterministic methods for reverse engineering biological pathways

While current high-throughput technologies are limited in resolution and scope, future advances could allow for the simultaneous measurement of a multitude cellular signaling components (metabolites, proteins and mRNA). When such technologies become available, the ability to "reverse engineering" cellular pathways from measurements of components concentration alone becomes a possibility. That is, time series of component signals could be used to infer the wiring diagram of the pathways. While important techniques to reverse engineering cell signaling have already been published, much work remains to be done. We will discuss on our research in this field including methods that can be divided into two classes. In one class, which we call the statistical approach, we attempt to infer the topology of a pathway in terms of the statistical associations between its components, without any attempt to infer the causal laws that govern the dynamics. These statistical associations include Bayesian methods, information theoretic methods, conditional expectation methods, and graph theoretic ideas. The second approach, which we call deterministic, attempts to deduce the kinetic interactions between components. We assume the component concentrations can be represented by state equations where the right hand sides are drawn from a limited class of functions. With this approach, the task of pathway reconstruction reduces to an optimization problem within the given class of functions. We have tested our methods using simulated data coming from simple kinetic models to more intricate models such as the yeast cell-cycle model.