A parallel solution to reverse engineering genetic networks

Abstract

The reverse engineering problem is the problem of, given a set of measurements over a genetic network, determine the function that fits the data. In this work we develop a solution to this problem that uses a finite field model and that takes advantage of the efficient algorithms for finite field arithmetic that have recently been developed for other applications such as cryptography. Our solution, which is very efficient for the very large networks that biologists would like to consider, is given by a univariable polynomial which is determined by a parallel version of Lipson's interpolation algorithm. © Springer-Verlag Berlin Heidelberg 2004.