Parallel computing for gene networks reverse engineering

Abstract

Gene networks provide a mathematical representation of gene interactions that govern biological processes in every living organism. Given a gene expression data, the goal of network inference is to reconstruct the underlying regulatory network. The problem is challenging owing to the convoluted nature of biological interactions and imperfection of experimental data. In many cases, the resulting computationalmodels are too complex to execute on a sequential computer and require scalable parallel approaches. In this chapter, we describe network inference methods based on information theory and show a parallel algorithm that enables whole-genome networks reconstruction.