Challenges in Reverse Engineering of Gene Networks from Algebraic Perspectives

Abstract

Reverse engineering of complex gene networks is a significant step towards understanding various biological processes. In this work, a novel algorithm for reverse engineering gene networks on a genome-wide scale using a noisy gene expression data is proposed. Under the proposed scheme, the challenges in reverse engineering of gene networks from these gene expression datasets are highlighted. Also, the parameter space describing gene interaction is partitioned into estimable and inestimable linear subspaces. The estimable linear subspace is obtained by using principal components analysis (PCA), the Akaike information criterion (AIC), and jackknifing. Furthermore, the approach is tested and validated using a simulated gene network model.