An efficient convex nonnegative network component analysis for gene regulatory network reconstruction

Abstract

A systems biology problem of reconstructing gene regulatory network from time-course gene expression microarray data via network component analysis (NCA) is investigated in this paper. Inspired by the idea that each column of the connectivity matrix can be estimated independently, we try to propose a fast and stable convex approach for nonnegative NCA (nnNCA). Compared with the existing method, our new method reduces the computational cost substantially, whereas maintains a reasonable accuracy. Both the simulation results and experimental results demonstrate the effectiveness of our method. © 2009 Springer Berlin Heidelberg.