Inferring Transcriptional Modules from Microarray and ChIP-Chip Data Using Penalized Matrix Decomposition

Chun-Hou Zheng, Wen Sha, Zhan-Li Sun, Jun Zhang

Abstract: Inferring transcriptional regulatory modules is a useful work for elucidating molecular mechanism. In this paper, we propose a new method for transcriptional regulatory module discovering. The algorithm uses penalized matrix decomposition to model microarray data. Which takes into account the sparse a prior information of transcription factors–gene (TFs–gene) interactions. At the same time, the ChIP-chip data are used as constraints for penalized matrix decomposition of gene expression data. Finally the regulatory modules can be inferred based on the factor matrix. Experiment on yeast dataset shows that our method can identifies more meaningful transcriptional modules relating to specific TFs.