A functional canonical correlation analysis approach to reverse engineering of in silico gene regulatory networks
BackgroundReverse engineering of gene regulatory networks via gene expression profiles across a series of perturbations is one of the most important topics in systems biology. Numerous methods have been proposed to deal with this problem, including Pearsons correlation coefficient (PCC), first and second order partial Pearsons correlation coefficient (O1PCC and O2PCC), entropy maximization (EM), mutual information (MI), least squares (LS) solution of ordinary differential equations (ODEs), total least squares (TLS) of ODEs, and many others. However, most of these approaches mainly focus on mining the structures of networks from steady state data. Although some methods have been trying to make use of the information of time series data, most of them deal with the data in the way of multivariate statistical analysis which ignores the latent functional patterns of the data. Furthermore, few of these methods are devoted into inferring large-scale regulatory networks of hundreds of genes. In usual cases of large-scale networks, the number of observations is less than the number of genes, raising ill-posed problems. Therefore, in this study we introduce the functional data analysis 1 into reverse engineering of large-scale regulatory networks from time series data.
Feng Zeng Xuebing Wu Xuegong Zhang Rui Jiang
MOE Key Laboratory of Bioinformatics and Bioinformatics Division,Tsinghua National Laboratory for Information Science and Technology/Department of Automation Tsinghua University, Beijing 100084, China
国际会议
The 7th Asia-Pacific Bioinformatics Conference(第七届亚太生物信息学大会)
北京
英文
894
2009-01-01（万方平台首次上网日期，不代表论文的发表时间）