Acta mathematica scientia,Series A ›› 2015, Vol. 35 ›› Issue (5): 1018-1024.

Previous Articles    

Link Prediction for the Gene Regulatory Network

Li Yan1, Zhang Xiaofei2, Yi Ming3, Liu Yanyan1   

  1. 1 School of Mathematics and Statistics, Wuhan University, Wuhan 430072;
    2 College of Mathematics and Statistics, Central China Normal University, Wuhan 430079;
    3 Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071
  • Received:2014-12-10 Revised:2015-06-20 Online:2015-10-25 Published:2015-10-25

Abstract:

In order to build gene regulatory network, we proposed an algorithm based on gene expression levels and network anti-delivery. The algorithm used the network anti-delivery idea to analyze indirect effect produced from the traditional correlated calculation methods, and added an norm penalty term for controlling network sparsity after considering the sparsity of regulatory networks. We use the algorithm on the E. coli experimental data. This method improves the edge predictive ability of the correlation analysis on the regulatory network, Pearson correlation coefficient increased by 6.42%, Spearman correlation coefficient increased by 5.92%, mutual information improves 9.35%. Overall, this model provides a new idea on modifying a large number of system-related data, and can be applied to network edge prediction and the control dynamics of biological network inference.

Key words: Gene regulatory networks, Direct correlation, Indirect correlation, Pearson correlation, Spearman rank correlation, Mutual information

CLC Number: 

  • O29
Trendmd