The Construction Of Gene Regulation Network By Mathematical Models

Motivation: Along with the recent advancements in genome science, information on gene sequences has been exhaustively clarified. In the post-genomic era, interest has arisen regarding the elucidation of interactions between genes, especially gene regulatory network on expression level. Advances in molecular biological, analytical and computational technologies are enabling us to investigate systematically the complex molecular processes underlying biological systems. The DNA microarray is capable of profiling the expression levels of many genes simultaneously, and is a promising technology for the elucidation of gene interactions. And how to extract the regulatory information hidden in the millions of data points that result from the microarray experiments has become a problem that researchers are eager to resolve. There have been quite a few methods including sophisticated mathematical models that are used to analyze gene regulatory network. However, these methods have their own limitation respectively and the research in this area is still in its infancy. So in our research, we bring up a new mathematical model as an exploration.Method: â‘ Correlation analysis is one of the most welcomed algorithm which have been used in gene expression data analysis. In this research, a new algorithm named Fuzzy Correlation Coefficient was brought up to construct gene regulatory network. â‘¡Fuzzy correlation coefficient and three other existed methods(Pearson, Spearman and information entropy correlation coefficient)were used to analyze three microarray data sets which were about human gene expression during the development of CNS in embryonic period. To accomplish the construction of gene expression regulatory network, a program was completed with MATLAB package.Result: â‘ Gene regulatory networks were respectively constructed above a threshold by using the four types of correlation coefficients. And it was found that the genes in the same biological pathway or the genes with similar function could be linked in a same network. â‘¡Pearson and Spearman corr. coef. could reflect the tendency relationship between two genes, while information entropy and fuzzy corr. coef. could reflect the dependency relationship of regulation. â‘¢Within the four correlation coefficients,information entropy corr. coef. was the most comprehensive, and then was Spearman corr. coef. As to Pearson corr. coef. and fuzzy corr. coef., high related data was needed.Conclusion: â‘ It is feasible to apply fuzzy corr. coef. to construct gene regulatory network. â‘¡The scale of regulatory network could be controlled by adjusting the threshold of correlation coefficient, which could provide clues for latter research.