Gene network analysis using measurement indices in microarray data

A great number of data mining methods have been widely made such as gene regulatory networks and gene set analyses to connect genes that reveal similar expression patterns. These methods generally fail to unveil gene-gene interactions in the same cluster. The aim of this study is to use several nonparametric correlation coefficient methods to transform the linear rank statistics into distance metrics on a Saccharomyces cerevisiae data set.

These nonparametric correlation coefficients, Kendall’s tau index and Gini rank correlation, were compared with common Pearson correlation method. The reliability and advantages of our proposed is satisfied using genetic website, http://www.yeast genome .org/. To address the interactions and characterize the gene–gene biological processes explicitly, the gene relationships are shown as a Pajek graph topology.

The results of biological interactions and characteristics demonstrated that the proposed nonparametric correlation coefficient methods have a strong capability to identify interaction genes. Moreover, suggested techniques could accurately detect the main genes and functional interactions in comparison to generally used Pearson correlation coefficient.

The two non-linear correlation coefficient techniques are proposed to measure the gene interactions more precisely.