Differentially Expressed Gene Analysis Of Citrus Huanglongbing Based On Maximum Information Coefficient

Citrus huanglongbing is a serious disaster in citrus production.It is called citrus "cancer" and has caused tremendous damage to the global citrus industry.Through the differentially expressed genes,it could insight the influence of the Candidatus liberibacter on biological processes and gene functions,and identify the genes associated with the disease.In this study,we introduce a novel nonparametric statistical method,maximum information coefficient,into the differentially expressed genes identifying in Huanglongbing.Two benchmarks,Rank Products and Wilcoxon's signed-rank test,are employed to test the novel method.In this paper,we discuss the feasibility and availiability of employing maximum information coefficient to identify the differentially expressed genes in Citrus huanglongbing.Our experiments based on the citrus huanglongbing gene expression data set GSE33459 show that the differentially expressed genes identified by the novel method include the same risk genes identified by the benchmarks.And the method also found 8437 and 8892 genes in 5-9 wai data set more than the two benchmarks,Wilcoxon's signed-rank test,and Rank Products.Similarly,in 13-17 wai data set,the numbers are 10189 and 11611 respectively.Although the numbers of differentially expressed genes are significantly different among the three methods,the ratio of the GO analyzes results of maximum information coefficient is almost same with the bechmarks',indicating that the novel method is not biased in identifying differentially expressed genes.In addition,the further experiments on the citrus huanglongbing gene expression data set GSE29633 support the conclusion.Our studies show that it is feasible and avaliable to identify the differentially expressed gene in citrus huanglongbing by using the maximum information coefficient.