Regulation Study Between MiRNA And Target Gene In Disease Based On Gene Expression Data

miRNA(microRNA) is a small molecule RNA which plays a very important role of regulation in cance and tumor. miRNA is non-coding 19-24nt RNA, targeted on mRNA by the principle of complementary base pairing to inhibit the expression of mRNA or to promote the degradation of mRNA. The study of the function of miRNA in regulating human complex diseases is a hot topic and a tough question currently. The methods in the research of the relationship of miRNA and disease and the target of miRNAs include sequence alignment and biological statistics, and all these methods generally take a lot of time and money. Therefore, in order to overcome these shortcomings, we analyze the regulation relationship between miRNA and target mRNA in disease based on existing expression data.In this paper, we integrate the expression data of miRNA and mRNA from TCGA database. First, we analyze the correlation of miRNA and mRNA of Chromophobe renal cell carcinoma(RCC) with statistical methods and find the correlation difference between normal person group and RCC patient group in order to predict the miRNA which might regulate RCC. Second, the latest research shows that there are many miRNAs involved the pathophysiological process of Renal diseases, such as mir-155, mir-106 a, mir-106 b, mir-200 b, mir-200 c, mir-21, mir-210, mir141, mir182 and mir-192, et al, and we divide these miRNA into two categories, one is the miRNA which regulate RCC and the other one does not. And some features are selected from the expression data, trained, modeled and cross valided in the method of feature selection of SVM. At last, we classify miRNA and mRNA into two categories based on whether or not the miRNA is targeted on the mRNA with predict data download from Target Scan, and choose three features from the expression data of miRNA and mRNA,which are the mean of miRNA expression minus the mean of mRNA expression, the mean of miRNA expression divided by the mean of mRNA expression, the variance of miRNA expression divided by the variance of mRNA expression. And the features been trained, modeled and cross validated in the method of SVM in order to predict the target mRNA of miRNA. Our experiment result shows that there is a significant correlation difference to distinguish miRNAs which regulate RCC from all miRNAs, and there are also a high cross validation accuracy to predict the miRNA which regulate RCC and the target of miRNA in the method of SVM.All the above results show that the method of expression analysis of miRNA and mRNA is a very perspective direction in the research of miRNA. It not only has a higher accuracy in prediction but aslo overcomes the disadvantages of long period, high cost and uncontrollable in biological experimental. And with the increase of the data of miRNA and mRNA. The analysis of expression will provide more possibilities for the study on regulation of miRNA to disease.