| Single Nucleotide Polymorphism (SNP) is a kind of DNA polymorphism caused by allele gene space mutation.It has already been proved by biology science that the relevance between SNP and human diseases is significant, which, if used wisely, would be a great help in foretelling the diseases even before some symptoms show. In medical research, it is an important way of providing not only a theoretical basis for the formation reasons of incurable diseases, but also in-depth diagnosis in the major diseases that threaten human health. Currently, the study is of great significance in the world frontier domain.In this paper, the maximum entropy method used in the SNP-disease association model is studied.For the problems of information crossing and insufficient experimental samples, data processing methods of bootstrap, stability calculation and decrossing algorithm are raised.Roles of the various methods played in the experiment are analyzed. Data expansion, stability calculation and information decrossing are done to the experimental data. At the same time, the applicability of the maximum entropy method in this paper is theoretically analyzed,and the SNP-disease association model is established using the maximum entropy method. Experiments using the SNP sample data on real cases are raised to obtain the pathogenesis probability of each SNP model, and the performance of the results is evaluated using the test samples. Optimization function of the bootstrap method using in the maximum entropy algorithm to the SNP-disease association model is verified by the experiment results. Performance evaluation algorithm is proposed to compare the maximum entropy results before and after the decrossing algorithm. The problems in the experiment are analyzed, and the SNP-disease association study in the next step is pointed out. |
