基于DNN位点的选择和验证  

SPN Selection and Validation Based on DNN

作　　者：张艳[1] 倪继锋[1] 魏浩然[1] 茅宏伟[1] 

机构地区：[1]上海师范大学信息与机电工程学院,上海200234

出　　处：《计算机仿真》2017年第7期335-338,399,共5页Computer Simulation

基　　金：上海市青年科技英才扬帆项目(14YF1409300);上海高校青年教师培养资助项目(ZZshsf14026)

摘　　要：定位与性状相关联的位点在染色体或基因中的位置,能够帮助研究人员了解一些疾病的遗传机理,从而达到防止一些遗传病发生的目的。结合统计学与机器学习知识对疾病与DNA致病位点的分析,采用随机森林算法、深度学习算法构建数据分析模型,实现了致病位点与单个疾病的关联性分析、致病基因片段的匹配、以及10个关联性状与致病基因位点的分析。进而采用深度神经网络对所得到的致病位点进行分析和检验,仿真结果表明所得位点准确、合理。Locating the traits associated with SPN in a chromosome or gene can help researchers understand the Genetic mechanism of some diseases to prevent the occurrence of some genetic diseases. Combining statistical and machine learning to analyze the disease and DNA pathogenic SPN, a random forest algorithm and a deep learning al- gorithm were used to construct the data analysis model, and the association analysis between pathogenic SPN and sin- gle disease, matching of pathogenic gene fragment, as well as 10 Related traits and pathogenic SPN analysis were re- alized. Furthermore, deep neural network and explicit rate were used to analyze and test the pathogenic SPN. The simulation results show that the obtained SPN is accurate and reasonable.

关 键 词：随机森林 深度神经网络 位点 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]