摘要
蛋白质二级结构预测是生物信息学中最重要的课题之一。本文从氨基酸序列的原子信息出发,以6种原子化学位移的均值作为新的特征参量,应用二次判别法对蛋白质二级结构进行预测,取得了较好的结果。3折交叉检验得到预测总精度为77.46%,其中alpha螺旋、beta折叠的预测准确率分别达到92.36%和81.39%。在同样的数据库中,与其它预测软件相比较,显示了我们算法的优越性。
Protein secondary structure prediction is one of the most important topics in bioinformatics. In this study,we used six nuclei chemical shifts( CSs) as features combined with the method of quadratic discriminant analysis( for short QD) to predict protein secondary structure,and achieved better results. The overall accuracy of the three- fold cross- validated test is 77. 46%,and the accuracies of alpha helices,beta stands reached 92. 36%,81. 39% respectively. Compared with the other predictor,the results shown that our method more effective in the same database.
出处
《内蒙古农业大学学报（自然科学版）》
CAS
2015年第2期164-167,共4页
Journal of Inner Mongolia Agricultural University(Natural Science Edition)
基金
国家自然科学基金(No.31360206)
内蒙古农业大学草原英才项目
内蒙古农业大学博士科研启动基金(BJ08-30)
内蒙古农业大学基础科研项目基金(JC2013004)
关键词
蛋白质二级结构
化学位移
二次判别法
误差允许范围
Protein secondary structure
chemical shifts
quadratic discriminant analysis
in error allowed scope
