基于打分函数的蛋白质二级结构的识别  

Identification of Protein Secondary Structure Based on Scoring Function

作　　者：冯永娥[1] FENG Yong-e(College of Science, Inner Mongolia Agriculture University, Hohhot Innermagolia 010018 China)

机构地区：[1]内蒙古农业大学理学院,内蒙古呼和浩特010018

出　　处：《生物数学学报》2016年第4期455-460,共6页Journal of Biomathematics

基　　金：内蒙古自治区高等学校科学研究项目(No.NJZY067);内蒙古农业大学基础科研基金(No.JC2013004)资助项目

摘　　要：本文首次将基于位置权重矩阵的打分函数用于蛋白质二级结构的预测中.我们选取CB513数据库作为基准数据库,首先在库中截取11残基和21残基片段,依据中心残基的二级结构类型分成3个集合;然后分别建立20种氨基酸、以及依据亲疏水性约化成的6种氨基酸在3个集合中的位置权重矩阵;对于任意一个待测的序列片段X通过和3个位置权重矩阵比较,应用打分函数得到3个不同的分值,比较哪个分值最大X就属于哪一类;最后在误差允许范围内对预测结果进行修正,得到的预测精度Q_3最高达到了80%.Based on scoring function of position weight matrix database for predict pro- tein secondary structure in CB513 database in this paper. Firstly, we selected 11-residue and 21-residue fragments respectively and grouped those into three dataset （alpha, beta, Coil） accord- ing to the secondary structure of center residue; Secondly, we constructed three position weight matrix by 20 amino acids and 6 amino acids in fragments, for any one to test will be predicted the maximal one among three scores, which is calculated by the scoring function for three position weight matrix; Finally, The predicted results are corrected in the range of allowable error, and the prediction accuracy of Q3 reached 80%.

关 键 词：蛋白质二级结构 位置权重矩阵 打分函数 误差允许范围 

分 类 号：Q61[生物学—生物物理学]