HMM用于G蛋白偶联受体超家族的识别  

Classifying G-protein Coupled Receptors with Hidden Markov Models

作　　者：侯永丰[1] 李通化[1] 

机构地区：[1]同济大学化学系,上海200092

出　　处：《同济大学学报（自然科学版）》2004年第12期1696-1700,共5页Journal of Tongji University:Natural Science

基　　金：国家自然科学基金资助项目 (2 0 2 75 0 2 6)

摘　　要：采用了一种HMM(隐马尔可夫模型 )的方法用于G蛋白偶联受体超家族层次各类别之间进行识别 ,具体考虑了ACDE与B类超家族 ,以及BCDE与A类超家族的分辨 ,取得了不错的效果 ,类之间的识别准确率可以达到10 0 % .研究过程中 ,考虑了G蛋白偶联受体一级结构信息和数据自身特性 ,合理地利用了G蛋白偶联受体一级结构序列的不等长特性 ,同时在模型的预测方面采用了双模型的预测机制 .G protein coupled receptors are the most numerous types of cell proteins in the body and are the target agents for approximately 50 percent of the drugs on the market today. We applied hidden Markov model in classifying G protein coupled receptors, tested the discrimination of ACDE and B superfamily , together with BCDE and A superfamily concretely, and a good result was achieved.Hidden Markov models can integrate G protein coupled receptors data features into a model.Experiment results suggest hidden Markov models is well fit to classifying G protein coupled receptors with length-variable primary sequences.

关 键 词：G蛋白偶联受体 隐马尔可夫模型 分类 信息系统 

分 类 号：Q-332[生物学]