利用多样性增量位置得分函数预测人类5'非翻译区剪接位点  

PREDICTION OF SPLICE SITES IN HUMAN 5'UTRs BY USE OF POSITION SCORE FUNTION BASED ON INCREMENT OF DIVERSITY

作　　者：陈丽萍[1] 吕军[1] 

机构地区：[1]内蒙古工业大学理学院,呼和浩特010051

出　　处：《内蒙古工业大学学报（自然科学版）》2009年第4期274-278,共5页Journal of Inner Mongolia University of Technology：Natural Science Edition

基　　金：内蒙古工业大学校重点基金项目(ZD200607)

摘　　要：5’非翻译区中的剪接位点两侧不存在由编码区到非编码区的状态转换,所以通常的识别剪接位点的算法在非翻译区的性能不太理想.本文把多样性增量的位置得分函数应用到5’非翻译区剪接位点的识别中.对于供体端,正负集样本数之比为1∶17,识别敏感性为66.91%,阳性预报值为68.54%,总精度为96.45%,ROC曲线下面积为97.23%;对于受体端,正负集样本数之比为1:24,识别敏感性为77.19%,阳性预报值为29.37%,总精度为91.78%,ROC曲线下面积为93.91%.这一结果要好于已有相似算法.As there exists no translation from protein coding to non-coding in human 5＇ untranslated regions （ITRs）,conventional splice site prediction methods do not perform well with UTRs. In this paper, position score function based on increment of diversity is used to predict splice sites in 5＇UTRs. Results show that with the donor sites,the positive set in proportion to the negative is 1 ： 17,the sensitivity--66.91% ,the precision--68.54% ,the accuracy--96.45%and the area under the Receiver Operator Characteristics curve--97.23 %. While with the acceptor sites ,the positive set in proportion to the negative is 1 ： 24,the sensitivity--77. 19% ,the precision--29. 37N,the accuracy- 91.78% and the area under the Receiver Operator Characteristics curve--93. 91%. Keyworfls： 5＇ untranslated regions;recognition of splice sites;position score fanction base on increment of diversity

关 键 词：5’非翻译区 剪接位点识别 多样性增量位置得分函数 

分 类 号：Q61[生物学—生物物理学]