基于模糊理论预测CpG岛新方法  

A new method to predict CpG-islands based on fuzzy theory

作　　者：王金龙[1] 苏建忠[1] 王芳[1] 崔颖[1] 张岩[1] 

机构地区：[1]哈尔滨医科大学生物信息科学与技术学院,哈尔滨150081

出　　处：《生物信息学》2009年第2期91-94,共4页Chinese Journal of Bioinformatics

基　　金：黑龙江省自然科学基金(D2007-35);黑龙江省教育厅海外学人科研资助基金(1152hq28);哈尔滨市科技创新专项基金(RC2007LX003004)

摘　　要：启动子区域的CpG岛的异常甲基化是识别癌症的重要标志之一。目前已经建立的一些CpG岛的预测思想和方法都有自身的缺点,基于模糊理论的预测思想从贴进度的角度来判定CpG岛,能更容易地找到被以往方法所忽略的具有更多生物学意义的CpG岛。本研究通过构建属于CpG岛集合的隶属函数,计算候选序列的隶属度,找出所有的可接受隶属程度的CpG岛。将该方法应用于UCSC数据库中的一段序列(hg18-chr1-31618510-31623510--1000)进行预测,发现提高了预测的精确度。可见应用模糊理论预测CpG岛具有一定的可行性,利用选取不同的截集,可以得到更为精确的CpG岛。Abnormal methylation of CpG islands in promoter region of genome is an important marker to identify the occurrence of cancer. The methods of predicting CpG islands all suffered their different drawbacks. Theoretic prediction based on fuzzy idea could predict more CpG islands which have some biological function from the perspective of the approach degree. We constructed membership function of the set of CpG islands, and calculated the membership degrees of the candidate sequences to find all CpG island under the threshold of the membership degree.We used the naethod to predict the sequence （hg18.chr1 31618510.31623510 - 1000） from UCSC,and found that the precision of the prediction was improved. Prediction of CpG islands by fuzzy theory has a feasibility and we can find more precision of CpG island by choose different cut sets.

关 键 词：CPG岛 模糊理论 隶属函数 截集 

分 类 号：Q756[生物学—分子生物学]