摘要
采用一种基于关联规则与功能标签的推荐算法,对蛋白质功能进行预测。根据蛋白质序列的循环排列匹配关系,构建蛋白质相关联数据集;应用此数据集里已知蛋白质的功能标签,搜索蛋白质数据库,寻找与待预测蛋白质相似的已知蛋白质,进一步预测未知蛋白质的功能;最后,通过准确率、召回率以及F1-measure三个指标来衡量结果,并对比传统的直接推荐注释方法,验证结果的有效性。
This paper used a recommended algorithm based on association rules and functional label for protein function prediction. According to the cyclic arrangement of protein sequence, it structured the protein associated data set. It used the data of the functional label of known protein tag to search for the protein database. It found the known protein which was similar to the unknown protein further to predict the functions of the unknown protein. Finally, it compared the traditional method of direct recommendation algorithm with traditional method by the accurate rate, recall rate and F1-measure measurements. The method was proved to be of feasibility.
出处
《莆田学院学报》
2015年第2期40-43,共4页
Journal of putian University
基金
国家自然科学基金面上项目(61472082)
福建省自然科学基金项目(2014J01220)
关键词
蛋白质功能
推荐算法
生物信息学
蛋白质序列
关联规则
protein＇s functions
recommendation algorithm
Bioinformatics
protein sequence
association rules
