流形排序算法预测microRNA  

MicroRNA prediction based on manifold ranking

作　　者：王常武[1] 刘兵强[1] 王宝文[1] 刘文远[1] 

机构地区：[1]燕山大学信息科学与工程学院,河北秦皇岛066004

出　　处：《计算机应用研究》2012年第3期819-822,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60970123)

摘　　要：在已知microRNA(miRNA)较少的情况下,为了提高算法预测的准确性,提出一种基于流形排序的miR-NA预测算法。该算法采用加权图模型描述序列,使用置信传播分配排序分数,降低了算法的时间复杂度;算法根据大规模数据内部全局流形结构进行排序,提高了排序结果的准确性。在人类和按蚊全基因组范围内的实验证明,流形排序算法的预测效果优于传统的预测方法,可以作为预测miRNA的一个有效工具。In order to improve the precision of microRNA prediction while the number of known microRNAs is small, this paper proposed a novel microRNA prediction algorithm based on manifold ranking. The algorithm adopted the strategy of modeling microRNA prediction process as belief propagation on a weighted graph, hence reduced the time complexity of the algorithm. The core idea of algorithm was to rank the data with respect to the intrinsic manifold structure collectively revealed by a great amount of data, hence enhanced the accuracy of the ranking results. Experiments on H. sapiens and anopheles gambiae genes show that manifold ranking algorithm is better than the traditional algorithm, and can be worked as an effective tool for predicting novel microRNAs.

关 键 词：微小RNA 加权图 置信传播 流形排序 预测 生物信息学 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]