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.
出处
《计算机应用研究》
CSCD
北大核心
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)
关键词
微小RNA
加权图
置信传播
流形排序
预测
生物信息学
microRNA
weighted graph
belief propagation
manifold ranking
prediction
bioinformatics
