针对蛋白质复合体检测的自学习图聚类（英文）

发布时间：2018-04-09 00:33

本文选题：图聚类　切入点：蛋白质复合体　出处：《控制理论与应用》2017年06期

【摘要】：蛋白质复合体是由两条或多条相关联的多肽链组成,在生物过程中起着重要作用.假如用图表示蛋白质蛋白质相互作用(protein-protein interactions,PPI)网络数据,那么从中找出紧密耦合的蛋白质复合体是非常困难的,特别是在近年来PPI网络的容量大大增加的情况下.在本文中,通过对称非负矩阵分解,针对蛋白质复合体检测问题提出了一种图聚类方法,该方法可以有效地从复杂网络中检测密集的连通子图.并且将此方法和当前最先进的一些方法在3个PPI数据集中用同一个基准进行比较.实验结果表明,本文的方法在3个拥有不同大小和密度的数据集中均显著优于其它方法.
[Abstract]:Protein complex is composed of two or more associated polypeptide chains and plays an important role in biological processes.If we use graphs to represent protein-protein interactions (PPI) network data, it is very difficult to find tightly coupled protein complexes, especially when the capacity of PPI networks is greatly increased in recent years.In this paper, a graph clustering method is proposed for protein complex detection by symmetric nonnegative matrix decomposition, which can effectively detect dense connected subgraphs from complex networks.This method is compared with some of the most advanced methods in three PPI datasets using the same benchmark.The experimental results show that the proposed method is superior to other methods in three datasets with different sizes and densities.
【作者单位】：华南师范大学计算机学院;仲恺农业工程学院信息科学与技术学院;
【基金】：Supported by Natural Science Foundation of Guangdong Province,China(2015A030310509) National Science Foundation of China(61370229,61272067,61303049) S&T Planning Key Projects of Guangdong(2014B010117007,2015B010109003,2015A030401087,2016A030303055,2016B030305004,2016B010109008)
【分类号】：O157.5;Q51

【相似文献】