NetCoffee: a fast and accurate global alignment approach to identify functionally conserved proteins in multiple networks
NetCoffee: a fast and accurate global alignment approach to identify functionally conserved proteins in multiple networks
复制标题
DOI：
10.1093/bioinformatics/btt715
复制
发表时间：
2014-02-15
期刊：
影响因子：
5.8
通讯作者：
Reinert, Knut
中科院分区：
文献类型：
--
作者：
Hu, Jialu;Kehr, Birte;Reinert, Knut
动机：由于高通量技术的最新进展，越来越多物种的蛋白质-蛋白质相互作用网络在公共数据库中可用。如何识别跨物种的功能保守蛋白质的问题在计算生物学中引起了很多关注。网络对齐提供了一个系统的方法来解决这个问题。然而，大多数现有的对准工具在解决这个问题时遇到限制。因此，对更快，更有效的比对工具的需求正在增长。结果：我们提出了一个快速，准确的算法，NetCoffee，它允许找到一个全球性的多个蛋白质相互作用网络的比对。NetCoffee通过在一组加权二分图上使用模拟退火最大化目标函数来搜索全局对齐，这些加权二分图是使用类似于T-Coffee的三重方法构建的。为了评估其性能，NetCoffee应用于四个真实的数据集。我们的研究结果表明，NetCoffee弥补了以前算法的几个局限性，在速度方面优于所有现有的对齐工具，但仍能识别具有生物学意义的对齐。
Motivation: Owing to recent advancements in high-throughput technologies, protein-protein interaction networks of more and more species become available in public databases. The question of how to identify functionally conserved proteins across species attracts a lot of attention in computational biology. Network alignments provide a systematic way to solve this problem. However, most existing alignment tools encounter limitations in tackling this problem. Therefore, the demand for faster and more efficient alignment tools is growing.Results: We present a fast and accurate algorithm, NetCoffee, which allows to find a global alignment of multiple protein-protein interaction networks. NetCoffee searches for a global alignment by maximizing a target function using simulated annealing on a set of weighted bipartite graphs that are constructed using a triplet approach similar to T-Coffee. To assess its performance, NetCoffee was applied to four real datasets. Our results suggest that NetCoffee remedies several limitations of previous algorithms, outperforms all existing alignment tools in terms of speed and nevertheless identifies biologically meaningful alignments.