基于关键功能模块挖掘的蛋白质功能预测

精确注释蛋白质功能是从分子水平理解生物体的关键.由于内在的困难和昂贵的开销，实验方法注释蛋白质功能已经很难满足日益增长的序列数据.为此，提出了许多基于蛋白质相互作用（Protein-protein interaction，PPI）网络的计算方法预测蛋白质功能.当今蛋白质功能预测的趋势是融合蛋白质相互作用网络和异构生物数据.本文提出一种基于多关系网络中关键功能模块挖掘的蛋白质功能预测算法.关键功能模块由一组紧密联系且共享生物功能的蛋白质组成，它们能与网络中的剩余部分较好地区分开来.算法通过从多关系网络的每一个简单网络中挖掘高内聚、低耦合的子图形成关键功能模块.关键功能模块中邻居蛋白质的功能用于注释待预测功能的蛋白质.每一个简单网络在蛋白质功能预测中的重要性各不相同.实验结果表明，提出的方法性能优于现有的蛋白质功能预测方法.

Prediction of Protein Functions Based on Essential Functional Modules Mining

The accurate annotation of protein functions is a key to understanding living organisms at the molecular level. With its inherent difficulty and expense, experimental characterization of protein functions cannot scale up to accommodate the vast amount of sequence data. As a result, many computational methods based on protein-protein interaction (PPI) networks have been proposed to predict the functions of proteins. Nowadays, the trend in protein functions prediction is to integrate PPI networks and heterogeneous biological data. A novel protein functions prediction algorithm was proposed based on mining essential functional modules from a multi-relational network. An essential functional module is a group of densely connected proteins with shared biological function and can be well-separated from the rest of the network. The proposed algorithm identified subgraph with high cohesion and low coupling on each single network derived from the multi-relational network to form essential functional modules. Functions of neighbor proteins within essential functional modules were used to annotate the testing protein. Each single network has different importance on the prediction of protein functions. Experiment results show that our method outperforms other protein functions prediction methods.