基于监督学习的蛋白质络合物抽取方法

Method of Predicting Protein Complex Based on Supervised Learning

蛋白质关系网络中存在着大量的蛋白质络合物,络合物对有利于深入探索生物细胞的组织原理和功能有着重要意义。然而传统的络合物发现算法多基于网络的拓扑结构,没有融合络合物本身的结构信息。针对这个问题,提出了监督学习的络合物发现方法,将多种能够标示络合物的信息作为特征,使用监督学习方法对样本集进行训练,将训练得到的模型应用在络合物发现算法中。实验证明,该方法能有效地从蛋白质关系网络中发现络合物。

Protein complexes are important for understanding principles of cellular organization and function. Predicting protein complexes from protein-protein interaction （PPI） networks is of great significance. Previous methods for complex prediction are usually based on topological structure without considering the structure of complexes. In this paper,a supervised learning method is used to solve this problem. The features are constructed by multiple information of complex and the model obtained by the supervised method is used in the algorithm of complexes detection. The experimental results show that the method is an effective approach to predict protein complex from protein interaction network.