基于拓扑和生物特征的权重网络中络合物抽取
Complex Extraction from the Weighted Network Based on Topological and Biological Characteristics

蛋白质关系网络的权重设置对蛋白质络合物的抽取有着较大的影响.综合考虑蛋白质关系网络的拓扑结构特征和生物信息特征,提出一种新的权重设置策略,并向原有蛋白质关系网络中添加高可信度关系.基于修正后的蛋白质关系网络,用监督学习方法抽取蛋白质络合物,在DIP数据集上,F值达到0.5705.
The weight setting of protein interaction network has a great effect on protein complex identification.A better strategy which considers both topological characteristics and biological characteristics of protein interaction network has been provided.Furthermore,some credible interactions have been added into the original network.Then,the updated protein interaction network has been used for complex identification based on a supervised method.F-score reached 0.570 5 in the DIP dataset