基于现有的蛋白质相互作用数据,提出利用邻居曲线方法来分析癌症基因产物在蛋白质相互作用网络中的中心度和聚集度,据此获取与癌症高度相关的候选致病基因。癌症基因大规模测试显示,有26%的目标基因在候选基因中排名前5%,90%的目标基因在候选基因中排名前50%,该方法能有效地识别癌症致病基因。
By analyzing the topological features in PPIs(protein-protein interactions) network based on nearest neighbor search,this paper evaluated the possibilities of candidate genes to be cancer-causative genes.A large scale test shows that 26% and 90% target genes are at the top 5% and top 50% of the list prioritized,the approach can identify cancer genes with high performance