基于复合物结构和网络拓扑特性的关键蛋白质识别算法  

Identification of essential proteins based on protein complex and topology properties

作　　者：卢鹏丽[1] 蔚京娟 LU Peng-li;YU Jing-juan(School of Computer and Communication,Lanzhou Univ.of Tech.,Lanzhou 730050,China)

机构地区：[1]兰州理工大学计算机与通信学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2022年第2期97-106,共10页Journal of Lanzhou University of Technology

基　　金：国家自然科学基金(11861045,11361033)。

摘　　要：关键蛋白质的识别有助于从分子水平上理解生命的活动过程,然而仅从拓扑特性角度来识别的关键蛋白质不够精准,因此为了提高识别准确率,结合复合物信息提出了确定蛋白质关键性的指标模型EIC,该模型是基于蛋白质复合物内的局部中心性特性以及网络的全局信息特性来考虑.使用DIP和MIPS两种蛋白质相互作用(PPI)网络作为实验数据集,选用“排序-筛选”方法、精准召回、折刀等评估方法,通过与现有方法DC、BC、EC、SC、LAC、NC以及UC作对比,实验结果表明EIC方法优于这些方法,能够提高预测关键蛋白质的准确率.The identification of essential proteins is helpful to understand the process of life at the molecular level,thus the identification of essential proteins has gradually become an important research work.However,the identification of essential proteins that only focused on the topological characteristics is not accurate enough.To improve the recognition accuracy,a new method EIC was proposed,which integrated the local central information and global topological features of protein complex for identifying essential proteins.Different Protein-Protein Interaction(PPI)networks of DIP and MIPS were used as materials.Some experiments of different assessment method were used including precise-recall curves and jackknife methodology.By comparing the proposed method with the existing methods such as DC,BC,EC,SC,LAC,NC and the recent method UC,the ability of recognizing essential proteins by the proposed method EIC is more effective,in that it may enhance the accuracy for identifying essential proteins.

关 键 词：蛋白质相互作用网络 关键蛋白质 蛋白质复合物 评估方法 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]