电子学报 ›› 2021, Vol. 49 ›› Issue (8): 1489-1497.DOI: 10.12263/DZXB.20200357

• 学术论文 • 上一篇    下一篇

基于动态图的PPI网络构建和复合物挖掘算法研究

李鹏1,2,3, 闵慧4, 罗爱静1,3   

  1. 1.中南大学湘雅三医院,湖南 长沙 410013
    2.湖南中医药大学信息科学与工程学院,湖南 长沙 410208
    3.医学信息研究湖南省普通高等学校重点实验室(中南大学),湖南 长沙 410006
    4.湖南信息职业技术学院软件学院,湖南 长沙 410200
  • 收稿日期:2020-04-13 修回日期:2020-08-27 出版日期:2021-08-25
    • 作者简介:
    • 李 鹏 男,1983年11月出生,湖南泸溪人.博士、讲师,中南大学公共卫生与预防医学博士后流动站在站博士后.主要研究方向为生物信息学、机器学习、中医药大数据. E-mail:lpchs617@csu.edu.cn
      闵 慧 女,1986年12月出生,湖南湘潭人.硕士、讲师,主要研究方向为生物信息学、网络优化. E-mail:mh1220@126.com
      罗爱静(通信作者) 女,1962年出生,湖南安乡人.博士、教授、博士生导师,主要研究方向为医药信息管理、卫生信息管理、医药信息检索. E-mail:805372510@qq.com
    • 基金资助:
    • 国家社会科学基金重点项目 (17AZD037); 国家重点研发计划 (2017YFC1703306); 湖南省卫生健康委科研项目 (202112072217); 湖南自然科学基金青年项目 (2019JJ50453); 湖南自然科学基金面上项目 (2018JJ2301); 湖南省科技厅重点项目 (2018JJ2301); 湖南省教育厅一般项目 (19C1318)

Research on PPI Network Construction and Complex Mining Algorithm Based on Dynamic Graph

LI Peng1,2,3, MIN Hui4, LUO Ai-jing1,3   

  1. 1.The Third Xiangya Hospital of Central South University, Changsha, Hunan 410013, China
    2.School of Informatics, Hunan University of Chinese Medicine, Changsha, Hunan 410208, China
    3.Key Laboratory of Medical Information Research (CSU), College of Hunan Province, Changsha, Hunan 410006, China
    4.Software Department, Hunan College of Information, Changsha, Hunan 410200, China
  • Received:2020-04-13 Revised:2020-08-27 Online:2021-08-25 Published:2021-08-25
    • Supported by:
    • Key Program of The National Social Science Fund of China (17AZD037); National Key Research and Development Program of China (2017YFC1703306); Research Program of Health Commission of Hunan Province (202112072217); Youth Project of Natural Science Foundation of Hunan Province (2019JJ50453); Natural Science Foundation Project of Hunan Province (2018JJ2301); Key Project of Hunan Science and Technology Department (2018JJ2301); General Program of Hunan Educational Committee (19C1318)

摘要:

动态蛋白质网络的构建和复合物挖掘问题是目前研究的热点.针对现有的算法在解决前述问题上的不足,文中考虑了蛋白质的活性周期和连接强度,首先提出了一种基于动态图的蛋白质网络构建算法.然后基于密度聚类设计了一种在动态蛋白质网络上挖掘复合物的算法(PCMA).整个挖掘过程包含三个步骤:基于DBSCAN(Density-Based Spatial Clustering of Applications with Noise)算法的蛋白质复合物生成;基于合并增益的蛋白质复合物合并和基于归属度的复合物调整.在多个公开的生物数据集上进行了实验,实验结果表明,所提算法在查全率、查准率和F-measure方面的性能都要优于现有的算法,且对输入参数不敏感.在保证蛋白质复合物挖掘准确性的前提下,算法的时间复杂度处于一个合理的范围之内.

关键词: 动态蛋白质网络, 蛋白质复合物, 动态图, 密度聚类, 查全率, 查准率, 时间复杂度

Abstract:

Dynamic protein network construction and complex mining problem is a hot topic. In view of the shortcomings of existing algorithms in solving the above problems, a protein network construction algorithm based on dynamic graph is firstly proposed by considering the active period and the connection strength of proteins in this paper. Then, a protein complex mining algorithm(PCMA) on dynamic protein network is designed based on the density clustering. The whole mining process consists of three steps: the generation of protein complex based on DBSCAN(density-based spatial clustering of applications with noise) algorithm; the combination of protein complex based on the combination gain and the adjustment of protein complex based on the degree of membership. Experiments are carried out on several open biological datasets. The experimental results show that the performance of the proposed algorithm is better than that of the existing algorithms in terms of recall, precision and F-measure, and it is not sensitive to the input parameters. On the premise of ensuring the accuracy of protein complex mining, the time complexity of the proposed algorithm is in a reasonable range.

Key words: dynamic protein network, protein complex, dynamic graph, density clustering, recall, precision, time complexity

中图分类号: