摘要
提出了一种多物种代谢网络同源功能模块的挖掘算法.依据物种间的亲缘关系,给出一种新的化合物相似度定义方式,使用AP算法进行模块的初始划分,然后按照模块的同源系数逐层进行重叠扩展.使用65种不同生物的代谢网络进行实验研究,结果表明:得到的保守功能模块与KEGG数据库提供的参考功能模块具有较高匹配率,外围功能模块体现了功能模块在不同物种内的分布差异,验证了算法的有效性.
An algorithm for extracting orthologous functional modules in metabolic networks of multiple species was developed. A similarity measure between the compounds integrating the phylogeny was firstly constructed. Based on the similarity, the AP （affinity propagation） algorithm was adopted with an immediate purpose to obtain a hard partition of the multiple species networks. A soft partition was finally obtained with an overlapping extension concerning the orthologous coefficient of modules. To verify the efficiency of our algorithm, the available metabolic networks were used from 65 diverse species and the experimental results demonstrate that the conserved functional modules match well with those proposed in the KEGG （Kyoto Encyclopedia of Genes and Genomes） database. Furthermore, the periphery modules show that the distribution of functional modules is different from various species.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2012年第1期35-39,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60933009)
高等学校博士学科点专项科研基金资助项目(200807010013)
中央高校基本科研业务费专项基金资助项目(50510030006)
关键词
代谢网络
数据挖掘算法
同源功能模块
重叠模块
相似度
metabolic network
data mining algorithm
orthologous functional module
overlap mod-ule
similarity
