Parallel Identifying Motifs in Biological Network on Multicore System
The required time to indentify the motifs in biological networks will be sharply increased with increase of the scale of the original network or the size of motifs in the networks.It is very necessary to develop the fast algorithms to indentify larger motifs in large-scale biological networks.A thread-level parallel indentifying motifs algorithm is proposed on multicore machine, in which the local hash tables are used to store the subgraph SIKs mined by parallel threads and a task queue is used to dlistribute the data blocks to parallel threads to balance their loads.A dynamic task scheduling strategy is used to balance the loads among nodes on multi-core cluster, and an efficient parallel algorithn is designed to identify the motifs in biological networks.The experiment results in undirected network YeastPPI and directed network yeast show that compared with the existing parallel indentifying motifs algorithm, the presented parallel algorithms are faster and obtain good speedup, and it can identify larger motifs in the biological networks.
biological network motif identification multicore architecture parallel algorithms subgraph enumeration subgraph isomorphism task scheduling
Cheng ZHONG Jing ZHAO Feng YANG Mengxiao YIN
School of Computer and Electronics and Information,Guangxi University,Nanning,China,Guangdong Key Laboratory of Popular High Performance Computers,Shenzhen Key Laboratory of Service Computing and Applications,Shenzhen,China
国内会议
金华
英文
1-10
2015-10-30（万方平台首次上网日期，不代表论文的发表时间）