A Network Alignment Algorithm with Structure Features and Representation Learning

• 1、School of Artificial Intelligence and Computer Science, Jiangnan University, Wuxi 214122
• 2、Jiangsu Provincial Engineering Laboratory of Pattern Recognition and Computing Intelligence, Jiangnan University, Wuxi Jiangsu 214122, China

Abstract：Protein-protein interaction (PPI) networks contain a large amount of unknown biological information, and the analysis of PPI network pairs by network alignment can discover unknown protein functions among different species for cross-species knowledge transfer. In recent years, numerous network alignment algorithms have been proposed, but the existing network alignment algorithms are difficult to achieve a good compromise between biological conservativeness and topological conservativeness. Therefore, in this paper, a new pairwise PPI network alignment algorithm EmbAlign is proposed.To address the limitations of existing algorithms in computing topological similarity, EmbAlign extracts the structural features of nodes in the network hierarchically and uses network embedding to embed and represent the nodes in a low-dimensional vector space. Meanwhile, to retain biological information, EmbAlign adds sequence similarity in the matching process. In order to verify the accuracy of EmbAlign algorithm, this paper conducts experiments on synthetic and real networks and compares them with existing algorithms respectively. The experimental results show that EmbAlign can generate results that outperform existing algorithms in both topological and biological metrics.

Keywords： network alignment network embedding node representation PPI network complex network