An algorithm for graph isomorphism and its application to KEGG Compound Search

Figure 1: KEGG Compound Search One of the significant challenges in bioinformatics is to represent and manipulate computationally complex biological information such as metabolic pathways, chemical compounds and protein-protein interactions. This challenge was successfully accomplished by representing biological information with graph objects[1]. On the other hand, these enormous graph structured databases (target graphs) require efficient algorithms for searching for a given subgraph (query graph). These kinds of search problems may be defined using the concept of subgraph isomorphism. Recently, subgraph isomorphism has come to be applied to biological analysis[2]. In this paper, we present a novel algorithm for solving the subgraph isormorphism problem by extending Ullmann’s algorithm[3], which is currently recognized as the most efficient algorithm[4]. We implemented our algorithm as the SUBCOMP program which is available on the KEGG Compound Search page (Fig. 1).