An Introduction to Metabolic Networks and Their Structural Analysis
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Dividing protein interaction networks for modular network comparative analysis
Pattern Recognition Letters
Transactions on large-scale data- and knowledge-centered systems III
Survey: Computational challenges in systems biology
Computer Science Review
Algorithms for subnetwork mining in heterogeneous networks
SEA'12 Proceedings of the 11th international conference on Experimental Algorithms
Hi-index | 3.84 |
Motivation: An important tool for analyzing biological networks is the ability to perform homology searches, i.e. given a pattern network one would like to be able to search for occurrences of similar (sub)networks within a set of host networks. In the context of metabolic pathways, Pinter et al. [Bioinformatics, 2005] proposed to solve this computationally hard problem by restricting it to the case where both the pattern and host networks are trees. This restriction, however, severely limits the applicability of their algorithm. Results: We propose a very fast and simple algorithm for the alignment of metabolic pathways that does not restrict the topology of the host or pattern network in any way; instead, our algorithm exploits a natural property of metabolic networks that we call ‘local diversity property’. Experiments on a test bed of metabolic pathways from the BioCyc database indicate that our algorithm is much faster than the restricted algorithm of Pinter et al.—the metabolic pathways of two organisms can be aligned in mere seconds—and yet has a wider range of applicability and yields new biological insights. Our ideas can likely be extended to work for the alignment of various types of biological networks other than metabolic pathways. Availability: Our algorithm has been implemented in C++ as a user-friendly metabolic pathway alignment tool called METAPAT. The tool runs under Linux or Windows and can be downloaded at http://theinf1.informatik.uni-jena.de/metapat/; Contact: florian.rasche@uni-jena.de Supplementary information: Supplementary data are available at bioinformatics online.