Combinatorial Optimization Algorithms for Metabolic Networks Alignments and Their Applications

Abstract

The accumulation of high-throughput genomic and proteomic data allows for reconstruction of large and complex metabolic networks. To analyze accumulated data and reconstructed networks, it is critical to identify network patterns and evolutionary relations between metabolic networks; finding similar networks is computationally challenging. Based on gene duplication and function sharing in biological networks, a network alignment problem is formulated that asks the optimal vertex-to-vertex mapping allowing path contraction, different types of vertex deletion, and vertex insertions. This paper presents fixed parameter tractable combinatorial optimization algorithms, which take into account the similarity of both the enzymes’ functions arbitrary network topologies. Results are evaluated by the randomized P-Value computation. The authors perform pairwise alignments of all pathways for four organisms and find a set of statistically significant pathway similarities. The network alignment is used to identify pathway holes that are the result of inconsistencies and missing enzymes. The authors propose a framework of filling pathway holes by including database searches for missing enzymes and proteins with the matching prosites and further finding potential candidates with high sequence similarity.