Applying Logic Programming to Derive Novel Functional Information of Genomes

Abstract

This paper describes an application of the logic programming paradigm to large-scale comparison of complete microbial genomes each containing four-million amino acid characters and approximately two thousand genes. We present algorithms and a Sicstus Prolog based implementation to model genome comparisons as bipartite graph matching to identify orthologs — genes across different genomes with the same function — and groups of orthologous genes — orthologous genes in close proximity, and gene duplications. The application is modular, and integrates logic programming with Unix-based programming and a Prolog based text-processing library developed during this project. The scheme has been successfully applied to compare eukaryotes such as yeast. The data generated by the software is being used by microbiologists and computational biologists to understand the regulation mechanisms and the metabolic pathways in microbial genomes.