Improved Homology-driven Computational Validation of Protein-Protein Interactions Based on Evolutionary Gene Duplication and Divergence Hypothesis

Outline

C. Frech, M. Kommenda, V. Dorfer, T. Kern, H. Hintner, J. W. Bauer, K. Oender - Improved Homology-driven Computational Validation of Protein-Protein Interactions Based on Evolutionary Gene Duplication and Divergence Hypothesis - BMC Bioinformatics, Vol. 10, No. 1, 2009, pp. 18

Abstract

BACKGROUND: Protein-protein interaction (PPI) data sets generated by high-throughput experiments are contaminated by large numbers of erroneous PPIs. Therefore, computational methods for PPI validation are necessary to improve the quality of such data sets. Against the background of the theory that most extant PPIs arose as a consequence of gene duplication and divergence, we investigated if traditional homology-based concepts for PPI validation can be further improved by a more comprehensive search for homologs. RESULTS: To validate a putative PPI we combine FASTA and PSI-BLAST to perform a sequence-based search for pairs of interacting homologous proteins within a large, integrated PPI database. A normalized scoring scheme that incorporates both the quality and quantity of all observed matches allows us (1) to consider also tentative paralogs and orthologs in the analysis and (2) to easily combine the search results obtained by FASTA and PSI-BLAST. ROC curves illustrate the high efficacy of this approach. CONCLUSIONS: The duplication-divergence model of PPI evolution suggests that for true PPIs many homologous PPIs exist, not only among close relatives but also among remote homologs. We demonstrated that a validation technique that consequently exploits this idea is very efficient and improves over traditional homology-based concepts. In particular, our insights might be useful in cases where traditional homology-based techniques are not an option owing to a lack of assured paralogs or orthologs.