Development of computer software for predicting associations among oral bacteria

Introduction: It is now accepted that more than 700 bacterial species may inhabit the oral cavity, and that ecological interactions among oral bacteria such as nutritional interdependences and co-aggregation may frequently occur. Taking this into consideration, particular sets of bacteria may interact and play synergistic roles in oral health and disease. Objectives: This study aimed at developing and applying a computer software to predict positive associations among oral bacteria, based on presence/absence data derived from a DNA-based technique, using the user-friendly Ruby programming language as building block. Methods: Terminal Restriction Fragment Length Polymorphism (T-RFLP) was applied to determine the microbial composition of 34 cases of endodontic infections. The command-line software FRAGMATE was designed with the Ruby programming language to investigate positive associations among all possible pairs of bacteria, as determined by the odds-ratio statistics (OR≥2.0), considering presence/absence as the test parameter. Taxonomical prediction of species was inferred by TAP-T-RFLP software, based on a ribosomal DNA database. Finally, a clustering algorithm was applied on every pair of species to investigate the existence of higher-order structures of associated bacteria. Results: Overall, 23 positively associated bacteria species were identified. Of these, special attention was drawn to Veillonella sp. oral clone ASCB03 and to Fusobacterium nucleatum, both displaying markedly higher prevalence of associations with other species. A strinkingly high positive association (OR=41.8) was found between F. nucleatum and Campylobacter sputorum. In general, interactions were preponderantly orchestrated by sets of two to three taxonomic units, some of which composing core groups capable of participating in more than one higher-order consortium. In all, 4 threesome and 1 foursome higher-order consortia were detected. Conclusion: Our results reiterate the hypothesis that oral bacteria may act sinergistically in the oral cavity. Further laboratory studies should be performed to validate the results obtained by our in silico analyses.