Modeling Mycobacterium tuberculosis H37Rv in silico

Abstract

Network reconstructions and constraint-based modeling have been shown to be effective methods for understanding complex processes, such as metabolism. These reconstructions are in fact biologically structured knowledge-bases that can be queried through computations, and thus have become valuable tools for Systems Biology. Strengths of this approach include flexibility in incorporating incomplete data measurements, the ability to incorporate different types of data (high-throughput as well as physiological), simultaneously, as well as the ability to make predictions with minimal reliance on parameter and curve fitting. Thus, this approach aims to move away from fitting data to describe experimental results using the current understanding of metabolism in order to interpret the data, make predictions, and to identify the gaps and bridges in knowledge. The critical components for creating genome-scale reconstructions of metabolism include a sequenced and annotated genome, reaction stoichiometry for the annotated enzymes, and a bibliome for the organism (combined primary and secondary literature sources). Network reconstructions of the devastating pathogen Mycobacterium tuberculosis have been developed and have enabled the ability to query functional capabilities using constraint-based modeling approaches. Since these networks are then structured in terms of gene-protein-reaction associations, these knowledge-bases can serve as biologically structured databases onto which various high-throughput data types can be directly mapped on. This chapter will focus on the model reconstruction process, methods that have been employed for analysis, and predictive applications of modeling the pathogen H37Rv strain of tuberculosis. Employing the existing analysis methods and available datasets there have already been a large number of applications for modeling constraint-based modeling of H37Rv. The reconstruction process is a time and resource intensive procedure and to date high quality reconstructions have not been possible without manual curation. The benefit of having a detailed and quality-controlled reconstruction procedure is to help determine a high quality model that will provide more meaningful predictions from simulations. Applications of M. tuberculosis models have included the prediction of growth rates, assessment of different growth media, prediction of gene knockouts, identification of new drug targets, identification of alternative drug targets for existing drugs, and modeling the interaction macrophages during different infectious states. Historically, technological advancements have driven biological discovery and have thus been a limiting factor in the development of methods to modify and alter biology, e.g., antibiotics. However, in the past decade with various high-throughput technologies (e.g., transcriptomics, proteomics, metabolomics, etc.) are being employed more frequently, thus there is a growing burden and need for means to integrate, interpret, and ideally make predictions for these datasets. Given the successes to date, with further development of new methods in conjunction with deeper experimental probing of tuberculosis in vitro and in vivo, constraint-based modeling will likely become even more important in the finding new targets and treatments for tuberculosis.