Employing Evolutionary History and Genetic Context to Better Infer Gene Function in Metagenomic and Whole-Genome Sequencing Data

McFarland, Alexander George

doi:https://doi.org/10.21985/n2-jxhd-0m85

Work

Employing Evolutionary History and Genetic Context to Better Infer Gene Function in Metagenomic and Whole-Genome Sequencing Data

Public

Download PDF

Download All Files (.zip)

To combat the rise of antibiotic-resistant pathogens in clinical settings, it is necessary to understand which environments and conditions select for antibiotic resistance. Analysis of environments hypothesized to select for antibiotic resistance has been revolutionized by metagenomic sequencing. The metagenomic pool of DNA sequences can be probed in silico using bioinformatic methods to identify antibiotic resistance genes. These methods attribute the function of antibiotic resistance to predicted genes based on their sequence homology to functionally characterized genes in a database. A challenge to these approaches is that sequence homology may not always signify function. Additionally, some true antibiotic resistance genes may be missed because functionally confirmed representatives in a database do not have sufficient homology. Finally, detection of all the components of multigene resistance systems, like multidrug efflux pumps, is rarely performed. Developing and employing accurate approaches to predict antibiotic resistance, and validating them with experimental methods, is crucial to fully leveraging the power of shotgun metagenomic sequencing to better assess the resistance potential of an environment.Antibiotic resistance may inadvertently emerge through exposure to antimicrobials, which are found in everyday products and may select for antibiotic resistance genes and systems. In this dissertation, I examined the microbiomes in environments hypothesized to select for antibiotic-resistant bacteria due to exposure to antimicrobials. I find that all environments contain antibiotic resistance genes, including multidrug efflux pumps. This is of particular concern as multidrug efflux pumps can confer resistance to multiple antibiotics. I examined the multidrug efflux pump content of Pseudomonas recovered from indoor dust and hospital sink biofilms and present a model for assessing the potential for antimicrobial exposure in selecting for antibiotic resistance. Because multidrug efflux pumps and other multigene resistance gene clusters feature prominently in metagenomic and whole-genome sequencing data, I developed a new bioinformatic tool, GeneGrouper. GeneGrouper searches for and bins gene clusters of interest and presents a population-level overview of their distribution and the conservation of their gene content, quantified as dissimilarity. To further understand the distribution of antibiotic resistance genes in other environments, I also helped develop a novel functional metagenomics method, Mosaic-Ends Tagmentation (METa). METa uses low DNA input to generate large functional metagenomic libraries that can be screened against antibiotics of interest. Taken together, the results in this dissertation help to better understand and characterize the antibiotic resistance gene content of environments and factors that may select for them.

Creator