The goal of this proposal is to develop a universal platform for a high throughput, phenotype-based identification of bacterial pathogens in complex environments like soil or freshwater. Our approach, which integrates innovations in environmental microbiology, droplet microfluidics, and machine learning/computational biology, represents a revolutionary advance in the science, devices, and systems needed for the isolation and characterization of known as well as potentially unknown or unculturable bacterial pathogens. The long-term impact of these activities will be the delivery of a technology platform that detects bacterial pathogens as, or even before, they emerge, and thereby reduces the attendant threats to warfighters and civilians.
Task 1. To develop novel techniques for extracting, isolating, and maintaining the viability of potentially unknown and/or unculturable bacteria from complex environments. This effort featured the encapsulation of extracted bacterial cells in gel microdroplets for massively parallel single-cell resolution microbial cultivation under defined nutrient conditions that mimic the natural environment.
Task 2. To develop a high-throughput, droplet microfluidics system in which the phenotypic outcome of pathogenic interactions between single isolated bacteria and human host cells will be non-destructively measured. Bacterial phenotypes interrogated included host cell toxicity, intracellular replication, host cell adherence, antibiotic resistance, resistance to reactive oxygen species, and immune activation and evasion.
Task 3. To develop a pipeline that will perform single-cell genome sequencing of bacteria that display pathogenic phenotypes, and use this information to develop machine-learning algorithms and bioinformatics/database systems that automatically link the various phenotypes and genotypes to correctly determine the pathogenic nature of target bacterial cells.
A microfluidics device was developed that is now being used by CDC and FDA. The UVA team helped tune the device by analyzing the bacteria that generated responses in that device.
We developed pipelines for single-cell genomic analysis that were used to associate a genomic signature with a bacterial phenotype. We also developed pipelines for single-cell RNA-Seq analysis to identify human immune genes that respond to bacterial interactions.