Project Summary:
Despite reductions in antimicrobial use and implementation of antibiotic stewardship programs, antimicrobial
resistant (AMR) pathogens continue to emerge and persist, causing millions of deaths every year. The long-term
goal of this proposal is to decrease clinical AMR incidence by advancing knowledge of how AMR develops,
persists and transmits to pathogens within the microbiome. The overall objective of this project is to develop a
novel high-throughput method for generating enriched long-read sequencing data that accurately reflects the in
situ dynamics of AMR genes across entire microbial communities. This is a major gap in methodology that
hampers knowledge discovery and leads to suboptimal patient outcomes. The rationale is that this method will
enable fundamental advances in both applied and basic research through: improved prediction about when and
how bacteria exchange AMR genes; and rapid, lower-cost, higher-throughput generation of metagenomic
resistome data. The project objective is achieved through three specific aims: 1) Increase the sensitivity and
specificity of probe-based pre-sequencing enrichment; 2) Generate contextualized metagenomic sequence data
with real-time enrichment of antimicrobial resistance genes; and 3) Evaluate the performance of triple-enrichment
using samples with diverse antimicrobial resistance gene composition. Under aim 1, probes for all potential off-
target (i.e., AMR) DNA will be designed using a novel compressed colored de Bruijn graph. Probes will be
manufactured and used to deplete off-target DNA from metagenomic samples prior to enrichment of AMR genes.
For aim 2, adaptive Nanopore sequencing will perform real-time rejection of off-target DNA using a compressed
pangenomic index of all off-target regions, followed by on-the-fly classification of antimicrobial resistance genes.
For aim 3, the potential bias of probe- and sequence-based enrichment will be quantified on a gene-by-gene
basis to fully evaluate the performance of the enrichment workflow for various sample types and use cases. The
proposed research is innovative for two major reasons: first, it enables the applied use of microbial ecology for
clinical and public health applications, which is a major gap in the field of metagenomics; second, it allows for
deep sequencing and robust genomic co-localization of rare and low-abundance metagenomic targets, which is
currently very difficult to obtain. The proposed methodology will open new horizons for basic research into how
and why bacterial communities evolve; and under what conditions diverse bacteria share critical functions such
as AMR. The proposed research is significant because it will have broad applicability to many critical activities
for combating AMR, including patient-centric decisions about antibiotic use and public health activities such as
AMR surveillance. The proposed advancements will help predict AMR emergence events in patients and
populations; identify hot-spots of AMR development and dissemination in hospitals and the community; and help
track critically important AMR genes that may be rare or low-abundance within a given environment or set of
samples.