The main goal of my research group is to understand how mechanisms of inheritance affect genetic variation, and conversely, how genetic variation affects mechanisms of inheritance (i.e., “the genetics of genetics”). Our primary model system is the human bacterial pathogen Haemophilus influenzae, an important agent of ear infections (otitis media) in children, as well as lung infections associated with chronic respiratory conditions. H. influenzae, like many other pathogens, is naturally competent, able to actively transport environmental DNA through its cell membranes and incorporate homologous molecules into its chromosomes. This pathway, called “natural transformation,” is a major mechanism of gene transfer across bacteria and has a profound effect on genome evolution, including spreading antibiotic resistances and other virulence determinants. Our current research seeks to answer three major questions using a combination of microbiology, molecular genetics and genomics/bioinformatics approaches:

We have generated high-resolution genome-wide maps of transformation, finding massive variation in rates at different chromosomal loci. Producing these maps required extremely deep DNA sequencing (>10,000-fold genomic coverage) and novel analytical tools to distinguish true events from sequencing errors. Unexpectedly, the two known determinants, local sequence identity and the proximity to known “uptake signal sequences,” explain only a small proportion of the variation (~15%). Other unknown factors must contribute. We are now using advanced optical mapping technology to reproduce these maps in the absence of genetic variation, in order to disentangle the role of chromosome structure from the potential for genetic incompatibilities (“speciation genes”) skewing our results.

This work has also uncovered a small number of extreme hotspots (>10% transformation frequency), which reside in genes undergoing strong diversifying selection that encode large membrane proteins, likely as an immune invasion tactic. Ongoing work is dissecting the underlying molecular mechanism of these hotspots.

We have developed a novel method for mapping genes in bacteria, exploiting natural transformation in combination with genome-wide deep sequencing. The approach exploits natural competence to generate complex pools of recombinants between a donor carrying a pathogenesis trait of interest and an avirulent recipient. Selection for recombinants that acquired the trait, followed by genome-wide profiling of donor-specific allele frequencies, we can rapidly identify the relevant genes. We have used this method to map an operon involved in intracellular invasion of airway epithelial cells, a trait with implications for chronic infection, bacterial persistence, and trafficking of cells to different body sites. Newer work is using the same approach to map the genes responsible for natural variation in other pathogenesis traits, including resistance to human complement-mediated killing, as well as investigating the possibility of identifying genes involved in in vivo pathogenesis in an animal model of otitis media.

In several ongoing collaborations, we are investigating how the genomes of bacteria isolated from patients with chronic infections change over time. The long-term goal of this research is to apply statistical genomic approaches developed by human geneticists to the identification of bacterial virulence factors that contribute to disease in natural populations. Most (but not all) of our current datasets are clinical isolates of H. influenzae, including mutators from pediatric cystic fibrosis, carriage isolates from healthy children, serially collected isolates from adult patients with chronic obstructive pulmonary disease, as well as isolates collected from the middle ear of children with otitis media upon insertion of tympanostomy tubes. We are applying a variety of genomic methods to identifying putative virulence genes, including machine learning and phylogenetic correlated evolution methods. Other organisms of interest include Gardnerella vaginalis, Burkholderia cenocepacia, and non-tuberculosis mycobacteria.

In addition to the projects described above, our group is actively involved in several collaborations as a member of the Center for Genomic Sciences and the Center for Advanced Microbial Processing. These include bioinformatic analysis of HIV and mitochondrial genomes in HIV/AIDS patients, transcriptome analyses in several mammalian and bacterial systems, and novel approaches to microbiome data analysis.