see also: Publications page

The big picture:

The human microbiome harbors a large capacity for within-person adaptive mutations. Bacterial strains stably colonize a person’s microbiome for decades, and during this time billions of bacterial mutations are generated daily. Rare adaptive mutations among these billions might impact the immune system, the metabolism of particular nutrients or drugs, or resilience of the microbiome. Despite this potential, little is known about the extent of evolution within and across human microbiomes.

Despite the potential of rational probiotics for treating disease and promoting wellness, our ability to employ them remains limited. (Biomedical impact)

Fecal transplants are famously effective in treating people with recurrent Clostridium dificile infections. Recurrent urinary tract infections have been treated with asymptomatic bacteria that outcompete pathogens. Even for diseases that are not caused by bacteria, microbiome-based therapies may one day improve drug metabolism, supply vital nutrients, and modulate the immune system (including for cancer therapy). Probiotics for the skin may one day improve wound healing, prevent the formation of acne, or outcompete Staphylococcus aureus to manage eczema.

However, we still remain limited in our ability to precisely manipulate microbiomes. We cannot reliably predict which strains, once introduced, will stably colonize an individual. This unpredictability stems from a lack of mechanistic understanding of how single strains and species colonize and evolve in human ecosystems. What conditions facilitate transfer of bacteria between people? Do probiotics need to be personalized to colonize stably? Does bacterial warfare influence the colonization process?

Understanding the interplay of rapid evolution and ecology in complex microbial environments. (Basic science impact)

Experimental evolution and genomic studies of evolution during infections have revealed that bacterial adaptation can be rapid. In addition, they revealed common patterns, including repeated recurrence of adaptive mutations (parallelism), the importance of genomic context for predicting whether or not a mutation will confer an advantage (epistasis), and the ability of a single genotype to rapidly diversity into coexisting strains with different life strategies (niche differentiation).

We seek to understand the extent to which these themes persist in more complex environments, and the ways in which ecological complexity impacts evolution. For example, opportunities for differentiation may be limited in microbiomes due to the occupation of alternative niches by other species.

Conversely, we seek to understand the impact of evolution on ecological interactions (between species). Traditionally, evolution is considered a slow process, with ecological models treating species as fixed entities. However, work from our lab and others have shown the potential for adaptive evolution within individual human microbiomes. What is the impact of adaptive mutations on other species, and might this alter emergent community properties?

Tools we are using:

Culture-based ‘single-organism’ evolutionary-inference to gain insights into community assembly in vivo. Bacteria replicate their DNA with high fidelity as they grow from a single cell into a colony, allowing rigorous identification of single mutations across the whole genome (rare mutations occurring in culture can usually be identified and removed from analysis). Whole genome information, with linkage, enables retrospective inference of on-person and across-person dynamics.

in vivo measurements on the microscale, for understanding the physiological determinants of ecology and evolution in the microbiome. We culture bacteria and phage from individual pores using commercially available tools (see image below). We are also developing new protocols for imaging microbial communities in biological tissues.

Genomic and metagenomic tool development and leveraging under-analyzed public data to mechanistic gain insights into microbiome assembly and function.

High-throughput measurements for studying strain-level variation and drivers of bacterial fitness. We leverage robotics and clever experimental design to assess microbial growth across a wide range of conditions.

in silico biophysical modeling of community assembly, incorporating salient features of abiotic and biotic environments experienced by microbes.

Experimental ecology and evolution to test targeted hypotheses.

Advantages of studying human skin microbiomes:

We study a variety of human-associated and environmental microbiomes, including the gut, colorectal cancer and vaginal microbiomes. Our largest focus is currently on the skin microbiome.

Stable and person specific. While human skin is indeed exposed to the environment, each person retains specific strains for years. We seek to understand why some strains are maintained and what allows them to exclude others.

Ease of spatiotemporal sampling. The skin surface can be repeatedly sampled from human subjects of all ages. From adult subjects, we use cosmetic tools to study microbiomes at various levels of resolution — down to the level of individual pores.

Low complexity. Skin microbiomes relatively few stable species, with many strains of each species coexisting on each person. This simplicity makes complete (including phage!) ecological and evolutionary characterization feasible.

In addition, many human skin diseases to not have good animal models; we hope to uncover the mechanisms of skin diseases, including atopic dermatitis and acne while asking fundamental questions about microbiome communities.


We strive to:

  • Address questions without clear answers in all real-world microbial communities
  • Enable strong inferences and clear answers by gathering the highest quality data
  • Choose approaches that are complex as required, but no more complex