Research

photo credit: Kelvin Ma, Tufts Photography

TB continues to kill millions of people every year. Mycobacterium tuberculosis remains the second deadliest infectious agent in the world. Our research focuses on understanding how mycobacteria tolerate stress and perturb host cell biology to evade killing by antibiotic treatment and the host immune response. We integrate single-cell measurements and computational modeling to quantitatively describe stress tolerance and virulence mechanisms of mycobacteria.  We aim to use this knowledge to engineer improved therapeutics for treating tuberculosis and other mycobacterial diseases.

Phenotypic mycobacterial drug tolerance:
Using live-cell microscopy and quantitative image analysis, we recently found that mycobacteria exhibit an asymmetric growth pattern that deterministically generates closely related cells with different growth properties and tolerance to drug treatment. We continue to use live-cell microscopy and computational modeling to quantify the relative contributions of key mycobacterial cell physiological properties on the ability of distinct subpopulations to tolerate different stressors. We also study how cell cycle progression and growth heterogeneity are mediated.

Deciphering cellular complexity:
The behaviors of both host and bacterial cells are determined by a combination of many dynamic factors. We aim to measure these important factors in time and space and navigate cell complexity by utilizing mathematical modeling and analysis tools. Key to our approach is the use of phase diagrams to visualize cell phenotype relative to multiple parameters. These approaches are broadly applicable and have application to a broad range of biological and disease systems.

Data-driven design of combination therapy:
TB must be treated with combination therapy. We use DiaMOND, a geometric optimization of traditional drug combination assays, to efficiently measure drug interactions and combination efficacies. We aim to use systematic measurement of drug combinations to optimize TB drug regimens. Our methods are broadly applicable to other disease models.