Research & Interests

Below, I briefly summarize my research to date and elaborate on my interests and what I hope to continue to work on in the future. At the end of this page, you will find current and past projects.


Infectious Disease Modeling

There are many approaches to modeling infectious diseases depending on the research question. In one framework, I have considered the “who”, “where”, “origin” of a pathogen in a region that was previously thought to be resilient to infection. To determine “who” was being infected, I estimated “risk factors” accounting for confounding using inverse probability treatment weights estimated through the supervised SuperLearner (machine learning) algorithm; “where” infections were occurring using Bayesian Gaussian spatial process models; and the “origin” using basic genetic analyses. For other research questions estimating the severity of an infectious disease, I have employed Bayesian statistical modeling or helped to implement mathematical models for infection scenario and burden planning. Most recently, I have been exploring how differing features in contact networks make pandemic prediction difficult. Through these efforts, I have explored using AI/ML agent-based models for better pandemic prediction. Overall, I am most interested in the intersection of methods, combining molecular techniques, bioinformatics/genomics, evolutionary biology, statistical modeling, machine learning/artificial intelligence and spatial methods to identify pathways of infection and antimicrobial resistance evolution. Using an interdisciplinary approach, I hope to identify interventions that block these pathways.


Bioinformatics, Genomics, Statistical Genetics

Molecular surveillance and genomics greatly enrich infectious disease epidemiology analyses, models, and intervention planning. To date, a large majority of my bioinformatic and genomic work has focused on antimicrobial resistance, with an emphasis on putative antimalarial resistance. In the past, I have used genome-wide barcodes to infer how antimalarial selectional pressures have shaped Falciparum malaria population structures in the Democratic Republic of the Congo as well as whole genome sequencing to identify novel antimicrobial mutations in Staphylococcus epidermidis. Separately, I have been engaged in statistical genetics method development with models focused on estimating local inbreeding (DISCent), important for sink-source dynamics and interventional planning, as well as inferring identity by descent in polyclonal malaria infections (polyIBD).


Global Health & Medicine, Translational Research

Global medicine has become local medicine and vice versa. I am interested in supporting capacity building efforts through research in low- and middle-income countries and historically underserved regions.



AMR in the Hospital

Case Study of the evolution of Daptomycin in SEpi following prolonged antibiotic exposure

Antimalarial Resistance in Ethiopia

Estimating the prevalence of antimalarial mutations in Ethiopia across time and space

COVID19 IFR Re-estimation

Estimating the COVID-19 first wave IFR using serologic data accounting for seroreversion

DISCent: Spatial Inbreeding Estimation

An algorithm for detecting deme inbreeding spatial coefficients from recombining pathogen genetic data

Vivax in the DRC

Determing who was being infected, where infections were occuring, and where they originated

COVIDCurve: COVID-19 Statistical Modeling

Bayesian statistical model for estimating age-based infection fatality ratios from serologic data

polySimIBD: A Simple Malaria Genetic Model

Spatial Structured Wright Fisher Model for Malaria Population Genetic Simulations

Antimalarial Resistance in the DRC

Exploring population structure attributed to antimalarial resistance in the Democratic Republic of the Congo

Tanzania-Zanzibar Transmission

Coupling epidemiologic analysis with relatedness metrics (identity by descent) and population dynamics, we show that there is active …

What Makes Pandemic Prediction Difficult?

Improving Pandemic Prediction using agent-based models and deep-neural networks