Anticipating the evolution and dynamics of emerging pathogen threats.
To prepare for future pandemic threats, we need better data and analytics for anticipating and detecting zoonoses with high pathogenic potential. Once a threat has been identified, we need better tools for rapid and accurate characterization of the threat, including its transmissibility, severity, and potential to evolve, as well as the likely efficacy of various containment strategies.
Global initiatives, like the CDC’s new Center for Forecasting and Analytics will expand access to critical data. Alongside, we aim to develop complex systems models that integrate pathogen dynamics and evolution with our observational processes themselves (data collection) to rapidly detect and predict anomalous activity from available data. We focus on molecular and structural responses of the pathogen, which will ultimately be integrated with signals from electronic health records, retail transactions, work and school absenteeism, media and social media, or GPS traces.
Pilot Study — Retrodicting the evolution of SARS-CoV-2 variants. We aim to develop methods to predict pathogen evolution at all stages of a pandemic, and to validate these methods retrospectively against the known history of SARS-CoV-2 emergence.
We are using machine-learning methods to ask whether we could have predicted the emergence of variants of concern given all the information we have today about circulating SARS-like viruses, the structure of the SARS-CoV-2 spike protein and its interaction with the ACE2 receptor. We aim to evaluate diverse prediction methods, identifying promising approaches, and vet pathogen retrodiction as a research tool.