Abstract: 

The number of infections reported by healthcare systems tend to be a subset of total number of infections. This may be because symptoms of infected individuals are too mild to seek medical care, or they are not officially diagnosed with the infection due to the lack of testing. Both of these factors play a role in the current COVID-19 pandemic.
In many places, increasing testing rates is not an option due to resource constraints. However, we can get an estimate of the number of undetected infections therefore requires inference using mathematical models. One particular approach I will focus on in this session is the use of phylodynamic models of infectious disease transmission, which describe not only how many infections we expect over time but also the expected changes in the viral genome.
The results of the analysis points to large numbers of undetected infections. The consequence of this is that the coronavirus can spread latently through the population, and across great distances when infected individuals are still able to travel when they are infectious. This also means that eliminating COVID-19 will require population-wide social distancing in the absence of a vaccine.

Bio: 

Lucy Li is a data scientist at Chan Zuckerberg Biohub whose interests are in the application and development of machine learning and mathematical modeling methods for infectious diseases, with a focus on producing actionable information for infection prevention and control teams. Prior to joining CZ Biohub, Lucy obtained a Ph.D. from Imperial College London and carried out postdoctoral research at Harvard University, both within the fields of infectious disease epidemiology and global health.