Collaborating Against COVID

A Conversation with Kelly Gaither, Director of Health Analytics, TACC

The COVID-19 pandemic created an opportunity for TACC experts to work closely with renowned epidemiologist Lauren Ancel Meyers and her team at UT Austin. Leading the charge from TACC was Kelly Gaither, whose long history in scientific modeling, data science, and visualization brought new skills to the team. Her experience is showing more than ever that computational experts have to be boundary spanners and bridge builders to advance science.

How would you characterize TACC’s relationship with UT epidemiologist Lauren Ancel Meyers?

Lauren is a leading epidemiologist who has spent her career planning for infectious disease outbreaks including SARS, Ebola, H1N1, Zika, and now COVID-19. We have worked with her team to provide advanced computing, portal development, visualization, software development and best practices as needed since 2008. Most recently, and prior to the COVID-19 pandemic, she reached out to us to collaborate with her on a Centers for Disease Control and Prevention proposal to expand work on the Texas Pandemic Flu Toolkit for other types of viruses. We were gearing up to work with her and her team on that project. In mid-March, COVID-19 hit. We dropped everything to embed with her team and work as hard as we could to help her provide production-level support to policy and health decision makers across the country.

What is the UT Austin COVID-19 Modeling Consortium?

The consortium is a group of UT Austin researchers and their collaborators from around the world who want to tackle COVID-19 from all directions. The membership is in the hundreds — physicians, epidemiologists, mathematicians, computer scientists, biologists, and students. This group meets every morning to discuss problems relevant to COVID-19 that need to be solved and share ideas to better understand the science and landscape of this pandemic.

How is the COVID-19 Pandemic Portal being used?

TACC worked with the Meyers team to build and launch the COVID-19 Pandemic Portal to predict the spread of infections and communicate risk to the public. New projections launch weekly. The portal gets researchers up and running on supercomputers, specifically Frontera and Stampede2 at TACC. It’s also where we host mortality projections for major metropolitan areas in the U.S. We envision a situation in which the portal will become a clearinghouse for testing computational models.

Why are computational models so powerful?

They allow you to run ‘what if’ scenarios. There’s no other way to know how things might play out without models. This team is using two specific models – one is purely statistical that looks at projections for mortalities due to COVID-19; the other is the SEIR (Susceptible, Exposed, Infectious, Recovered) model, which analyzes transmission during the different stages of an epidemic outbreak.

What has this effort meant for you personally?

I have enjoyed working with Lauren’s team and the collaborators in the COVID19 consortium. It’s really gratifying to be able to plug into work that has such a profound impact on human health. Additionally, I have thoroughly enjoyed working so closely with the Meyers lab, soaking up as much as I can about epidemiology. Without my team members at TACC, we would not have been able to plug in so effectively.

What are you learning from this unique situation of COVID-19 in 2020?

This has been a scenario in which the floor was moving underneath us every day. We were responding to the urgency as it was happening. We will take what we have learned from our work with COVID-19 and adapt the models, tools, and technologies to respond to future outbreaks.

---

View the latest projections: covid-19.tacc.utexas.edu