Computer modeling to understand, track and prepare for COVID-19

Friday, May 08, 2020

Enabling local modeling of COVID-19

National models of COVID-19’s spread may be of limited use to state and local planners because the models generally average population data instead of considering the distinctive character of smaller regions. Sandia Labs researchers are using regional data provided by the State of New Mexico to determine the most effective regional distribution of medical supplies and resources and where pop-up testing centers would be best located. Sandia researchers have applied these techniques to other states and localities across the United States.

High-Performance Computing Consortium

Sandia’s Solo high-performance computing system has been made available to users through the COVID-19 High-Performance Computing Consortium within the NVBL consortium. In May at the consortium’s request, Solo began analyzing the possibility of repurposing antibodies used to fight previous coronaviruses for use in mixtures that could neutralize virus variants that cause COVID-19. Initial results look promising. The research continues to refine predictions and increase the number of antibodies that could be used to target COVID-19, simultaneously combining biological tests with computational experiments. The project has utilized more than 111,450 processor hours on Solo. COVID-related projects selected for study by the consortium and allocated to Solo will receive the highest priority on this system, which is designed specifically for external collaborations. Consortium projects that are complementary to ongoing Sandia COVID-19 projects bring the possibility of broader research collaborations.