Machine learning to accelerate diagnostics and therapeutics for COVID-19

The COVID-19 pandemic demands a rapid response of science, diagnostics, and therapeutics. Debora Marks’ lab at Harvard Medical School aims to accelerate those efforts using predictive models of the SARS-CoV2-19 sequences and 3D structures. The lab’s existing methods for statistical modeling of RNA and protein sequences enable them to predict the function of new virus sequences, solve 3D structures of virus proteins and RNA, and identify virus-host interactions. Currently, Prof. Marks’ team is completing the first-pass models of all SARS-CoV2-19 proteins and their interactions and designing sensitive multiplexed viral detection and optimized library designs for therapeutic nanobodies and peptides for ACE2 receptors.

To support as much translational research as possible, the lab is providing those models and predictions as a SARS-COV2-19 resource available to anyone on the web. Prof. Marks’ group is also working with wet-labs to develop sequence-based diagnostics, select optimal epitopes for vaccine design, and create optimal libraries of antibodies against the virus.

She is looking for partners in any area where discoveries can be accelerated.

Intellectual Property Status: Patent(s) Pending