SARS-CoV-2 variants with enhanced transmissibility and evasion of humoral immunity have propagated globally, representing a serious threat to global health. There is great interest in understanding the impacts of SARS-CoV-2 mutations on transmissibility. We have developed machine learning models that predict receptor-binding domain (RBD) mutations that modulate receptor (ACE2) affinity and escape from RBD-specific human antibodies. Importantly, these models predict many of the known impacts of RBD mutations, including in Variants of Concern, on receptor affinity and antibody escape. Additionally, the models identify novel single and multisite RBD mutations that are predicted to strongly modulate both properties linked to transmissibility. These models are expected to serve as valuable tools for continued SARS-CoV-2 surveillance and may be employed for guiding the development of next-generation vaccines and biotherapeutics.

Mutational analysis of SARS-CoV-2 variants using machine learning reveals key tradeoffs between receptor affinity and antibody escape

ACS Spring 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.