The impact of viral mutations on antibodies used to treat COVID-19


Jonathan Li, @DrJLi

Jesse Bloom, @jbloom_lab

Tyler Starr, @tylernstarr

Allison Greaney, @AllieGreaney

Manish Choudhary, @Dr_MChoudhary


These slides at


Classical way to identify escape mutations is to grow virus with antibody in lab

"... novel spike mutants rapidly appeared after in vitro passaging in the presence of individual antibodies..."

Limitation of this approach is you only find some of the possible escape mutants

We express SARS-CoV-2 receptor binding domain (RBD) on surface of yeast


fluorescent antibody

fluorescent tag on RBD

Library of yeast, each expressing different RBD mutant

We map escape mutations by sorting for RBD variants that don't bind antibody

In maps, tall letters indicate strong escape mutations

Sort done at single antibody concentration, which determines sensitivity!

Escape map for REGN10987

Tall letters = strongly escape binding. For instance, mutations at 446 escape, but those at 484 don't. However, quantitative relationship of letter height to IC50 depends on FACS gate: we confidently capture mutations with >100 fold effects, but can miss smaller ones.

Image shows key sites of interest, click here for interactive map of all mutations.

We now have escape maps for many lead clinical antibodies

Note: these are high-throughput maps generated using yeast display. We have found excellent correlation between the maps and pseudotyped viral neutralization assays (see papers above), but still recommend validating key conclusions.

In persistently infected patient, five RBD mutations after REGN-COV2 treatment

Patient described in Choi, Choudhary, ...,. Cernadas, Li. New England J Medicine (2020)


Note extensive genetic hitchhiking and competition among viral lineages; also described in within-patient evolution of influenza in persistent infections: Xue et al, eLife (2017)

Importantly, our complete maps show that four of these mutations reduce binding by one or the other of the REGN-COV2 antibodies, only one of which was identified by Regeneron's escape selections. Illustrates value of complete maps.

Mutations in emerging variant lineages:

Studies validating these conclusions from the mapping:

Crowe lab (Vanderbilt):

  • James Crowe
  • Seth Zost
  • Pavlo Gilchuk



Bloom lab (Fred Hutch)

  • Tyler Starr

  • Allie Greaney

  • Adam Dingens

  • Amin Addetia

  • Will Hannon

Li lab (Brigham & Women's):

  • Jonathan Li
  • Manish Choudhary


By Jesse Bloom


Viral mutations and antibodies used to treat COVID-19

  • 679