Integrating artificial intelligence-based epitope prediction in a SARS-CoV-2 antibody discovery pipeline: caution is warranted.

Background: SARS-CoV-2-neutralizing antibodies (nABs) showed great promise in the early phases of the COVID-19 pandemic. The emergence of resistant strains, however, quickly rendered the majority of clinically approved nABs ineffective. This underscored the imperative to develop nAB cocktails targeting non-overlapping epitopes.
Methods: Undertaking a nAB discovery program, we employed a classical workflow, while integrating artificial intelligence (AI)-based prediction to select non-competing nABs very early in the pipeline. We identified and in vivo validated (in female Syrian hamsters) two highly potent nABs.
Findings: Despite the promising results, in depth cryo-EM structural analysis demonstrated that the AI-based prediction employed with the intention to ensure non-overlapping epitopes was inaccurate. The two nABs in fact bound to the same receptor-binding epitope in a remarkably similar manner.
Interpretation: Our findings indicate that, even in the Alphafold era, AI-based predictions of paratope-epitope interactions are rough and experimental validation of epitopes remains an essential cornerstone of a successful nAB lead selection.
Funding: Full list of funders is provided at the end of the manuscript.
Competing Interests: Declaration of interests Ghent University has filed for patent protection on the antibody sequences described herein, and D.D.A., M.W., R.W., W.W., S.G. and L.V. are named as co-inventors on this patent (European Patent Application: 21186206.5). A.P. is employee of the MAbSilico, H.R. holds a patent regarding neutralizing VHH antibodies binding the Spike RBD (PCT/EP2021/052885) and has filed a priority application for neutralizing VHH antibodies binding Spike S2 (EP 23160838.1). X.S. is a recipient of FWO research project COVID-19 (G0G4920N) and FWO-FNRS project VIREOS (EOS ID: 30981113) grants.
(Copyright © 2023. Published by Elsevier B.V.)