A computational pipeline to discover potential cross-reactive antibodies: a case study on coronavirus

Cross-reactive antibody (crAb), which recognizes conserved viral epitopes across diverse strains, has emerged as powerful tools for diagnostics, therapeutic targeting, and pandemic preparedness. Yet conventional crAb discovery primarily relies on hybridoma technology and phage display which are time-consuming and labor-intensive.

Here, we present an integrative computational pipeline that leverages conformational epitope prediction and epitope immunogenic similarity analysis to compute cross-reactive epitopes, so as to prioritize potential crAbs undisclosed before.

Taking coronavirus as an example, we demonstrate that this pipeline successfully predicted candidate crAbs for five coronavirus antigens, including SARS-CoV-2 variants (WT, Beta, Omicron B.1, XBB.1.5) and one SARS-CoV variant. Out of Top20 predicted candidates, 45% were validated in CoV-AbDab as cross-reactive across coronavirus variants, including experimentally confirmed crAbs such as P17 (IC₅₀ = 0.165 ng/mL against SARS-CoV-2 WT), BG7-15 (IC₅₀ = 16 ng/mL against SARS-CoV-2 WT), and Beta-54 (IC₅₀ = 1 ng/mL against Gamma variant).

Though in pilot-study, this pipeline might serve as a scalable and efficient strategy for rapidly prioritizing potential crAbs in research of infectious disease.