SEPPA-mAb: spatial epitope prediction of protein antigens for mAbs

Potent and broadly neutralizing antibodies against sarbecoviruses induced by sequential COVID-19 vaccination

The current SARS-CoV-2 variants strikingly evade all authorized monoclonal antibodies and threaten the efficacy of serum-neutralizing activity elicited by vaccination or prior infection, urging the need to develop antivirals against SARS-CoV-2 and related sarbecoviruses. Here, we identified both potent and broadly neutralizing antibodies from a five-dose vaccinated donor who exhibited cross-reactive serum-neutralizing activity against diverse coronaviruses. Through single B-cell sorting and sequencing followed by a tailor-made computational pipeline, we successfully selected 86 antibodies with potential cross-neutralizing ability from 684 antibody sequences. Among them, PW5-570 potently neutralized all SARS-CoV-2 variants that arose prior to Omicron BA.5, and the other three could broadly neutralize all current SARS-CoV-2 variants of concern, SARS-CoV and their related sarbecoviruses (Pangolin-GD, RaTG13, WIV-1, and SHC014). Cryo-EM analysis demonstrates that these antibodies have diverse neutralization mechanisms, such as disassembling spike trimers, or binding to RBM or SD1 to affect ACE2 binding. In addition, prophylactic administration of these antibodies significantly protects nasal turbinate and lung infections against BA.1, XBB.1, and SARS-CoV viral challenge in golden Syrian hamsters, respectively. Importantly, post-exposure treatment with PW5-5 and PW5-535 also markedly protects against XBB.1 challenge in these models. This study reveals the potential utility of computational process to assist screening cross-reactive antibodies, as well as the potency of vaccine-induced broadly neutralizing antibodies against current SARS-CoV-2 variants and related sarbecoviruses, offering promising avenues for the development of broad therapeutic antibody drugs.

Single-residue linear and conformational B cell epitopes prediction using random and ESM-2 based projections

B cell epitope prediction methods are separated into linear sequence-based predictors and conformational epitope predictions that typically use the measured or predicted protein structure. Most linear predictions rely on the translation of the sequence to biologically based representations and the applications of machine learning on these representations. We here present CALIBER 'Conformational And LInear B cell Epitopes pRediction', and show that a bidirectional long short-term memory with random projection produces a more accurate prediction (test set AUC=0.789) than all current linear methods. The same predictor when combined with an Evolutionary Scale Modeling-2 projection also improves on the state of the art in conformational epitopes (AUC = 0.776). The inclusion of the graph of the 3D distances between residues did not increase the prediction accuracy. However, the long-range sequence information was essential for high accuracy. While the same model structure was applicable for linear and conformational epitopes, separate training was required for each. Combining the two slightly increased the linear accuracy (AUC 0.775 versus 0.768) and reduced the conformational accuracy (AUC = 0.769).