Accepted_test

Using a neural network algorithm to predict binding affinity in ACE2-RBD protein-protein complexes
by Elizaveta Bogdanova | Chernukhin Artem | Shaitan Konstantin | Novoseletsky Valery | Moscow State University | Mendeleev University of Chemical Technology of Russia | Moscow State University | Shenzhen MSU-BIT University
Abstract ID: 459
Event: BGRS-abstracts
Sections: [Sym 3] Section “Structural biology of proteins nucleic acids and membranes”

In this work, we assessed the experimentally obtained structures of 48 complexes of the ACE2 receptor with the RDB S protein of the coronaviruses SARS-CoV and SARS-CoV-2 (including mutant forms of the latter), for which the dissociation constant was calculated. To predict binding affinity, we used the neural network algorithm ProBAN, which we had previously developed, as well as a few other algorithms for estimating the Gibbs free energy: Prodigy, FoldX, DFIRE and ROSETTADOCK. A comparison of the evaluation results shows that ProBAN has the best prediction quality (Pearson correlation 0.56, MAE 0.7 kcal/mol) of all the analyzed algorithms. The results obtained suggest better quality of affinity prediction for other protein-protein complexes.