Accepted_test

Monoclonal neutralizing antibodies show promise in therapeutic drug development. Recently, generative diffusion models have been popularly used for designing new sequences and structures of these antibodies. To enhance the design efficiency, a quick and effective method for assessing antibody-antigen complex formation is required. Protein-protein interaction assessment is a computationally complex task. Therefore, we propose an evaluation method using the metadynamics method in the Martini 3 force field representation using the example of neutralizing antibody complexes targeting RBD epitopes of the SARS-CoV-2 S-protein. We have demonstrated that this approach allows us to obtain the initial structures of complexes. From these, a wider variety of antibodies can be produced, targeting different epitopes. Subsequently, the generated designs are evaluated based on the binding free energy of the antibody and antigen.