The novel coronavirus (SARS-CoV-2) pandemic resulted in the largest public health crisis in recent times. Significant drug design effort against SARS-CoV-2 is focused on the receptor-binding domain (RBD) of the spike protein, although this region is highly prone to mutations causing therapeutic resistance. We applied deep data analysis methods on all-atom molecular dynamics simulations to identify key non-RBD residues that play a crucial role in spike−receptor binding and infection. Because the non-RBD residues are typically conserved across multiple coronaviruses, they can be targeted by broad-spectrum antibodies and drugs to treat infections from new strains that might appear during future epidemics.
The MD trajectories are available from Zenodo at [doi.org/10.528c1/zenodo.5052691][1]. The codes and the residue correlation data used in this study are available from GitHub at <https://github.com/dhimanray/COVID-19-correlation-work>. All further details about the methods and the data are available within the article and [ SI Appendix ][2].
After the paper was accepted we became aware of a concurrent study by Mugnai et al. in long-range allosteric communication between the SARS-CoV-2 spike and ACE2 receptor ([85][3]).