Benchmarking ANI potentials as a rescoring function and screening FDA drugs for SARS-CoV-2 Mpro

Abstract

Here, we introduce the use of ANI-ML potentials as a rescoring function in the host-guest interaction in molecular docking. Our results show that the "docking power" of ANI potentials can compete with the current scoring functions at the same level of computational cost. Benchmarking studies on CASF-2016 dataset showed that ANI is ranked in the top 5 scoring functions among the other 34 tested. In particular, the ANI predicted interaction energies when used in conjunction with GOLD-PLP scoring function can boost the top ranked solution to be the closest to the x-ray structure. Rapid and accurate calculation of interaction energies between ligand and protein also enables screening of millions of drug candidates/docking poses. Using a unique protocol in which docking by GOLD-PLP, rescoring by ANI-ML potentials and extensive MD simulations along with end state free energy methods are combined, we have screened FDA approved drugs against the SARS-CoV-2 main protease (M^pro). The top six drug molecules suggested by the consensus of these free energy methods have already been in clinical trials or proposed as potential drug molecules in previous theoretical and experimental studies, approving the validity and the power of accuracy in our screening method.

Keywords: Binding mode; Interaction energy; Molecular docking; Rescoring; Scoring function.

Fig. 1

The success of different scoring functions on predicting the first poses as the closest structure to the crystal with RMSD less than 2.0 Å. ANI is ranked 8th among 34 functions

Fig. 2

The success of different scoring functions on predicting the first poses as the exact crystal structure. ANI is ranked 5th among 34 functions

Fig. 3

The Spearman correlation coefficients between RMSD and scores for each scoring function being tested. The x-axis indicates the RMSD bins (≤ 2, ≤ 3, etc.) Heat map ranging between 0 and 1 indicates low/high correlations

Fig. 4

The structure of the SARS-CoV-2 main protease (PDB ID:7L14) used in GOLD docking. The nearby residues around ligand by 10 Å are highlighted.ANI scores were calculated using $\mathrm{\Delta }{G}_{\mathit{ANI}}=E_{\mathit{complex}}-(E_{\mathit{protein}}+E_{\mathit{ligand}})$

Fig. 5

GOLD-PLP and ANI scoring on the poses generated by Gold docking on 14 selected inhibitors of SARS-CoV-2 main protease. ANI shows superiority to GOLD in the top one solution

Fig. 6

Docking scores generated by a GOLD-PLP and b ANI with respect to experimental BFEs. ANI scores were produced by scaling with 0.127 as outlined in [41]

Fig. 7

Our unique screening protocol involving consensus scoring of ANI and GOLD, classical MD simulations and consensus BFEs by end-state methods

Fig. 8

Application of elimination (red)/selection (blue) criteria to MD trajectories. Δ(RMSD) ≤ 3.0 Å for the ligand, and Δ(#H-bonds) ≥ -0.5 between ligand and protein