Absolute binding free energy calculations improve enrichment of actives in virtual compound screening

Abstract

We determined the effectiveness of absolute binding free energy (ABFE) calculations to refine the selection of active compounds in virtual compound screening, a setting where the more commonly used relative binding free energy approach is not readily applicable. To do this, we conducted baseline docking calculations of structurally diverse compounds in the DUD-E database for three targets, BACE1, CDK2 and thrombin, followed by ABFE calculations for compounds with high docking scores. The docking calculations alone achieved solid enrichment of active compounds over decoys. Encouragingly, the ABFE calculations then improved on this baseline. Analysis of the results emphasizes the importance of establishing high quality ligand poses as starting points for ABFE calculations, a nontrivial goal when processing a library of diverse compounds without informative co-crystal structures. Overall, our results suggest that ABFE calculations can play a valuable role in the drug discovery process.

Figure 1

Comparison of the ability of ABFE calculations versus docking to distinguish active compounds from inactives (decoys), shown as Receiver Operating Characteristic (ROC) curves with the Area Under Curve (AUC) statistics for the 30 Tier 1 and 30 Tier 2 compounds of all three protein targets, as labeled. Red: docking results. Blue: ABFE results. These ABFE calculations omit the free energy term for ligand protonation state changes that were incorporated into the docking calculation. However, adding this term to the ABFE results leads to negligible changes in the AUC statistics (maximum change 0.02, mean change 0.00).

Figure 2

Comparison of distributions of computed ABFE values for inactive (decoy) and active compounds for Tier 1 and Tier 2 of all three protein targets, as labeled.

Figure 3

Initial ligand conformations from docking and for the two independent runs of BACE1 active compound CHEMBL1090542. The same initial ligand pose (left) from docking relaxes to two different conformations during the MD equilibration step. In the run that yields the more favorable BFE (middle), the ligand stays close to the initial pose, whereas in the less favorable run (right) the ligand drifts away and loses interaction with the catalytic Asp residues, resulting in much less favorable binding free energy. BACE1 is shown in ribbon representation, the two catalytic Asp residues in ball and stick representation, and the ligand in licorice representation.