An alternative method for the evaluation of docking performance: RSR vs RMSD

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Molecular docking is crucial for predicting ligand binding modes.
• Current evaluation methods like root-mean-square distance (RMSD) have limitations in assessing pose accuracy.
• Experimental electron density provides a direct measure of ligand fit within a protein's binding site.

Purpose of the Study:

• To introduce and validate a new assessment criterion for molecular docking poses.
• To evaluate the performance of docking programs using experimental electron density.
• To compare the efficacy of the real space R-factor (RSR) with RMSD.

Main Methods:

• Utilized three docking programs (Gold, Glide, Fred) on 88 protein-ligand complexes.
• Generated docking poses and compared them to known crystal structures.
• Calculated the real space R-factor (RSR) to measure ligand fit against experimental electron density.
• Compared RSR with the traditional root-mean-square distance (RMSD) metric.

Main Results:

• The RSR-based criterion effectively assesses docking pose accuracy.
• Identified shortcomings of the RMSD metric that are overcome by RSR.
• Demonstrated that RSR provides a more meaningful evaluation of docking results.
• Showcased the practical utility of RSR in evaluating docking methods.

Conclusions:

• The RSR criterion offers a superior method for evaluating molecular docking poses compared to RMSD.
• Incorporating experimental electron density enhances the reliability of docking pose assessment.
• This approach has significant implications for the development and application of docking methodologies in drug discovery.