Towards automated binding affinity prediction using an iterative linear interaction energy approach.

Area of Science:

• Computational chemistry
• Molecular modeling
• Drug discovery

Background:

• Accurate binding free energy calculations are vital for drug development.
• Accounting for protein dynamics and ligand orientations is computationally challenging.
• The iterative Linear Interaction Energy (LIE) approach offers a balance of accuracy and efficiency.

Purpose of the Study:

• To demonstrate the automation of binding affinity prediction using the iterative LIE method.
• To apply this automated workflow to predict drug-protein interactions for Cytochrome P450 2D6.
• To explore methods for assessing prediction quality based solely on simulation data.

Main Methods:

• Development of an automated workflow integrating protein conformation selection, ligand docking, and molecular dynamics simulations.
• Application of the iterative LIE approach with Boltzmann weighting for binding free energy calculation.
• Utilizing preselected protein conformations and automated ligand docking/clustering.

Main Results:

• Successful automated prediction of binding affinities for aryloxypropanolamines to Cytochrome P450 2D6.
• Demonstrated ability to predict affinities without prior knowledge of specific protein-ligand conformations.
• Achieved predictions without compromising the computational efficiency of the LIE approach.

Conclusions:

• The iterative LIE approach is effective for automated, accurate binding affinity prediction.
• This automated workflow facilitates drug development by efficiently assessing drug-target interactions.
• Future work will focus on developing simulation-based quality assessment for LIE predictions.