Empirical scoring functions for advanced protein-ligand docking with PLANTS

Area of Science:

• Computational chemistry
• Structural biology
• Bioinformatics

Background:

• Protein-ligand interactions are crucial in drug discovery.
• Accurate prediction of binding poses is essential for virtual screening.
• Existing scoring functions and docking algorithms have limitations in pose prediction accuracy.

Purpose of the Study:

• To develop and validate novel empirical scoring functions for the PLANTS docking algorithm.
• To improve the accuracy of protein-ligand binding pose prediction.
• To optimize parameter settings for the PLANTS algorithm to balance speed and reliability.

Main Methods:

• Development of two empirical scoring functions: PLANTS(CHEMPLP) and PLANTS(PLP).
• Parametrization of scoring functions using two test sets (298 complexes).
• Utilized ant colony optimization (ACO) within the PLANTS docking algorithm.
• Comparison with the GOLD docking tool on the Astex diverse set.

Main Results:

• Achieved excellent performance in pose prediction on test sets.
• Reproduced 87% of Astex diverse set complexes and 77% of CCDC/Astex clean list complexes within 2 Å RMSD.
• Demonstrated improved pose prediction performance compared to the GOLD docking tool, particularly for drug-like molecules.
• Identified optimized parameter settings for the PLANTS search algorithm.

Conclusions:

• The PLANTS(CHEMPLP) and PLANTS(PLP) scoring functions represent a significant advancement in protein-ligand docking.
• The PLANTS algorithm with optimized parameters offers a reliable and efficient tool for pose prediction.
• These findings contribute to more accurate virtual screening and drug design efforts.