Q-fit: a probabilistic method for docking molecular fragments by sampling low energy conformational space

Area of Science:

• Computational chemistry
• Molecular modeling
• Drug discovery

Background:

• Fragment-based drug discovery is crucial for identifying novel lead compounds.
• Accurate prediction of molecular fragment binding modes to protein receptors is computationally challenging.

Purpose of the Study:

• To present a new, computationally efficient method for docking molecular fragments to rigid protein receptors.
• To validate the method's accuracy in predicting binding modes using established test cases.

Main Methods:

• A probabilistic procedure based on statistical thermodynamics is employed to identify low-energy binding sites for ligand atom triplets.
• The method incorporates energy minimization and utilizes geometric hashing or pose clustering for conformational sampling.
• Search constraints are introduced to enhance computational efficiency in exploring conformational space.

Main Results:

• The method achieved high success rates (8/10 and 10/10) in correctly ranking binding modes for standard test fragments.
• For previously problematic fragment sets, success rates of 3/10 and 4/10 were observed.
• Both geometric hashing and pose clustering methods demonstrated high agreement and computational speed (approx. 11-13 seconds per placement).

Conclusions:

• The developed probabilistic docking method is computationally efficient and accurate for predicting molecular fragment binding modes.
• The approach shows promise for facilitating virtual screening and accelerating the identification of novel small molecule lead compounds.