GOLEM: Automated and Robust Cryo-EM-Guided Ligand Docking with Explicit Water Molecules

Area of Science:

• Structural biology
• Computational chemistry
• Drug discovery

Background:

• Understanding ligand-protein interactions is crucial for rational drug design.
• Cryo-electron microscopy (cryo-EM) provides high-resolution structural data but struggles with accurate ligand modeling due to ligand flexibility.
• Accurate small-molecule coordinate assignment in cryo-EM maps remains a significant challenge.

Purpose of the Study:

• To develop an automated and robust method for predicting ligand pose and conformation in cryo-EM maps.
• To address the limitations of current methods in modeling flexible ligands within cryo-EM density.
• To provide a tool that enhances the utility of cryo-EM data for drug discovery.

Main Methods:

• Development of GOLEM (Genetic Optimization of Ligands in Experimental Maps), a ligand docking method.
• Utilized a Lamarckian genetic algorithm for a hybrid global/local search of ligand conformational, orientational, and positional space.
• Incorporated molecular energetics and cryo-EM map correlation into the scoring function, including explicit consideration of water molecules.

Main Results:

• GOLEM successfully predicted ligand poses and conformations in cryo-EM maps.
• Validation against multiple cryo-EM structures with varying resolutions and ligand types demonstrated excellent agreement with experimental densities.
• The method effectively handles the dynamic and flexible nature of small-molecule ligands.

Conclusions:

• GOLEM is an effective and automated tool for ligand modeling in cryo-EM data.
• The method improves the accuracy of ligand coordinate assignment, facilitating structure-based drug design.
• As a free VMD plugin, GOLEM is accessible to the scientific community, advancing cryo-EM applications in drug discovery.