MIMUMBA revisited: torsion angle rules for conformer generation derived from X-ray structures

Area of Science:

• Computational chemistry
• Structural biology
• Drug discovery

Background:

• Accurate prediction of molecular conformation is crucial for drug discovery.
• Existing methods for conformational analysis may lack sufficient quality and coverage.
• The Cambridge Crystallographic Database (CSD) provides valuable experimental structural data.

Purpose of the Study:

• To develop an automated method for generating rules of preferred torsion angles for drug-size molecules.
• To improve the quality and coverage of conformational space explored during molecular modeling.
• To enhance the accuracy of conformer generation for drug design and analysis.

Main Methods:

• Automatic generation of SMARTS patterns for four-atomic torsional fragments.
• Searching experimental structures within the Cambridge Crystallographic Database (CSD).
• Deriving rules for preferred torsion angles applicable to drug-size molecules.
• Utilizing the OMEGA conformer generator with the newly derived rules.

Main Results:

• The developed method significantly improves the quality and coverage of conformational space.
• The percentage of structures with a conformation within 0.5 Å of the X-ray structure increased from 84% to 92%.
• The average RMS distance of the closest conformation to the X-ray structure improved from 0.30 Å to 0.22 Å.

Conclusions:

• The automated generation of torsion angle rules enhances conformer generation accuracy.
• This approach offers a significant improvement over the default OMEGA setup for conformational analysis.
• The method provides a valuable tool for more reliable conformational exploration in drug discovery.