An Improved Receptor-Based Pharmacophore Generation Algorithm Guided by Atomic Chemical Characteristics and...

Area of Science:

• Computational Chemistry
• Drug Discovery
• Structural Bioinformatics

Background:

• Pharmacophore-based virtual screening is crucial for identifying novel drug candidates.
• Existing algorithms face challenges like ligand dependency and excessive feature extraction.

Purpose of the Study:

• To develop an improved, receptor-based pharmacophore generation algorithm.
• To overcome limitations of current ligand-dependent methods and reduce feature complexity.

Main Methods:

• An algorithm guided by atomic chemical characteristics and hybridization types was developed.
• Six chemical characteristics were extracted, including hydrophobic and aromatic features, with spatial constraints.
• The new algorithm was compared against the LigandScout algorithm.

Main Results:

• The developed algorithm successfully generated pharmacophores with six chemical characteristics.
• The method demonstrated fewer pharmacophore features and ligand independence.
• Effective differentiation between active and decoy molecules was achieved on a benchmark dataset.

Conclusions:

• The improved algorithm offers a more efficient and robust approach to pharmacophore generation.
• This method enhances virtual screening accuracy by reducing computational complexity and ligand bias.
• The findings contribute to advancing compound discovery in medicinal chemistry.