DeCAF-Discrimination, Comparison, Alignment Tool for 2D PHarmacophores

Area of Science:

• Cheminformatics
• Computational Chemistry
• Drug Discovery

Background:

• Small molecule comparison is crucial for drug design, repurposing, and predicting side effects.
• Current methods often rely on structural fingerprints or computationally expensive 3D conformer generation.
• Existing approaches may not fully capture spatial and physicochemical properties efficiently.

Purpose of the Study:

• To develop a conformation-free method for augmenting molecule representation with spatial and physicochemical properties.
• To enable efficient large-scale comparison of small molecules in cheminformatics.
• To introduce an open-source tool for discrimination, comparison, and alignment of 2D pharmacophores.

Main Methods:

• Representing molecules as undirected graphs where nodes are atoms with pharmacophoric properties and edges are feature distances.
• Developing the DeCAF (Discrimination, Comparison, Alignment tool for 2D PHarmacophores) open-source Python module.
• Utilizing a combination of pharmacophoric properties and spatial information without 3D conformer generation.

Main Results:

• DeCAF successfully augments molecule representation with spatial and physicochemical data.
• The method provides a conformation-free approach, reducing computational cost.
• Usage examples and statistical evaluation demonstrate DeCAF's capabilities and limitations.
• The tool can be manually adjusted for specific task optimization.

Conclusions:

• DeCAF offers an efficient and effective alternative for small molecule comparison in cheminformatics.
• The graph-based approach successfully integrates spatial information without conformational analysis.
• The open-source availability of DeCAF facilitates its adoption in various drug discovery workflows.