SURFCOMP: a novel graph-based approach to molecular surface comparison

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Analyzing physicochemical properties on molecular surfaces is crucial for understanding compound behavior.
• Identifying similarities between molecules aids in rational drug design.
• Existing methods may lack efficiency in detecting local similarities.

Purpose of the Study:

• To develop and present a novel computational approach for detecting locally similar regions between molecular surfaces.
• To enhance the process of rational drug design by improving the analysis of molecular similarities.
• To reduce the complexity of comparing molecular surfaces with associated properties.

Main Methods:

• Utilizing maximal common subgraph comparison and harmonic shape image matching.
• Augmenting molecular surfaces with properties like electrostatic potential and lipophilicity.
• Employing geometric and physicochemical heuristics as filters to reduce computational complexity.

Main Results:

• The approach successfully detects locally similar regions between molecular surfaces.
• Tested on dihydrofolate reductase and thermolysin inhibitors, it accurately recovers compound alignments.
• Demonstrated effectiveness in identifying key interactions within active sites.

Conclusions:

• The presented method offers an efficient way to compare molecular surfaces based on geometric and physicochemical properties.
• This technique can significantly contribute to rational drug design by revealing subtle molecular similarities.
• The approach shows promise for applications in drug discovery and development.