sensaas: Shape-based Alignment by Registration of Colored Point-based Surfaces.

Area of Science:

• Computational chemistry
• Structural bioinformatics
• Cheminformatics

Background:

• Accurate molecular shape alignment is crucial for understanding structure-activity relationships and drug discovery.
• Existing methods for molecular shape comparison have limitations in accuracy and efficiency.

Purpose of the Study:

• To introduce SENSAAS, a novel tool for aligning and comparing molecular shapes and sub-shapes.
• To evaluate the performance of SENSAAS against established shape-alignment methods.

Main Methods:

• Utilizes 3D point-based representations of van der Waals surfaces for alignment.
• Employs a two-stage rigid-body superimposition: initial alignment via pose-invariant FPFH descriptors, refined by colored point matching.
• Points are colored based on the physico-chemical class of their closest atom.

Main Results:

• SENSAAS successfully reproduced the superimposition of X-ray structures from the AstraZeneca (AZ) dataset.
• Demonstrated superior performance compared to the shaep and shafts shape-alignment approaches.
• Illustrated effective sub-matching capabilities for substructures and bioisosteric fragments.

Conclusions:

• SENSAAS offers an efficient and accurate method for molecular shape alignment and comparison.
• The tool's ability to handle sub-shapes and colored points enhances its utility in cheminformatics.
• SENSAAS provides a valuable addition to the toolkit for molecular modeling and drug design.