ROSHAMBO: Open-Source Molecular Alignment and 3D Similarity Scoring.

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Virtual screening is crucial for identifying drug candidates.
• Existing proprietary tools are effective but costly.
• Structure-based virtual screening has limitations.

Purpose of the Study:

• To present an open-source package for molecular alignment and 3D similarity calculations.
• To provide a cost-effective alternative to proprietary virtual screening software.
• To complement existing structure-based virtual screening methods.

Main Methods:

• Utilized the PAPER software for Gaussian volume overlap-based molecular alignments.
• Implemented GPU acceleration to enhance computational speed and reduce resource usage.
• Calculated shape and pharmacophore-based color scores for molecular similarity assessment.

Main Results:

• Benchmarked the package using the DUDE-Z dataset, demonstrating near state-of-the-art performance.
• Showcased robustness across various target classes.
• Achieved significant speed improvements, facilitating routine ligand-based drug discovery.

Conclusions:

• The open-source package offers an accessible and efficient solution for virtual screening.
• It provides a viable alternative to proprietary software for drug discovery workflows.
• The tool's performance and speed support large-scale molecular similarity assessments.