Molecular Similarity Based on Surface Representations
Author Information
Author(s): Rahul Singh
Primary Institution: San Francisco State University
Hypothesis
Can molecular similarity be effectively determined using complex surface-based representations?
Conclusion
The proposed method allows for efficient and accurate determination of molecular similarity by directly comparing molecular surfaces and their properties.
Supporting Evidence
- The proposed method achieved 100% accuracy in query-retrieval settings with 5000 molecules.
- It outperformed existing methods in terms of computational speed.
- The method was validated through applications in structure-activity modeling.
Takeaway
This study shows a new way to compare molecules based on their shapes, which helps in finding similar molecules faster and more accurately.
Methodology
The study used a novel spherical representation to map molecular surfaces and computed similarity using histogram intersection of surface-property distributions.
Limitations
The method may not scale well with extremely large molecular databases and relies on the quality of the molecular surface representations.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website