Identifying Similar 3D Shapes of Molecules in PubChem
Author Information
Author(s): Evan E. Bolton, Sunghwan Kim, Stephen H. Bryant
Primary Institution: National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health
Hypothesis
Can a new 'Similar Conformers' relationship improve the identification of biologically relevant small molecules based on their 3D shapes?
Conclusion
The 'Similar Conformers' method effectively identifies similar small molecules that may be overlooked by traditional 2D methods, achieving a throughput of over 150,000 conformers per second.
Supporting Evidence
- The study identified 8.16 billion conformer neighbor pairs and 6.62 billion compound neighbor pairs.
- Using more diverse conformers per compound increases the number of 3D neighbors found.
- The method achieved a throughput rate of over 150,000 conformers per second.
Takeaway
This study helps find small molecules that look similar in 3D, which can help scientists understand how they might work in the body, even if they look different in 2D.
Methodology
The study used a series of filters to compute 3D similarity between conformers of small molecules, focusing on shape and feature similarities.
Limitations
The method relies on a limited number of diverse conformers per compound, which may not capture all possible 3D shapes.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website