Related Experiment Videos
Why do we need so many chemical similarity search methods?
Robert P Sheridan1, Simon K Kearsley
1Dept of Molecular Systems, RY50SW-100 Merck Research Laboratories, Rahway, NJ 07065, USA. sheridan@merck.com
Drug Discovery Today
|January 28, 2003
Summary
Searching chemical structure databases with similarity methods is key for drug discovery. Different methods yield varied results, so using multiple approaches improves lead identification for therapeutic targets.
Area of Science:
- Computational chemistry
- Medicinal chemistry
- Drug discovery
Background:
- Computational tools are vital for identifying potential drug leads in early-stage drug discovery.
- Similarity methods are a diverse and powerful class of computational tools for searching chemical databases.
Purpose of the Study:
- To share practical insights from years of experience using in-house similarity search methods.
- To highlight the variability and complementarity of different similarity search techniques in drug discovery.
Main Methods:
- Utilized in-house computational similarity search methods.
- Applied methods across various therapeutic areas and chemical structure databases.
- Analyzed the performance and output of different similarity metrics.
Main Results:
- The effectiveness of similarity methods is highly dependent on the specific biological activity and difficult to predict.
- Different similarity methods identify distinct subsets of active compounds from a database.
- No single method consistently outperforms others across all therapeutic problems.
Conclusions:
- Employing multiple similarity search methods is recommended to maximize the identification of diverse active compounds.
- Understanding the limitations and variability of each method is crucial for effective lead discovery.
- A combined approach using several search strategies enhances the comprehensiveness of chemical database exploration.