Related Experiment Videos
Guiding molecules towards drug-likeness.
William J Egan1, W Patrick Walters, Mark A Murcko
1Vertex Pharmaceuticals Inc, 130 Waverly Street, Cambridge, MA 02139-4242, USA. bill_egan@vrtx.com
Current Opinion in Drug Discovery & Development
|August 29, 2002
Summary
This review explores computational methods for predicting drug-likeness, covering historical approaches and modern algorithms. It highlights areas for improvement in datasets, model interpretability, and optimization techniques for drug discovery.
Area of Science:
- Computational chemistry
- Medicinal chemistry
- Drug discovery
Background:
- Drug-likeness is crucial for successful drug development.
- Historical lead generation and optimization practices are reviewed.
- Properties of known drugs, fragments, and scaffolds are surveyed.
Purpose of the Study:
- To review computational methods for predicting drug-likeness.
- To identify areas for advancement in drug-likeness prediction.
- To discuss optimization techniques and pattern recognition algorithms.
Main Methods:
- Review of published literature on computational drug-likeness prediction.
- Analysis of methods for delineating chemical space.
- Examination of similarity metrics and advanced pattern recognition algorithms.
Main Results:
- Comprehensive overview of existing computational drug-likeness prediction strategies.
- Identification of limitations in current datasets and model interpretability.
- Discussion of multivariate optimization and statistical methods.
Conclusions:
- Computational drug-likeness prediction is an evolving field with significant potential.
- Improvements in dataset scope, model interpretability, and statistical methods are needed.
- Advanced algorithms and multivariate optimization can enhance drug design.