Related Experiment Videos
Binary formal inference-based recursive modeling using multiple atom and physicochemical property class pair and
Cho1, Shen, Hermsmeier
1Combinatorial Drug Discovery, Bristol-Myers Squibb Company, Wallingford, Connecticut 06492-7660, USA. chos@bms.com
Journal of Chemical Information and Computer Sciences
|June 13, 2000
Summary
This study introduces a new computational method for analyzing large datasets from high-throughput screening. The binary formal inference-based recursive modeling approach effectively identifies structure-activity relationships using multiple molecular descriptors.
Area of Science:
- Computational chemistry
- Cheminformatics
- Bioinformatics
Background:
- Analyzing large datasets from high-throughput screening (HTS) is challenging.
- Identifying structure-activity relationships (SAR) is crucial for drug discovery.
Purpose of the Study:
- To develop an efficient computational method for analyzing large HTS datasets.
- To improve the identification of SAR using molecular descriptors.
Main Methods:
- Developed binary formal inference-based recursive modeling.
- Utilized atom and physicochemical property class pair and torsion descriptors.
- Implemented recursive partitioning with statistical hypothesis testing and multi-feature extraction.
Main Results:
- The method successfully distinguished random from real data sets.
- Partitioning using multiple descriptors proved advantageous for SAR analysis.
- Tested on 27,401 National Cancer Institute (NCI) compounds against the NCI-H23 cell line.
Conclusions:
- Binary formal inference-based recursive modeling is an effective tool for analyzing large chemical datasets.
- Employing multiple descriptors enhances the analysis of structure-activity relationships.
- The developed method aids in understanding compound activity in drug discovery contexts.