Related Experiment Video
Updated: Jul 30, 2026
11:13
Identification of Kinase-substrate Pairs Using High Throughput Screening
Published on: August 29, 2015
Kinase inhibitor recognition by use of a multivariable QSAR model
D G Sprous1, John Zhang, Lei Zhang
1CytRx Laboratories, 1 Innovation Drive, Worcester MA 01605, USA. dsprous@cytrx.com
Journal of Molecular Graphics & Modelling
|October 29, 2005
Summary
Kinase inhibitors possess distinct chemical properties, enabling a quantitative structure-activity relationship (QSAR) model to differentiate them from other drugs. This model aids in evaluating commercial compound libraries for drug discovery screening.
Area of Science:
- Medicinal Chemistry
- Computational Drug Discovery
- Pharmacology
Background:
- Kinase inhibitors are a crucial class of drugs, but distinguishing them from non-kinase drug-like molecules can be challenging.
- Understanding the unique chemical space and physical properties of kinase inhibitors is essential for efficient drug discovery.
Purpose of the Study:
- To identify distinct chemical scaffolds and R-group properties of kinase inhibitors compared to non-kinase drugs.
- To develop a quantitative structure-activity relationship (QSAR) model for distinguishing kinase inhibitors.
- To evaluate commercial compound libraries for their kinase inhibitor content to aid high-throughput screening (HTS) efforts.
Main Methods:
- Applied a retrosynthetic program to analyze chemical space of known kinase inhibitors and non-kinase drugs.
- Developed a multivariable QSAR model using distinct chemical fragments and physical properties as descriptors.
- Validated the QSAR model using reserved test sets and applied it to assess commercial compound libraries (Asinex, BioFocus, ChemDiv, LifeChemicals).
Main Results:
- Kinase inhibitors exhibit unique chemical fragment and physical property profiles compared to non-kinase drugs.
- The developed QSAR model achieved 98% recognition of training set kinase inhibitors with a 15% false positive rate.
- Model performance remained robust down to 70% data reserve for testing, indicating good generalization.
- Significant variations in kinase inhibitor populations were observed across commercial vendors, with BioFocus showing the highest proportion.
Conclusions:
- A QSAR model effectively differentiates kinase inhibitors based on their distinct chemical characteristics.
- The model provides a valuable tool for prioritizing compound libraries for HTS in kinase inhibitor drug discovery.
- Vendor analysis reveals differential enrichment of potential kinase inhibitors in commercial libraries, guiding procurement decisions.
