Classification of kinase inhibitors using BCUT descriptors.

Area of Science:

• Computational Chemistry and Cheminformatics
• Molecular Descriptors and Drug Design
• Biochemistry and Kinase Inhibitor Research

Background:

• Molecular descriptors are crucial for understanding structure-activity relationships in drug design.
• Buried Connectivity (BCUT) descriptors encode topological and electronic information of molecules.
• Accurate classification of kinase inhibitors is vital for targeted cancer therapy.

Purpose of the Study:

• To evaluate the efficacy of BCUT descriptors for classifying ATP site-directed kinase inhibitors.
• To assess BCUTs' performance against diverse protein kinase families (serine/threonine and tyrosine kinases).
• To compare BCUTs with other descriptor types for insights into their information content.

Main Methods:

• Utilized BCUT descriptors to represent kinase inhibitors.
• Employed Partial Least Squares (PLS) discriminant analysis for classification.
• Tested classification accuracy against five protein kinases, including EGFR.

Main Results:

• BCUTs, when combined with PLS discriminant analysis, accurately classified inhibitors based on their target kinase.
• A novel class of kinase inhibitors was correctly predicted to target the EGFR tyrosine kinase.
• BCUTs demonstrated comparable or superior performance to 2D fingerprints and 3D pharmacophore descriptors.

Conclusions:

• BCUT descriptors provide valuable information for discriminating between kinase inhibitors targeting different protein kinases.
• BCUTs offer a computationally efficient approach for kinase inhibitor classification and drug discovery.
• The findings highlight the utility of BCUTs in identifying novel kinase inhibitor activities and guiding drug development.