Computational Modeling of Kinase Inhibitor Selectivity.

Area of Science:

• Pharmacology
• Computational Biology
• Medicinal Chemistry

Background:

• Kinase inhibitors are crucial therapeutics, but off-target effects can cause toxicity.
• Understanding kinase inhibitor selectivity is essential for drug development and repurposing.

Purpose of the Study:

• To computationally predict off-target kinase interactions for 15 therapeutic kinase inhibitors.
• To validate predictions against available experimental data and assess accuracy.
• To explore the potential for repurposing known kinase inhibitors to new targets.

Main Methods:

• Utilized computational methods to predict inhibitor selectivity against approximately 480 human kinases.
• Compared predictions with experimental data for around 280 kinase targets.
• Validated predictive models using recent experimental data for sorafenib and sunitinib.

Main Results:

• Achieved an average prediction accuracy and specificity of approximately 90% for kinase off-target interactions.
• Demonstrated high predictive accuracy for sorafenib and sunitinib against new experimental data.
• Successfully predicted interactions across the human kinome.

Conclusions:

• Computational predictions of kinase inhibitor selectivity are highly accurate.
• This approach facilitates the identification of new therapeutic opportunities by repurposing existing kinase inhibitors.
• The findings support the development of targeted therapies with reduced off-target effects.