KinasePred: A Computational Tool for Small-Molecule Kinase Target Prediction.

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology and Cheminformatics
• Pharmacology and Drug Discovery

Background:

• Protein kinases regulate crucial cellular functions and are key targets in diseases such as cancer.
• Drug discovery efforts heavily focus on identifying small molecules that can modulate kinase activity.

Purpose of the Study:

• To develop KinasePred, a computational workflow integrating machine learning and explainable AI for predicting small molecule kinase activity.
• To provide insights into structural features governing ligand-target interactions and kinase selectivity.

Main Methods:

• Development of a machine learning model for kinase activity prediction.
• Integration of explainable AI to elucidate structure-activity relationships.
• Validation through virtual screening and development of target-focused models.

Main Results:

• KinasePred demonstrated significant predictive performance.
• Successful identification of six kinase inhibitors via virtual screening.
• Identification of molecular determinants for kinase selectivity.

Conclusions:

• KinasePred accelerates the screening and identification of kinase-targeting compounds.
• The framework supports target identification, polypharmacology, and off-target effect analysis.
• Provides a versatile tool for streamlining the drug discovery process.