Molecular insights on ABL kinase activation using tree-based machine learning models and molecular docking

Area of Science:

• Biochemistry and Molecular Biology
• Medicinal Chemistry
• Computational Drug Discovery

Background:

• Abelson kinase (c-Abl) is a non-receptor tyrosine kinase crucial for cellular processes like differentiation, migration, proliferation, and survival.
• Activating c-Abl presents a therapeutic strategy for conditions including chemotherapy-induced neutropenia, prostate cancer, and breast cancer.
• Recent identification of c-Abl activators offers new avenues for drug development.

Purpose of the Study:

• To characterize the critical chemical properties and interactions of identified c-Abl activators.
• To understand the structural features necessary for c-Abl activators to bind to the myristoyl pocket.
• To develop predictive structure-activity relationship (SAR) models for c-Abl activators.

Main Methods:

• Integrated structure-based drug design (SBDD) and ligand-based drug design (LBDD) methodologies.
• Employed molecular docking simulations to predict binding interactions.
• Utilized tree-based machine learning models (decision tree, AdaBoost, random forest) for analysis.

Main Results:

• Developed predictive and robust machine learning models with Matthews correlation coefficient values > 0.4.
• Identified key structural characteristics of c-Abl activators responsible for binding to the myristoyl pocket.
• Established a structure-activity relationship (SAR) model based on these identified characteristics.

Conclusions:

• The study successfully combined SBDD and LBDD with machine learning to elucidate c-Abl activator mechanisms.
• The identified structural features and SAR model provide a foundation for designing novel c-Abl activating drugs.
• This work facilitates the development of targeted therapies for diseases involving c-Abl dysregulation.