Machine Learning and Integrative Structural Dynamics Identify Potent ALK Inhibitors from Natural Compound Libraries

Area of Science:

• Computational chemistry
• Drug discovery
• Oncology

Background:

• Anaplastic lymphoma kinase (ALK) is a key oncogenic driver in non-small cell lung cancer and other malignancies.
• Targeting ALK with small molecules is a clinically relevant strategy for cancer treatment.

Purpose of the Study:

• To identify novel ALK inhibitors from natural product-derived compounds using a computational approach.
• To leverage machine learning and molecular dynamics for prioritizing drug candidates.

Main Methods:

• Structure-based virtual screening of the ZINC20 database.
• Machine learning model training and benchmarking for structure-activity relationship analysis.
• 100 ns molecular dynamics simulations and binding free energy calculations.

Main Results:

• Six potential ALK inhibitors were shortlisted based on multiple criteria.
• ZINC3870414 and ZINC8214398 showed stable ligand engagement and favorable binding free energy.
• Molecular dynamics indicated these compounds restrict protein conformational sampling and alter residue communication pathways.

Conclusions:

• ZINC3870414 and ZINC8214398 represent promising scaffolds for ALK inhibitor development.
• The integrated computational approach, combining ML with dynamic and network metrics, is effective for early-stage kinase inhibitor discovery.