Machine Learning Approach for Predicting Drug-Like Molecules Targeting Calmodulin Pathway Proteins.

Area of Science:

• Computational chemistry
• Pharmacology
• Biotechnology

Background:

• Predicting drug interactions and biological activity is crucial but challenging.
• Integrating diverse data sources for accurate drug efficacy prediction remains a significant hurdle.

Purpose of the Study:

• To develop a novel machine learning model for predicting drug and assay efficacy for calmodulin-related diseases.
• To identify key factors influencing drug-target interactions using a predictive modeling framework.

Main Methods:

• Compiled a comprehensive dataset of drugs and compounds targeting calmodulin (CaM) complexes.
• Developed and validated the IFPTML-XGB machine learning model for efficacy prediction.
• Synthesized novel riluzole derivatives and performed experimental and computational assays, including molecular docking.

Main Results:

• The IFPTML-XGB model achieved high predictive performance with 89.1% test accuracy and 89.0% sensitivity.
• The model successfully predicted the bioactivity of novel riluzole derivatives.
• Identified key factors influencing drug-target interactions through IFPTML modeling.

Conclusions:

• The developed framework provides a robust method for assay efficacy prediction.
• This approach can accelerate the identification of promising drug candidates for calmodulin-related diseases.
• Enhances understanding of molecule-CaM interactions and facilitates new drug formulation development.