ALPACA: A machine Learning Platform for Affinity and selectivity profiling of CAnnabinoids receptors modulators

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Pharmacology

Background:

• Targeting cannabinoid receptor 2 (CB2R) offers therapeutic potential for neurodegeneration and cancer.
• Selective CB2R modulation is crucial to avoid undesired psychotropic effects associated with CB1R.
• In-silico tools are needed to predict CB2R affinity and selectivity.

Purpose of the Study:

• To develop and validate machine learning classifiers for predicting CB2R affinity and selectivity.
• To create a user-friendly platform for accessing these predictive models.
• To facilitate the design of selective CB2R modulators.

Main Methods:

• Machine learning classifiers were trained using high-quality bioactivity data from ChEMBL v30.
• Data included small molecules acting on CB1R and CB2R.
• Random forest algorithms were employed, achieving high predictive performance (AUC ≥0.96).

Main Results:

• Developed accurate machine learning models for predicting CB2R affinity, CB1R affinity, and CB2R/CB1R selectivity.
• Random forest models demonstrated superior performance.
• A web platform, ALPACA, was created for free access to these predictive tools.

Conclusions:

• The developed classifiers exhibit strong predictive power for CB2R affinity and selectivity.
• ALPACA provides a valuable, user-friendly resource for drug discovery.
• This tool can significantly reduce time and resources in designing selective CB2R modulators.