Data-Driven Taxonomy for Antipsychotic Medication: A New Classification System.

Area of Science:

• Pharmacology
• Neuroscience
• Data Science

Background:

• Current antipsychotic classification (typical/atypical) is insufficient due to diverse pharmacological and clinical profiles.
• A need exists for a data-driven classification system that links receptor affinity to clinical effects.
• Receptor affinity data offers a foundation for a more nuanced antipsychotic classification.

Purpose of the Study:

• To develop a data-driven classification of antipsychotic medications based on receptor affinity.
• To evaluate if this classification can predict clinical effects of antipsychotics.
• To provide a framework for improved treatment guidance and drug development.

Main Methods:

• Analyzed in vitro receptor binding affinities of 27 antipsychotics across 42 receptors.
• Employed a clustering algorithm to group antipsychotics by receptor affinity.
• Utilized a machine learning model to predict clinical effects based on the derived groups.

Main Results:

• Identified four distinct antipsychotic groups based on receptor affinity.
• Group 1: Muscarinic antagonism (cholinergic/metabolic side effects).
• Group 2: Dopamine D2 partial agonism/adrenergic antagonism (low side effects).
• Group 3: Serotonergic/dopaminergic antagonism (moderate side effects).
• Group 4: Dopaminergic antagonism (EPS/hyperprolactinemia).
• Groups 1 and 4 showed higher efficacy than Groups 2 and 3.
• The classification accurately predicted out-of-sample clinical effects.

Conclusions:

• Receptor affinity-based antipsychotic grouping reflects pharmacology and predicts clinical differences.
• This classification can guide clinical treatment decisions for antipsychotic medications.
• The approach aids researchers in drug development and understanding antipsychotic profiles.