Dendrimer Pharmacokinetics Prediction (dendPoint)

Dendrimer Pharmacokinetics Prediction (dendPoint) predicts intravenous pharmacokinetics of dendrimers from their physicochemical and structural properties using machine learning.

Key Features:

• In silico pharmacokinetic model: Predicts intravenous pharmacokinetic parameters of dendrimers from physicochemical and structural properties without requiring animal data.
• Machine learning-based approach: Employs machine learning models to relate dendrimer structure and properties to pharmacokinetic outcomes.
• Manually curated relational database: Built on a comprehensive, manually curated relational database containing the largest collection of dendrimer pharmacokinetic parameters and corresponding structural and physicochemical properties.
• Predictive performance: Demonstrates correlation coefficients (r) up to 0.83 and cross-validated Q² values up to 0.68 for predicting half-life, clearance, volume of distribution, and dose recovered in the liver and urine.

Scientific Applications:

• Nanotherapeutic design guidance: Guides the design and refinement of dendrimer constructs prior to resource-intensive in vivo testing.
• Pharmacokinetic forecasting: Forecasts key intravenous PK parameters (half-life, clearance, volume of distribution, liver and urine dose recovery) for dendrimer formulations.
• Reduction of animal studies: Supports reduction of reliance on animal experiments by providing in silico PK predictions for dendrimers.

Methodology:

Machine learning models trained on a manually curated relational database of dendrimer structural and physicochemical properties and pharmacokinetic parameters; model performance evaluated using correlation coefficient (r) and cross-validated Q² (reported up to r = 0.83 and Q² = 0.68).