truPK -- human pharmacokinetic models for quantitative ADME prediction

Area of Science:

• Pharmacology
• Computational Chemistry
• Drug Discovery

Background:

• In silico prediction of Absorption, Distribution, Metabolism, and Excretion (ADME) properties is crucial for early drug candidate assessment.
• Current in silico ADME and toxicity models have limited predictive power for clinical outcomes.
• Existing tools often focus on physicochemical properties or provide qualitative classifications, lacking quantitative human pharmacokinetic (PK) predictions.

Purpose of the Study:

• To develop a reliable in silico tool for predicting human pharmacokinetic (PK) properties.
• To address the limitations of existing methods in predicting clinical drug behavior.
• To provide quantitative predictions of key molecular properties influencing human dosage and frequency.

Main Methods:

• Development of truPK, a novel in silico tool by Strand Genomics.
• Utilizing sophisticated machine learning methods to build five predictive models.
• Validation of models using external datasets to assess accuracy.

Main Results:

• The truPK tool accurately predicts human pharmacokinetic properties.
• Models demonstrated greater than 75% accuracy in external validation sets.
• Predictions include bioavailability, protein binding, volume of distribution, elimination half-life, and absorption rate.

Conclusions:

• The truPK tool offers a significant advancement in predicting human PK properties.
• Accurate in silico PK prediction can reduce drug candidate failures and optimize dosing strategies.
• This tool enhances the reliability of early-stage drug discovery assessments.