Actionable predictions of human pharmacokinetics at the drug design stage

06 March 2024, Version 1
This content is a preprint and has not undergone peer review at the time of posting.


We present a novel computational approach for predicting human pharmacokinetics (PK) that addresses the challenges of early-stage drug design. Our study introduces and describes a large-scale dataset of 11 clinical PK endpoints, encompassing over 2700 unique chemical structures to train machine learning models. To that end multiple advanced training strategies are compared, including the integration of in vitro data and a novel self-supervised pre-training task. In addition to the predictions, our final model provides meaningful epistemic uncertainties for every data point. This allows us to successfully identify regions of exceptional predictive performance, with an Absolute Average Fold Error (AAFE/GMFE) of less than 2.5 across multiple endpoints. These advancements represent a significant leap towards actionable PK predictions, which can be utilized early on in the drug design process to expedite development and reduce reliance on nonclinical studies.


machine learning
graph neural networks
drug design
uncertainty quantification
clinical pharmacokinetics
clinical pharmacology
epistemic uncertainty

Supplementary materials

Supporting Tables & Figures
PDF file with additional descriptors and of the human PK data set per source, tables containing all discussed metrics from the cross-validation experiments, additional descriptors and metrics for the test set predictions, and an overview of the GPS architecture and hyperparameters.


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.