FLEX Method for Predicting Human Pharmacokinetic Clearance Published in “Pharmaceutical Research”

Fraction‑based Linear Extrapolation (FLEX) Method for Predicting Human Pharmacokinetic Clearance: Advanced Allometric Scaling Method and Machine Learning Approach

Published Date : September 10, 2025

Publication

Summary
Accurate prediction of human pharmacokinetic (PK) parameters, particularly clearance (CL), is critical for improving success rates in early-stage drug development. Conventional Single Species Scaling (SSS) methods, which use rat PK data and plasma protein binding fraction (fu,plasma), have been widely applied; however, validation using external datasets has been limited.
In this study, we generated a novel dataset through in-house experiments and developed the Fraction-based Linear Extrapolation SSS (FLEX-SSS) method, which optimizes dataset performance by applying an fu threshold. Importantly, this work represents the first validation of such an approach using an independent external dataset. Moreover, combining FLEX-SSS with machine learning demonstrated further improvements in predictive performance, suggesting its utility for more accurate human CL prediction in drug discovery.

Axcelead’s Solution
Axcelead provides seamless services from compound synthesis to in vitro and in vivo pharmacokinetic evaluations. Leveraging our extensive experience and expertise in drug discovery, we deliver accurate predictive models. Beyond experimental validation, we also support innovative solutions in applying machine learning during the drug discovery process. We welcome inquiries from partners seeking advanced strategies to overcome challenges in drug research and development.

Related Publication
・QSAR model to predict Kp,uu,brain with a small dataset, incorporating predicted values of related parameter (SAR QSAR Environ Res. 2022 Nov;33(11):885-897.)
・Development of a Novel In Silico Classification Model to Assess Reactive Metabolite Formation in the Cysteine Trapping Assay and Investigation of Important Substructures (Biomolecules. 2024 Apr 30;14(5):535.)


Yuki Umemori, M.S. (Pharmaceutics), Discovery DMPK
After completing his Master’s degree in Pharmaceutical Sciences at Kumamoto University, Umemori joined Teijin Pharma Ltd. in 2016 and has been affiliated with Axcelead Tokyo West Partners since 2024. His research spans early-stage compound screening to lead profiling, with expertise in pharmacokinetics. He develops machine learning–based prediction models tailored to specific research challenges, advancing drug discovery from both experimental (Wet) and computational (Dry) perspectives.