Jia Xuelian, Teutonico Donato, Dhakal Saroj, Psarellis Yorgos M, Abos Alexandra, Zhu Hao, Mavroudis Panteleimon D, Pillai Nikhil
Center for Biomedical Informatics and Genomics, Tulane University, New Orleans, Louisiana 70112, United States.
Department of Chemistry and Biochemistry, Rowan University, Glassboro, New Jersey 08028, United States.
J Med Chem. 2025 Apr 10;68(7):7737-7750. doi: 10.1021/acs.jmedchem.5c00340. Epub 2025 Mar 27.
Accurate prediction of new compounds' pharmacokinetic (PK) profile in humans is crucial for drug discovery. Traditional methods, including allometric scaling and mechanistic modeling, rely on parameters from or testing, which are labor-intensive and involve ethical concerns. This study leverages machine learning (ML) to overcome these limitations by developing data-driven models. We compiled a large data set of small molecules' physicochemical and PK properties from public sources and digitized human plasma concentration-time profiles for approximately 800 compounds from the literature. We introduced a hybrid modeling framework that combines ML with physiologically based pharmacokinetic modeling and a hierarchical ML framework that employs two steps of learning to directly estimate PK profiles. Tested on 106 drugs, these frameworks demonstrated prediction accuracies within a 2-fold and 5-fold error for 40-60% and 80%-90% of compounds, respectively, in both AUC and . Proposed approaches could enhance early molecular screening and design, advancing drug discovery capabilities.
准确预测新化合物在人体中的药代动力学（PK）特征对于药物研发至关重要。传统方法，包括异速生长缩放和机理建模，依赖于来自体内或体外测试的参数，这些方法劳动强度大且涉及伦理问题。本研究利用机器学习（ML）通过开发数据驱动模型来克服这些限制。我们从公共来源汇编了一个小分子物理化学和PK性质的大数据集，并将文献中约800种化合物的人血浆浓度-时间曲线数字化。我们引入了一个将ML与基于生理的药代动力学建模相结合的混合建模框架，以及一个采用两步学习来直接估计PK特征的分层ML框架。在106种药物上进行测试时，这些框架在AUC和峰浓度方面分别对40%-60%和80%-90%的化合物显示出2倍和5倍误差范围内的预测准确性。所提出的方法可以加强早期分子筛选和设计，提高药物研发能力。
