Drug-induced Liver Injury Prediction Based on Graph Convolutional Networks and Toxicogenomics
Tong Xiao, Ying Liu, Kaimiao Hu, Kaimin Guo, Mengying Zhang, TingTing Wang, Weihua Lei, Wenjia Wang, Shuiping Zhou, Yunhui Hu, Ran Su.
Abstract
Drug-induced liver injury is a leading cause of high attrition rates for both candidate drugs and marketed medications. Previous in silico models may not effectively utilize biological drug property information and often lack robust model validation. In this study, we developed a graph convolutional network embedded with a biological graph learning (BioGL) module—named BioGL-GCN(Biological Graph Learning-Graph Convolutional Network)—for drug-induced liver injury prediction using toxicogenomic profiles.
Introduction
Drug-induced liver injury (DILI) often leads to rejections of new drug applications, forces pharmaceutical companies to adjust dosing guidelines and issue medication warnings, and sometimes results in the withdrawal of drugs from the market. Between 1990 and 2010, 133 drugs meeting the inclusion/exclusion criteria were withdrawn from the market due to safety concerns, and 36 of these drugs (27.1%) were specifically recalled due to hepatotoxicity problems.
Materials and Methods:
Ting Li et al. curated a drug-induced transcriptome profiles dataset from the NIH LINCS L1000 dataset. They matched the Level 5 transcriptomic data from LINCS L1000 with the DILIst database using the PubChem Identifier service based on drug names and synonyms, obtaining 23,791 transcription profiles involving 69 cell lines. An improved Kennard-Stone algorithm was then used to extract transcription profiles with maximum explanatory variance.
Discussion
DILI is a critical safety consideration throughout the entire drug development process, encompassing all stages from preclinical to clinical studies. Driven by the rapid development of high-throughput technologies like the L1000 and the establishment of standardized frameworks for DILI classification, such as DILIst, many machine learning and deep learning-driven studies have made substantial advancements. In spite of that, we notice that within the gene networks associated with drugs, there might exist latent biological attribute correlations among the genes themselves.
Citation: Xiao T, Liu Y, Hu K, Guo K, Zhang M, Wang T, et al. (2025) Drug-induced liver injury prediction based on graph convolutional networks and toxicogenomics. PLoS Comput Biol 21(9): e1013423. https://doi.org/10.1371/journal.pcbi.1013423
Editor: Juilee Thakar, University of Rochester Medical Center, UNITED STATES OF AMERICA
Received: March 28, 2025; Accepted: August 11, 2025; Published: September 5, 2025.
Copyright: © 2025 Xiao et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: We have uploaded the data and code used in this study to GitHub, with the repository available at https://github.com/RanSuLab/BioGLGCN-DILI.
Funding: This study was supported by the National Natural Science Foundation of China (Grant No. 62222311 to RS). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
