Enhancing Drug Repositioning: A Multi-class Ensemble Model for Drug-target Interaction Prediction with Action Type Categorization
Leila Jafari Khouzani, Soroush Sardari, Soheila Jafari Khouzani, Horacio Pérez-Sánchez, Fahimeh Ghasemi
Abstract
Accurate prediction of drug–target interactions (DTIs) is critical for accelerating drug repositioning and reducing the cost of pharmaceutical development. Most existing studies frame DTI prediction as a binary task and often neglect the pharmacological action types and the quality of non-interaction data. This study introduces a multi-class classification framework that categorizes interactions into activators, inhibitors, and non-action classes. A novel zero-interaction selection algorithm is proposed, based on weighted drug–drug and protein–protein similarity scores, to improve dataset diversity and reliability.
Introduction
Despite significant investments and advancements, developing FDA-approved treatments still takes over 12 years and costs an average of $1.8 billion. In this context, drug repositioning—identifying new uses for existing drugs—has emerged as a key strategy in the pharmaceutical industry. This approach offers a faster and more cost-effective pathway to address undetected medical needs and expand treatment options.
Materials and Methods:
Building upon the outputs of the data acquisition process, the following steps were undertaken to create the dataset. DrugBank was chosen as the primary data source due to its comprehensive coverage of approved drugs and associated information. UniProt provided the necessary protein-related data, while PubChem was utilized to fill in any missing drug details.
The dataset construction began by extracting the general characteristics of 15,498 compounds listed in DrugBank as of August 1, 2020. These compounds were either FDA-approved or classified as nutritional. Key attributes such as type, group, molecular weight, and chemical structure were obtained.
Discussion:
This study introduces a robust framework for multi-class DTI prediction with action-type categorization, combining feature engineering, zero sample design, and classifier benchmarking. The proposed method enhances DTI prediction by addressing common challenges such as class imbalance, lack of true zero labels, and high feature dimensionality.
Citation: Jafari Khouzani L, Sardari S, Jafari Khouzani S, Pérez-Sánchez H, Ghasemi F (2025) Enhancing drug repositioning: A multi-class ensemble model for drug-target interaction prediction with action type categorization. PLoS One 20(12): e0333553. https://doi.org/10.1371/journal.pone.0333553
Editor: Minh Le, Montefiore Medical Center, UNITED STATES OF AMERICA
Received: March 19, 2025; Accepted: September 13, 2025; Published: December 15, 2025.
Copyright: © 2025 Jafari Khouzani 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: All relevant data for this study are publicly available from the GitHub repository (https://github.com/ljafari/A-Multi-Class-Ensemble-Model-for-Drug-Target-Interaction-Prediction-with-Action-Type-Categorization/).
Funding: This study was supported by the Office of Vice Chancellor for Research of the Isfahan University of Medical Sciences (Grant No. 3991125 to FG). The total funding amounted to approximately 280, all of which was allocated to computational costs.
Competing interests: The authors have declared that no competing interests exist.
