Graph based models compared for chemical toxicity prediction in drug discovery
Benchmarking graph-based models for in-silico toxicity prediction in drug discovery
Machine Learning
Summary
Predicting if a new drug will be toxic is really important but hard to do early on. The authors studied many computer models that look at drugs as graphs to see which ones best predict toxicity. They tested over 20 different models fairly and carefully using the same data and methods. Their work helps clarify which graph-based approaches work well and which do not. They also shared their testing tools publicly so others can check and improve research.
What this means in practice
- •For pharmaceutical data scientists: Evaluate and select graph-based models more reliably for early toxicity screening in drug candidate pipelines using a standard benchmarking framework.
- •For chemical safety analysts: Use consistent graph learning models to predict toxic effects of chemicals before expensive lab testing to improve safety prioritization.
Authors
Noel Suarez-Barro, Manuel Lama, Juan C. Vidal
Abstract
Drug discovery is a costly and high-risk process, where toxicity-related failures remain a major cause of attrition in both preclinical and clinical stages. As a result, accurate early prediction of chemical toxicity is essential to reduce downstream costs and improve compound prioritization. In this context, graph deep learning (GDL) has emerged as a powerful paradigm for toxicity prediction, leveraging molecular graph representations to learn directly from chemical structure with improved expressivity over traditional approaches. Despite the growing number of proposed models, current literature-based comparisons are often difficult to interpret due to inconsistencies in datasets, preprocessing pipelines, and evaluation protocols. To address this limitation, we introduce a unified and standardized benchmarking framework for GDL-based toxicity prediction. We systematically evaluate more than 20 representative approaches under consistent experimental conditions and across multiple datasets and partitioning strategies, enabling a fair and reproducible comparison of model performance. In addition, we complement this empirical study with a structured literature analysis to contextualize existing methodological trends and performance claims. Our results provide a clearer and more reliable assessment of the current state of the field, highlighting both the strengths and limitations of existing graph-based approaches. To support transparency and reproducibility, we release our benchmarking framework as open-source software https://gitlab.citius.gal/noel.suarez/benchtox, allowing the community to evaluate and compare models under consistent conditions.