Graph-Based Explainable AI for Next-Generation Medicine: Researchers from Grupo de Aplicación de Telecomunicaciones Visuales (GATV) Publish a Breakthrough in Drug Repurposing in Nature Journal

Bringing a new drug to the market from scratch is a complex journey that can cost billions of euros and span over a decade of research. Facing this challenge, drug repurposing—finding new therapeutic uses for existing, safe medications—has emerged as a crucial strategy for healthcare sustainability worldwide. However, the current hurdle lies in predicting accurately and comprehensibly which existing drugs will effectively treat new pathologies.

At the Grupo de Aplicación de Telecomunicaciones Visuales (GATV) at UPM, we have achieved a major scientific milestone in this domain. Our researchers Pablo Perdomo-Quinteiro and Alberto Belmonte-Hernández, alongside co-author Emre Guney, have developed an innovative methodology based on Graph Neural Networks (GNN) for high-performance Explainable Artificial Intelligence (XAI). Their scientific study has been recently published in the prestigious journal Nature (Scientific Reports) under the title “Generating explainable hypotheses for drug repurposing with graph neural networks”.

This outstanding research has been carried out within the framework of the European project REPO4EU, an EU-funded initiative aiming to establish a comprehensive European platform for mechanism-based drug repurposing.

What Does This Research Deliver?

The core strength of the methodology designed at the Grupo de Aplicación de Telecomunicaciones Visuales (GATV) lies in successfully coupling two factors that rarely coexist in traditional machine learning models:

  1. High Predictive Performance through Graphs: Utilizing Graph Neural Networks (GNNs) allows advanced mapping of heavily interconnected biomedical networks, processing massive datasets to optimally prioritize the best therapeutic candidates.
  2. Mechanism-Based Explainability (XAI): Unlike conventional “black-box” AI, this model opens up the decision-making process, providing human-readable biological hypotheses explaining why and how a specific drug interacts at a molecular level with a target disease.

Impact and Technology Transfer to the Pharmaceutical Industry

The industrial application of this technological development is immediate. By delivering transparent, mechanism-driven explanations, it provides vital strategic support for R&D departments in pharmaceutical and biotech companies. This significantly facilitates critical decision-making prior to clinical trials, lowers financial risks, and safely accelerates precision medicine workflows.

With this breakthrough, the Grupo de Aplicación de Telecomunicaciones Visuales (GATV) reinforces its strategic commitment to leveraging cutting-edge AI research for high-impact social and commercial solutions, directly driving the sustainability and efficiency of tomorrow’s global healthcare systems.

🔗 Access the Full Scientific Article (Open Access): You can read the original paper published in Nature directly through this link: https://www.nature.com/articles/s41598-026-50149-2