PRECISION
Explainable graph convolutional networks for innovation in precision oncology and drug resistance in cancer
The PRECISION project represents a significant advancement in research on triple-negative breast cancer (TNBC), a particularly aggressive subtype known for its challenging prognosis and limited treatment options. The project aims to integrate precision medicine with cutting-edge Artificial Intelligence (AI) methodologies to improve diagnosis and personalized treatment.
The project focuses on developing and applying various AI methods, Graph Convolutional Networks (GNN)—specifically Kolmogorov-Arnold Graph Networks (KAGNN)—and Explainable AI (XAI). These tools allow for the modeling and interpretation of complex biological interactions at the genomic level, transforming “black box” models into transparent systems that build trust among clinicians.
Main Objectives
- Objective 1: Integrating prior knowledge for explainable novel insights. Focuses on integrating biological components using KAGNN and XAI to identify novel therapeutic targets and drug repurposing opportunities for TNBC.
- Objective 2: Enhancing synergy between medical expertise and AI. approach to foster continual bidirectional human-machine interaction, ensuring clinical insights are integral to model development.
- Objective 3: Computational mapping of tumor microenvironment interactions. Explores cellular heterogeneity and diversity within tumors using both bulk and single-cell sequencing technologies to enhance targeted therapy development.
- Objective 4: Mining patterns of drug resistance mechanisms. Identifies graphical patterns associated with resistance by integrating transcriptomic and drug resistance data (DRMref) with machine learning models.
- Objective 5: Bridging Theory and Practice. Aims to transition from abstract Kolmogorov-Arnold theoretical representations to practical applications by exploiting the structural advantages of GNNs.
Project Details
| Full Title | Explainable graph convolutional networks for innovation in precision oncology and drug resistance in cancer (PRECISION) |
| Identifier | PID2024-162441OA-I00 |
| Total Funding | 118,500 € |
| Funding Agency | Ministerio de Ciencia, Innovación y Universidades. Agencia Estatal de Investigación. EDRF/EU |
| Programme | Programa Estatal para la Investigación y el Desarrollo Experimental |
| Subrogramme | Subprograma Estatal de Generación de Conocimiento Científico-Técnico y Desarrollo Experimental |
| Duration | 01/09/2025 - 31/08/2028 |
Research Team
The project is led by The Machine Learning Laboratory in Life Sciences Lab. The multidisciplinary team comprises computer scientists, bioinformaticians, biologists, and clinical oncologists. Notable collaborators include experts from the Breast Unit at Hospital Universitario Lucus Augusti (HULA) and the University of Cambridge.
Grant PID2024-162441OA-I00 funded by MICIU/AEI/ 10.13039/501100011033 and by “ERDF/EU”.
