Biomedical research is at a turning point: With BioPathNet have teams of Helmholtz Munich and the Mila - Quebec Artificial Intelligence Institute a novel AI method developed that revolutionizes the analysis of massive biological data networks. The tool, whose details in Nature Biomedical Engineering were published, helps researchers to specifically search for hidden connections in gene, disease, and drug data.

Biomedical knowledge graphs that link genes, proteins, diseases, and therapies are indeed a valuable asset but are often incomplete. BioPathNet differs from previous methods by analyzing not just individual data points but entire cascades of connections – for example, from a gene through a signaling pathway to a disease and further to a potential drug.

This approach enables the AI to predict new, biologically plausible connections from thousands of such patterns. The particular advantages of BioPathNet lie in its Traceability: Each prediction can be explained via the identified "paths" in the data network, which allows for well-founded verification in the lab or in studies.

In tests, BioPathNet was able not only to identify known therapies for complex diseases such as leukemia and Alzheimer's, but also to suggest substances that are already being tested in clinical trials. The procedure is designed as a tool for generating hypotheses that accelerates biomedical research and contributes to the development of more comprehensive "base models" for drug discovery and understanding disease mechanisms.

Thanks to the interdisciplinary collaboration of Computational Biology, Mathematics, Biophysics, and Computer Science, BioPathNet is now available as Open-source solution available worldwide for researchers. It promises to optimize the use of existing knowledge and generate new ideas for experiments and therapies.