Breakthrough biomarker discovery for dementia
Dementia is a growing global challenge, with 40 million individuals affected worldwide, a number projected to double every 20 years. In Europe alone, 8.8 million people live with the condition. Despite its profound personal and societal impacts, dementia remains difficult to diagnose in its early stages, as clinical symptoms typically emerge years after the onset of pathological changes, and there is considerable overlap in clinical symptoms between different causes of dementia.
Integrative approach towards biomarker identification
Undertaken with the support of the Marie Skłodowska-Curie Actions(opens in new window) programme, the MIRIADE(opens in new window) project aimed to accelerate the discovery of novel biomarkers for dementia(opens in new window). “We established a unique research and training programme to enable precision health approaches for early diagnosis, disease monitoring and better treatment outcomes,” explains project coordinator Charlotte Teunissen from Amsterdam University Medical Centre(opens in new window) (UMC).
Breakthroughs with machine learning
A central achievement of the MIRIADE project lies in its integrative approach to data analysis and biomarker identification. The consortium unified omics data(opens in new window) from cerebrospinal fluid (CSF) and blood samples and forwarded over 30 candidate biomarkers for development. It combined artificial intelligence (AI), big data analysis, cutting-edge assay technologies and entrepreneurial skills training, advancing both scientific discovery and clinical application. The AI tools developed were designed to support the project’s PhD students and researchers in narrowing down which proteins out of thousands were worth investigating. They drew on existing datasets, literature reviews and experimental data collected within the project itself. In Luxembourg, for instance, researchers applied refined statistical models and text mining techniques. This multilayered data approach enabled machine learning models to predict which proteins were not only biologically relevant but also feasible to measure in CSF or blood. Another powerful use of AI was in evaluating whether a given protein could be detected using antibodies – reagents that bind to target proteins. “Normally there is a bit of trial and error in this process,” says Teunissen. “But the AI models helped predict whether an antibody-protein interaction was likely to work before we even tested it in the lab.” For three of the project’s 15 PhD students, AI wasn’t just a method of investigation, it was the main focus of their research. The benefits of this have extended well beyond the scope of MIRIADE. These models are now informing new projects across neurodegenerative research, offering a smarter way to target biomarkers that work in real-world clinical settings.
Biomarker validation and assay development
Thanks to this big data approach, remarkable progress was made in biomarker assay development and clinical validation. Using advanced protein technologies, MIRIADE successfully developed and clinically validated assays for 16 biomarkers. Among the prioritised biomarkers were dopamine decarboxylase for dementia, aquaporin-4 for Alzheimer’s disease and synaptic biomarkers(opens in new window) SNAP-25 and VAMP-2 in blood for frontotemporal dementia and Alzheimer’s disease.
Translating findings into clinical practice
The translation of MIRIADE’s findings into clinical and diagnostic use is now well under way. Biomarkers such as dopamine decarboxylase are being implemented in initial clinical trials, and plasma pTau assays are undergoing prospective evaluation. For other biomarkers, retrospective and prospective validation studies are planned to refine their clinical application. Furthermore, the project has created a comprehensive roadmap to guide future biomarker development. This framework includes best practices for integrating AI, big data, assay development and clinical validation, as well as approaches for fostering stakeholder collaboration to accelerate the translation of biomarkers into practice. Looking ahead, the project will continue its work by refining biomarker validation, scaling up assay production and seeking funding to advance its discoveries. MIRIADE is shaping a future where timely diagnosis and effective treatment significantly improve patient outcomes and reduce the socio-economic burden of dementia worldwide.
