14
Partner research institutions
3
Use cases in rare diseases
8
European countries involved
Why we do it
Sensitive health data cannot be centralised, but research depends on scale.
GDPR and national regulations prohibit centralized sharing of sensitive health data across institutions and borders. At the same time, research questions about rare diseases require large sample sizes.
The solution is fundamental gap how to enable cross-border health research studies while maintaining the promise of GDPR.
The better approach
The Personal Health Train: bringing computation to data
Research algorithm defined
A federated algorithm is developed on researcher workstations.
Train dispatched to Stations
The algorithm is dispatched to participating institutions.
Local computation at each station
Computations run on local infrastructure on FAIR-prepared data.
Aggregated results returned
Only privacy-preserving aggregated results leave institutions.
Use cases at launch
Three clinical research domains
Three diverse areas selected for their unmet therapeutic needs and real world healthcare relevance, building upon existing European expertise.
Paediatric Intellectual Disability
Use case 1
Integrating genomic and phenotypic data from diverse European cohorts to identify shared pathways and potential therapeutic targets.
View use caseInherited Retinal Dystrophies
Use case 2
Combining clinical and genomic information across rare disease registries and biobanks to characterize genotype-phenotype correlations.
View use caseAutism Spectrum Disorders
Use case 3
Multi-modal data integration — clinical, genomic, metabolic and environmental — identifies predictive risk factors.
View use case
Technology & standards
Built on open standards and FAIR principles
Federated Learning
Train AI models across institutions without moving patient data.
FAIR Data
Findable, Accessible, Interoperable and Reusable health datasets.
PADME & Vantage6
Platforms for Analytics and Distributed Machine Learning for Enterprises.
Vantage6
Open-source Personal Health Train infrastructure for federated analysis.
GDPR Compliance
Privacy-by-design architecture aligned with European data regulation.
AI / ML Methods
Federated training of clinical, genomic and multi-modal models.
Consortium
14 organisations across 8 countries
A pan-European consortium uniting academic research institutions, clinical centres, technology SMEs and a coordinating AI company.
