TOOL-01: AI-derived predictive biomarkers for chronic non-communicable diseases

Call goal: use established AI methods to develop and clinically validate multimodal biomarkers that predict disease progression and treatment response in chronic non-communicable diseases (cancer excluded). Proposals must show the clinical data is already available.

What we contribute:

·       Molecular biomarker candidates from genomic, transcriptomic and methylation data, combined with clinical markers.

·       Adaptation of existing Geneyx AI tooling to multimodal biomarker prediction.

·       Trustworthy-AI evaluation: performance metrics, bias assessment and external benchmarking, as in GAIC's benchmarking design.

·       PGx markers as predictors of treatment response.

·       Automated cohort-scale analysis, and decision-support visualization for clinicians.

·       Copy-number, structural and long-read-derived variants (including methylation) as additional molecular biomarker candidates.

·       API-based integration with partners' clinical data systems, so AI models run on new cohort data automatically.

·       ISO 13485 certification, CE-IVDR certification in progress and ISO/IEC 27001 security, with GDPR-aligned EEA hosting, supporting privacy-compliant AI and the regulatory strategy for qualifying biomarkers and AI tools.

Why it strengthens the consortium: the call favors adapting established AI over building new methods, and it weighs trustworthiness heavily. Geneyx brings AI tooling already deployed in a clinical decision support platform, plus hands-on experience with bias-aware design (GAIC deliberately leaves phenotype out of training to avoid bias toward common diseases) and external benchmarking. Adding genomic, epigenomic, transcriptomic and PGx layers makes the biomarkers multimodal, as required, and an industry partner supports the exploitation plan and the regulatory qualification route.