Genomic Analysis

Genomic data is analyzed to connect variants, phenotypes, and disease patterns for diagnosis and treatment development.

Business impact

• Prediction accuracy — The DTU National Food Institute model achieved up to 97% accuracy in predicting resistance.
• Diagnosis rate — The hospital identifies genetic variants in 30% of its patients, enabling accurate diagnoses and tailored treatment plans.

Data requirements

• Genomic data (tabular) — The systems use genomes from pediatric patients and from more than 1,600 Listeria monocytogenes strains.
• Clinical data (tabular) — The hospital integrates patient records, symptoms, phenotypes, and related biomedical information.
• Multiomic data (tabular) — The hospital combines genomes, proteomes, and transcriptomes in its knowledge graph.
• Phenotype data (tabular) — Deep phenotyping links observed disease manifestations to genetic variants.

AI methods and techniques

• Graph queries — Researchers use Cypher to uncover relationships and patterns in the clinical knowledge graph.
• Graph machine learning — Algorithms establish causal links between data points and support prediction work on connected biomedical data.
• Federated machine learning — Models learn from partner data without sharing live patient data.
• Machine learning — A model is trained on bacterial genomes to predict resistance to disinfectants.

Industries

Real-world evidence

3 documented case studies on record.

Companies using this: Anthropic, Dr. von Hauner Children's Hospital at Ludwig Maximilian University of Munich (LMU), DTU National Food Institute.