Multi-model approach for discovery of associations between gene variants and diseases.
- Mentors
- Tim Jackson
- Organization
- Global Alliance for Genomics and Health
- Technologies
- python, postgresql, tensorflow
- Topics
- machine learning, Computational Genomics
To identify possible relations between gene variants, phenotypes, and diseases - an ideal way would be a multi-model approach that involves:
• Supervised learning methods to discover diseases associated with combinations of gene variants and clinical features.
• Also unsupervised learning methods like clustering to understand the underlying patterns among diseases that went unnoticed before.
This could be a strong evidence-based approach for clinical treatments.