Establishing extensive biobanks like the UK Biobank has opened new frontiers for disease prediction and biomarker discovery by integrating genomic, proteomic, metabolomic, and clinical data. At the newly founded Genomic and Proteomic Centre at MBRU, we leverage cutting-edge methodologies to validate and enhance predictive models for various diseases.
Building on innovative frameworks like MILTON, which successfully utilized multi-omics data to improve disease prediction, our focus is to integrate these insights with local validation datasets. This effort aims to refine predictive accuracy and identify novel biomarkers. Our project objectives are as follows:
1- Apply machine learning models to UK Biobank’s multi-omics datasets (genomics and proteomics) to predict disease risk across diverse phenotypes.
2- Discover novel biomarkers and combinations that surpass polygenic risk scores in disease detection.
3- Validate and optimize predictive models using local genomic and proteomic data from our centre at MBRU.
We aim to complete this project within three years of gaining access to the UK Biobank data. Through this research, we will explore biomarker diversity across populations, expecting to identify population-specific biomarkers in our local cohort that differ from those in the UK Biobank. Integrating and validating these models with local data will enable us to assess biomarker reliability, uncover new associations, and examine how diversity influences disease prediction across populations.
This comparison is essential for tailoring predictive algorithms to diverse groups, ultimately advancing precision medicine and ensuring broader applicability of these models for global healthcare challenges.