Establishing individual metabolic health from clinical biochemistry, metabolomics and genome data for precision health
Approved Research ID: 71521
Approval date: September 21st 2021
The health impact of aging demographics within the EU and the UK leads to a dramatic increase in chronic diseases in the last decade. This is only accelerated by the current pandemic, leading to an increased burden on those of healthy and working-age to provide for the health and social care expenditures for a range of related services for the entire population. This catalyzes an important change, potentially a transition towards stratified prevention and digital care in the healthcare market.
The understanding of human health and disease is evolving from a symptom-based definition of independent diseases perspective towards defining personal health via quantifiable biomarkers and definition of development of (chronic) diseases on a continuum. The availability of medical health technologies, coupled with personal genomes has the potential to support Data-Driven Physicians in healthcare practice making timely and tailored decisions. In this project we aim to use data from UK Biobank as a reference dataset to complement our in-house I Am Frontier cohort data, for developing algorithms and methodologies that can recognize individuals before the onset of symptoms in chronic diseases, identify key risk factors and offer customized therapy options that are best matching their personal situation. The results of our study (after study completion in 36 months) will benefit clinicians in assisting them in their decision-making as well as citizens in understanding stratified risks and managing their health. It will also empower public health officials to review, stratified preventive health measures to be included within the coverage of national healthcare systems.
Healthcare systems in Europe are faced with an aging population and an increasing need to support people with chronic diseases (such as type II diabetes, cardiovascular diseases, and hypertension) in addition to constant pandemic threats, very much like those currently being experienced. In this project, we focus on establishing methodologies and metrics to determine metabolic function at the individual level and its link to chronic diseases. We would like to address this in different layers: First, by analyzing genome variant associations with molecular and clinical data that contributes to metabolic function and Biological Age (BA). Second, we would like to assess to what extent can metabolic function calculations be used for stratification and risk assessment of various chronic disease sub-populations? Third, we aim to develop a graph neural network model by utilizing the available patient outcome information for predicting chronic diseases with a metabolic component. Finally, we would like to use our in-house developed methodology to estimate personal reference intervals (PRI) for clinical biochemistry and metabolomics data and assess its potential for the prevention of chronic diseases.
Metabolomics data is within the original focus and data access application of the project however could not be provided previously (See Data Dictionary). To be able to realize our study aims, access to metabolic data will be crucial.
We would like to expand the focus of the project with WGS and Genotyping arrays (See reasoning below) and the OLINK proteomics data that will be available later this year.
We aim to complement genetic information with the proteomics data in calculating pQTLs for relevant cases. Furthermore, we are going to calculate inter-individual distances in the context (chronic) diseases at the clinical & protein interaction network level. We plan to perform dimensionality reduction to represent a multidimensional dataset in lower-dimensional networks where each node is an individual with a select number of parameters. This allows us to develop insights into how individuals are similar or different from each other, allowing us to explore potential personal health trajectories leading to disease profiles.