Daniele Bizzarri | 11-09-2024 | Metabolomics- and methylomics-based predictors for estimating health and biological aging
Over the past decade, the abundance of high-throughput omics approaches coupled with the use of machine learning techniques, has made it possible to investigate the full molecular complexity of health and aging. The primary forcus of this thesis was to study and improve biological aging prediction. To achieve this we developed, evaluated, and deployed state-of-the-art models predicting different aspects of human health risks by employing multiple omics measurement, with a particular attention given to 1H-NMR metabolomics. Availability, affordability, interpretability, and robustness of the 1H-NMR metabolomics platform by Nightingale Health makes it a powerful tool with implications in the risk prediction of common diseases. We explored this research line in epidemiological settings within the BBMRI-nl consortium, which incorporates 28 cohorts with various specific characteristics. Hence, we took advantage of the wide range of health statuses when examining the extensive BBMRI datasets, investigated specific subgroups such as elderly or night-working individuals respectively recruited for the Leiden Longevity Study (LLS) and LIFELINES, and even explored the potential complementarity and interaction of different omics (e.g., 1H-NMR metabolomics, DNA methylome) available within the subset known as BIOS Consortium.