In Work Package 1, urban environmental exposure data from birth to 12 years were compiled and pre-processed, covering air pollution, natural spaces, built environment characteristics, traffic, meteorology, and social environment indicators. Data cleaning procedures included management of missing data, outlier detection, and standardisation of exposure variables. Multimorbidity indicators were developed based on seven health outcomes across cardiometabolic, respiratory/allergic, and neurodevelopmental domains. A multimorbidity risk score was constructed and clustering approaches were explored to identify patterns of disease co-occurrence. Machine learning methods were implemented to evaluate the predictive contribution of early-life urban environmental exposures to multimorbidity.
In Work Package 2, DNA methylation data underwent quality control and normalization procedures to remove low-quality samples and probes and to correct for technical variation and batch effects. The dataset was prepared for downstream analyses. Reproducible analytical pipelines were developed for epigenome-wide association studies and epigenetic clock analyses. Epigenetic research questions were also further refined and integrated into new project applications to support the continuation of research in this area.
In Work Package 3, analytical work focused on the assessment of associations between specific childhood urban environmental exposures and multimorbidity indicators and cardiometabolic health outcomes. Longitudinal trajectories of exposure to green and blue spaces were modelled, and annual and period-specific exposures to air pollutants were estimated. Regression-based approaches were applied to evaluate associations with multimorbidity risk score, multimorbidity clusters, and cardiometabolic health indicators.
