Fig. 1: Overview of study design.
From: Personal, social, and natural co-exposure pattern and plasma proteins in cardiometabolic diseases
We assessed the contributions of individual, social, and natural environmental exposures to cardiometabolic diseases using Cox regression and XGBoost model. Participants were then classified into 5 distinct co-exposure pattern subgroups via K-prototypes clustering, and the effects of these patterns were estimated using Cox regression. Leveraging high-dimensional proteomic data, we analyzed 2,911 proteins to identify proteomic signatures associated with each co-exposure pattern using LASSO regression. Additionally, mediation analysis was applied to pinpoint key proteins that mediate the relationship between these exposure patterns and CMD outcomes. CMD cardiometabolic disease; PM particulate matter; IMD Index of Multiple Deprivation; PA physical activity. Created in BioRender. Xw, T. (https://BioRender.com/ebkarjr).
