(A) Corrected P values [controlling the false discovery rate (FDR) at 0.05; dashed lines] from linear regression analyses using LFC (orange), ALT (purple), AST (green), or GGT (blue) as explanatory variables and plasma (upper panel) or urine metabolites (lower panel) as outcome. Results were separated by association direction: positive (β > 0) or negative (β < 0). Corresponding β estimates and FDR values are given in Supplemental Tables 1 and 2. Metabolites marked with a triangle exceeded the plotting range. (B) Box plots for the estimated proportion mediated between the exposure and metabolites by serum glucose, hsCRP, a measure of insulin resistance (HOMA-IR), and total TGs. (C) Color-coded FDR values from linear regression analyses using LFC, ALT, AST, or GGT as explanatory variables and plasma or urine metabolites as outcome, limited to nonmediated metabolites. Significant associations (FDR < 0.05) are framed black. Orange shading indicates positive associations, and blue shading indicates negative associations. Hatched boxes indicate mediation of the association by at least one of the following measures: insulin resistance (HOMA-IR), total TGs, hsCRP, or serum glucose. *Metabolites were annotated based on in silico prediction.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.