Figure 3.

The effect of age on protein expression levels. (a) A heatmap visualizing the average age-associated changes of the 122 proteins with age effect. Child-specific Z-scores were first calculated for each protein at all timepoints based on label-free proteomics data, and the averages of these Z-scores across all children were plotted in the heatmap. Example visualizations of (b) age-associated GP models and (c) corresponding measured protein intensities (in log₂ scale) for selected proteins present in the six main clusters. Each line represents one child, red = girls, blue = boys. (d) Covariate contributions for PI16. Estimated covariate contributions shown on the top row are calculated as the posterior mean predictions of the corresponding components in the age model. The bottom row shows the cumulative contributions these additive components. SERPIND1 = Heparin cofactor 2, PLG = Plasminogen, IGFALS = Insulin-like growth factor-binding protein complex acid labile subunit, VWF = von Willebrand factor, MASP2 = Mannan-binding lectin serine protease 2.