Additional file 2: Figure S2. Peptide array fluorescent intensity is highly reproducible across reagent lots, independent cohorts, and assay batch. (A) To estimate reagent lot and assay batch impact, 66 donors from the Discovery Cohort were selected for repeat assay using arrays from an independent manufacturing synthesis batch. The initial quantification (x-axis) is shown for each probe (dot) for all donors (single scatter plot) compared to re-assay on an independent reagent lot (y-axis). Values shown are normalized intra-array by median and inter-array by probe mean. This normalization decreases artifactual correlation as a result of probe absolute quantification, which tends to be similar in many conditions. Thus, each scatter plot shows the donor-specific residuals, which are highly correlated in the initial and repeated quantification. (B,C) Age (B) and BMI (C) statistics are reproducible across multiple independent cohorts. Age is highly reproducible with nearly exact same large effect size for each probe across cohorts. BMI is reproducible, but with smaller overall effect size and increased variation in BMI log10 fold change. Axes show average log10 ratio and each data point is a peptide probe on the array. (D) Probe correlations and association with age are reproducible across reagent lot and assay batch. Donors are the 66 selected for technical replication (described in main text), probes are same as shown in main-text heatmap, axis clustering uses new technical replicate data using independent reagent and array synthesis lots.