Fig. 2: Comparisons of significant genomic loci for 40 circulating cytokines across three proteomics assays.

a Number of reproducible and non-reproducible SNPs per cytokine (depicted as saturated and light-colored bars, respectively) for deCODE, SCALLOP, and YFS & FINRISK cohorts. The saturated portion of the bars represents the number of SNPs that were replicated in at least one other cohort, where reproducibility is defined as SNPs confined to significant loci (p-value < 0.05) and directionally concordant. b Proportion of replicated SNPs across cytokine datasets visualized for all possible combinations of cohorts. The matrix at the bottom left shows the comparison each of the vertical bars at the top represents. Arrows in the comparison matrix illustrate the direction of comparison—from the reference dataset where significant SNPs were identified, to the dataset in which SNPs were replicated. For example, the first bar displays the percentage of loci in deCODE that were replicated in SCALLOP (with the arrow pointing from deCODE to SCALLOP). The horizontal bars on the bottom right show the number of cytokines for which significant SNPs were found. The sample sizes used to derive the statistics in b are from left to right 73, 141, 46, 79, 101, and 70. Median, IQR (error bars represent the 25th and 75th percentiles). Colored bars represent the deCODE consortium in red, the SCALLOP consortium in blue and YFS & FINRISK cohorts in yellow. IQR inter-quartile range, SNP single-nucleotide polymorphism, SCALLOP Systematic and Combined AnaLysis of Olink Proteins, YFS & FINRISK Cardiovascular Risk in Young Finns Study.