Fig. 2

Variance in circulating inflammatory protein levels explained by common genetic variation. a In this panel, the variance explained (r²) by consensus SNPs (n = 13 SNP, 1 per protein) in the ordinary least squares regression model was compared against the variance explained by the same SNP set identified in the Bayesian penalised regression approach. b The proportion of variance explained in Olink® inflammatory protein levels by common genetic variants genotyped in the LBC1936 participants is shown. Only those proteins which had significant pQTL associations in both the ordinary least squares and Bayesian methods are presented (n = 13). Additionally, the proportion of variance explained attributable to medium effects (prior: variance of 1% explained) and large effects (prior: variance of 10% explained) are demonstrated in purple and green, respectively. Error bars represent 95% credible intervals