Fig. 6: Impact of reporting error on SNP effects and trait heritability.

a, Explained variance (β_std²) per SNP reaching genome-wide significance in error-corrected GWA analyses (y axis; phenotype obtained using means across multiple measurement occasions) or error-uncorrected GWA analyses (x axis; phenotype obtained from a single baseline measure). The colour scheme highlights in which GWA the genetic variant was identified, including error-corrected GWA (in red), error-uncorrected GWA (in green) or both (blue). b, The y axis shows the differences in SNP heritability estimates obtained from error-corrected GWA analyses and error-uncorrected GWA analyses (h²_diff = h²_{repeated measure} − h²_{single measure}). The x axis gives the degree of repeatability per phenotype, estimates as the variance (R²) explained by models regressing P_T2 on P_T1, while controlling for follow-up time (time_{T2 − T1}) and age. The sample size per phenotype is included in Supplementary Table 8. All error bars shown in the figure represent the 95% CIs.