Fig. 1: Predicted change in phenotypic means and disease prevalence among the edited genomes.

Left, common diseases. Right, quantitative biomarkers. For each trait, a list of published GWS loci was taken. For the left panel: AD⁶⁸, MDD⁶⁹, SCZ⁷⁰, T2D⁷¹ and CAD⁷². For the right panel: FG⁷³, LDL cholesterol⁷⁴, TG⁷⁴, SBP⁷⁵ and DBP⁷⁵. The GWS loci were ordered by the product of the estimated effect size and the frequency of the undesirable allele (that is, decreasing effects for disease and biomarkers, up to a maximum of ten loci) (Methods). The x axis represents the ordered number of edited loci. The y axis represents the predicted phenotypic change among edited genomes compared to the mean of unedited genomes. For disease, the predicted change is expressed as a fold change in lifetime prevalence. For the quantitative traits, the predicted change is in phenotypic standard deviations (s.d.). In both panels, the dotted lines correspond to standard deviations below or above the predicted changes. We calculated the predicted s.d. of gain on the liability scale as the square root of the expected variance explained by the edited loci in the general population. Expected changes of one s.d. above/below the predicted change were converted on a disease risk scale using a probit transformation. Data underlying this figure are given in Supplementary Table 1. The source code used to generate the figure is provided in ‘Code availability’.