Fig. 4: Signatures of selection in urban populations.

a Manhattan plots showing the distribution of cross-population nSL scores (XP-nSL) for each urban–rural population pair in 200 kb sliding windows with a 50 kb step size. Positive scores depict selection in urban populations. Significant autosomal outlier windows are highlighted in red and outlier windows on the Z-chromosome in blue (see “Methods”). Autosomal and sex-chromosome outlier windows were detected separately. b Violin plots showing the distribution of correlation coefficients (Pearson’s r) for the different inter-population summary statistics for all pairwise comparisons. Note that all distributions overlap zero and are not consistent across pairs. The shaded grey areas of the violin plot denote the kernel density estimation of the data distribution. The bottom and top edges of the white boxplot denote the upper and lower interquartile range (25th and 75th) with the vertical black line showing the median, and the thin horizontal line representing the rest of the distribution. Black dots represent values outside the distribution. c Violin plots for the distribution of correlation coefficients between inter-population summary statistics with proxies for recombination rate (within population LD, GC-content). The boxplot shows the interquartile range (white area), median (vertical black line) and the remaining distribution (horizontal black line). d Example signatures of divergence and proxies for recombination along chromosome 1 and 6, to highlight the heterogeneity of divergence across populations. Window-based measures were loess-smoothed across each chromosome. Populations with selective sweeps on the respective chromosomes are highlighted, with Glasgow (blue) on chromosome 1 and Lisbon (orange) on chromosome 6. All other populations are in grey in the background. Note that the highest sweep window, marked by the dotted line, corresponds to a likely low-recombination region on chromosome 6, but a normal recombination region on chromosome 1. e Non-significant correlation of the geographic distance between urban and rural populations from the same locality and their genetic differentiation. (linear model: F_1,7 = 0.23, P = 0.645). Shaded area denotes the 95% CI. BCN Barcelona, GLA Glasgow, GOT Gothenburg, LIS Lisbon, MAD Madrid, MAL Malmö, MIL Milan, MUC Munich, PAR Paris. Source data are provided as a Source data file.