PMC

(a) Information summarized in Isx. (b) Example of the correlation between Isx and other statistics, here DVxy-coding. (c) Summary of the correlations between Isx and other population-genetic statistics. All the correlation coefficients are high and statistically significant.

(a) Geographical locations of samples. The base map was plotted in R using the mapdata package and circles were added using Photoshop. (b) PCA using common variants. (c) PCA using low-frequency variants. (d) Sharing of rare variants within and between populations. Upper left triangle: f₂ variants; lower right triangle f₃–f₁₀ variants. (e) Effective population size (Ne) inferred from IBDNe for UKO and UKG during the past nine KY. (f) The lowest Ne inferred by IBDNe for all populations for the past three KY, plotted as a function of the time at which it occurred.

(a) Rxy-missense statistic in each isolate, showing no evidence for increased genetic load in the isolates. The mean and s.d. for each Rxy value from 100 bootstraps are shown. (b) DVxy-wg (DVxy-whole genome) statistic in isolates and general populations, stratified by CADD score, showing enrichment of highly functional low-frequency variants. (c) DVxy-coding statistic in isolates and general populations, showing enrichment of low-frequency missense variants in isolates. (d) SVxy-missense statistic in each isolate, showing relaxation of purifying selection in isolates in singletons. The s.e.'s for both DVxy and SVxy were calculated by randomly sampling data from 20 chromosomes 100 times. All of these analyses are based on the minimum-sample-size data set (36 individuals from each population).