Fig. 2

Population structure and evidence of recombination. A Colours correspond to ancestral populations making up individuals. Country of origin (above) is Au = Australia, Mo = Morocco, SoA = South Africa, Ca = Canada, Fr = France, No = Norway, and UK = UK. Below, states within Australia and Canada are indicated, where NSW = New South Wales, SA = South Australia, WA = Western Australia, AB = Alberta, MB = Manitoba, and SK = Saskatchewan. B Linkage disequilibrium (y axis) decay with physical distance (x axis). Points are averages for unique distance measurements, and the red line is a general additive model fit. C The first two principal components of genotypic variance. Colours indicate geographical origin and point shapes the four population sub-samples used for recombination analysis. D Across chromosomes and population sub-samples, the distribution of Spearman’s correlations between chromosome end distance and recombination rate. E Correlation between coding DNA sequence content (x axis) and recombination rate (y axis) of 50-kb sliding windows for population-4. The line is a general additive model fit. The same plots for all populations are shown in Supplementary Fig. 2. F Boxplot showing percent gene content of 50-kb windows containing and not containing recombination hotspots (*** = P < 2e⁻¹⁶). Boxes and whiskers show interquartile range. G Circles show where windows containing putative centromeres lie on a plot of recombination rate (y axis) against log recombination rate (x axis). Putative centromeres are in regions of low recombination, before the inflection point. H The y axis is scaled (division by maximum) recombination rate, amount of methylation or GC content for sliding windows. The x axis shows position (Mb) across chromosome 6 (all chromosomes and population samples are in Additional file 20: Supplementary File 1). All chromosomes had a dip in GC coincident with a spike in methylation, almost always coincident with a recombination cold spot