Fig. 3

Signatures of gene flow among Nubian, reticulated, and Masai giraffe s. str. populations. a Admixture graph of giraffe populations with migration events. The okapi was used as an outgroup but was omitted from the image for a better resolution of the divergence between giraffe populations. Nubian, reticulated, and Masai s. str. giraffe populations were defined following the best fit admixture model (K = 9; Fig. 1c–e). Numbers within parentheses in Nubian, reticulated, and Masai s. str. giraffe population labels indicate sampling locations according to Fig. 1a. Migration arrows are colored according to their weight and marked following the number of migration events (m = 1 or m = 2) allowed in the model in which they were first inferred. The complete admixture graphs inferred for m = 1 and m = 2 and their corresponding residual fits are shown in Additional file 2: Fig. S6. b Heatmap showing the f-branch (f_b) statistic estimated based on the topology recovered by OrientAGraph. Statistical significance was assessed through block-jackknifing with 100 equally sized blocks. f_b values are shown for tests where the p-value of the associated D statistic is < 0.01. Gray boxes indicate tips/branches which cannot be tested under a ((P1, P2) P3, outgroup) topology. c Contemporary migration rates among Nubian, reticulated, and Masai giraffe s. str. The posterior mean migration rates were estimated based on 8137 randomly sampled unlinked SNPs from 97 wild giraffe. The resulting circular plot is based on the run with the smallest Bayesian deviance out of three independent replicates. Links with arrow tips indicate migration direction. Link widths are proportional to the fraction of individuals in the recipient population with ancestry in the source population (per generation). Scale ticks represent the cumulative fraction of migrants (per generation). Posterior estimates and 95% credible sets for migration rates are provided in Additional file 1: Table S2