Genetic Landscapes Reveal How Human Genetic Diversity Aligns with Geography

Abstract

Geographic patterns in human genetic diversity carry footprints of population history and provide insights for genetic medicine and its application across human populations. Summarizing and visually representing these patterns of diversity has been a persistent goal for human geneticists, and has revealed that genetic differentiation is frequently correlated with geographic distance. However, most analytical methods to represent population structure do not incorporate geography directly, and it must be considered post hoc alongside a visual summary of the genetic structure. Here, we estimate "effective migration" surfaces to visualize how human genetic diversity is geographically structured. The results reveal local patterns of differentiation in detail and emphasize that while genetic similarity generally decays with geographic distance, the relationship is often subtly distorted. Overall, the visualizations provide a new perspective on genetics and geography in humans and insight to the geographic distribution of human genetic variation.

Keywords: geographic structure; geography; human genetics; isolation-by-distance; population genetics; population structure.

FIg. 1

Large-scale patterns of population structure. (a) EEMS posterior mean effective migration surface for Afro-Eurasia (AEA) panel. Regions and features discussed in the main text are labeled. Approximate location of troughs is annotated with dashed lines (see supplementary fig. 2, Supplementary Material online). (b) PCA plot of AEA panel: Individuals are displayed as gray dots, colored dots reflect median of sample locations; with colors reflecting geography and matching with the EEMS plot. Locations displayed in the EEMS plot reflect the position of populations after alignment to grid vertices used in the model (see Materials and Methods). For exact locations, see annotated supplementary figure 2, Supplementary Material online and supplementary table S1, Supplementary Material online. The displayed value of F_ST emphasizes the low absolute level of differentiation in human SNP data.

FIg. 2

Regional patterns of genetic diversity. (a) Scale bar for relative effective migration rate. Posterior effective migration surfaces for (b) Western Eurasia (WEA) (e) Central/Eastern Eurasia (CEA) (g) Africa (AFR) (h) South East Asian (SEA) (k) Southern African KhoeSan (SAKS) (l) Southern African Bantu (SAB) analysis panels. In panel g, red circles indicate Nilo-Saharan speakers. Approximate location of troughs is shown with dashed lines (see supplementary fig. 4, Supplementary Material online). PCA plots: (c) WEA (d) Europeans in WEA (f) CEA (i) SEA (j) AFR (m) SAHG+SAB. Individuals are displayed as gray dots. Large dots reflect median PC position for a sample; with colors reflecting geography matched to the corresponding EEMS figure. In the EEMS plots, approximate sample locations are annotated. For exact locations, see annotated supplementary figure 4, Supplementary Material online and supplementary table S1, Supplementary Material online. Features discussed in the main text and Supplementary Material online are labeled. F_ST values per panel emphasize the low absolute levels of differentiation.