Human migration from the Levant and Arabia into Yemen since Last Glacial Maximum

Abstract

While a broad consensus about the first successful migration modern humans out of Africa seems established, the peopling of Arabia remains somewhat enigmatic. Identifying the ancestral populations that contributed to the gene pool of the current populations inhabiting Arabia and the impact of their contributions remains a challenging task. We investigate the genetic makeup of the current Yemeni population using 46 whole genomes and 169 genotype arrays derived from Yemeni individuals from all geographic regions across Yemen and 351 genotype arrays derived from neighboring populations providing regional context. Principal Component Analysis shows stratification between Yemen districts but also with respect to nearby populations: Yemeni, other Arabian and Bedouin samples form a continuum towards the populations of the Levant, whereas East Africa and India appear strongly differentiated. This finding is further supported by higher Principal Components, admixture and haplogroup analyses, and F-statistics. Moreover, two-reference linkage disequilibrium decay estimates are most significant for Yemeni admixture from an ancient northern influx (up to 5220BP from Palestine) and East Africa (750BP). We show that the initial gene flow into the Yemeni populations of today came from the rest of Arabia and the Levant, and a less substantial and more recent genetic impact into coastal Yemen from East Africa, particularly.

Keywords: Genomics; Human mobility; Populations; Yemen.

Fig. 1

Map of Yemen. It shows Yemen governorates (black borders). Sampling efforts are shown with scaled triangles at their respective geographic locations. Figure was created using the Python libraries matplotlib and basemap using the basemap method ‘shadedrelie (URL: https://pypi.org/project/basemap/ Version: 1.3.2.).

Fig. 2

Admixture (individuals) from populations within 3000 km of Yemen, ordered by geographic location, showing up to K = 5 (Lowest validation error). Hadramaut and Hudaydah regions exhibit strongest diversity. The Yemen governorates were grouped into six districts as follows: Aden (Abyan, Dali, Lahij), Azal (Amran, Dhamar, Saada, Sanaa), Sheba (Bayda, Jawf, Marrib), Hadramaut (Hadramaut, Shabwa), Hudaydah (Hajjah, Hudaydah, Mahwit), Aljanad (Ibb, Taizz).

Fig. 3

Global contextualization of Yemen samples A) Yemen forms a cluster (ellipse) and spreads between Levant and Arabian population (towards the bottom). The magnification of this cluster (elliptic inlet, above) shows a complex population structure of Yemen, which is further disentangled by regional contextualization. B) PC1 vs PC3.

Fig. 4

(a) forest plots for Sheba district (North) with outgroup Mbuti, (b) forest plots for Aden district (South) with outgroup Mbuti.

Fig. 5

Y-chromosome phylogeny. 46 Yemen samples with fully sequenced Y chromosome were subjected to phylogeny construction.

Fig. 6

FineSTRUCTURE tree for Yemen individuals including surrounding populations. The various populations are color coded by continental world population association. Numbers in parenthesis refer to the number of samples representing the respective collapsed clade.