Population structure of indigenous inhabitants of Arabia

Abstract

Modern day Saudi Arabia occupies the majority of historical Arabia, which may have contributed to ancient waves of migration out of Africa. This ancient history has left a lasting imprint in the genetics of the region, including the diverse set of tribes that call Saudi Arabia their home. How these tribes relate to each other and to the world's major populations remains an unanswered question. In an attempt to improve our understanding of the population structure of Saudi Arabia, we conducted genomic profiling of 957 unrelated individuals who self-identify with 28 large tribes in Saudi Arabia. Consistent with the tradition of intra-tribal unions, the subjects showed strong clustering along tribal lines with the distance between clusters correlating with their geographical proximities in Arabia. However, these individuals form a unique cluster when compared to the world's major populations. The ancient origin of these tribal affiliations is supported by analyses that revealed little evidence of ancestral origin from within the 28 tribes. Our results disclose a granular map of population structure and have important implications for future genetic studies into Mendelian and common diseases in the region.

Fig 1. Saudi population structure.

(A) Map of sampled tribes and admixture average proportions. Pie charts summarize per-tribe average proportions of Indigenous Arab samples at K = 7. Assumed ancestries are Arab, European, African, East Asian, South Asian, American and Oceanian. The map was made with Natural Earth (public domain). (B) Global ancestry proportions at K = 7 (top) and K = 18 (bottom) estimated with ADMIXTURE including African, European, Asian, American, Oceanian and Arab. (C) Cross-validation error for K runs from 1 to 25, K = 18 has the lowest cross-validation error.

Fig 2. Genetic differentiation of indigenous Arab population.

(A) Principal component analysis of 957 indigenous Arab samples. Tribal affiliations for T01-T28 are represented by different symbols. Colors correspond to the geographical location of Arabian Peninsula. (B) Pairwise F_ST values among indigenous Arab tribes. Values at each combination show the pairwise F_ST, calculated by BEDASSLE package in R. Colored intensity indicates the degree of pairwise F_ST at each combination. Brighter red indicates more differentiated between tribes compared.

Fig 3. The average of inbreeding coefficients for 28 tribes.

The bar chart shows the average of inbreeding coefficient for each tribe shown in decreasing order. The color indicates the regional group in the Arabian Peninsula. The inbreeding coefficient for each individual was calculated as “Fhat2” estimate using PLINK.

Fig 4. TreeMix representation of indigenous Arab populations with reference populations.

Saudi tribes (T01-T29) are coded by regional colors. Bootstrap confidence value is indicated by distinct symbols at each node. AFR; Africa, AMR; America, CAS; Central Asia, EAS; East Asia, EUR; Europe, OCN; Oceania, SAS; South Asia, WAS; West Asia, Q1; Bedouin in Qatar genome project, Q2; Persian-South Asian in Qatar genome project, Q3; African in Qatar genome project, Q0; unassigned in Qatar genome project.

Fig 5

(A) The estimated date of African admixture, (B) The ratio of African ancestry for 28 tribes. Tribes are sorted from right to left by decreasing proportion of estimated African admixture.