Genetic Stratigraphy of Key Demographic Events in Arabia

Abstract

At the crossroads between Africa and Eurasia, Arabia is necessarily a melting pot, its peoples enriched by successive gene flow over the generations. Estimating the timing and impact of these multiple migrations are important steps in reconstructing the key demographic events in the human history. However, current methods based on genome-wide information identify admixture events inefficiently, tending to estimate only the more recent ages, as here in the case of admixture events across the Red Sea (∼8–37 generations for African input into Arabia, and 30–90 generations for “back-to-Africa” migrations). An mtDNA-based founder analysis, corroborated by detailed analysis of the whole-mtDNA genome, affords an alternative means by which to identify, date and quantify multiple migration events at greater time depths, across the full range of modern human history, albeit for the maternal line of descent only. In Arabia, this approach enables us to infer several major pulses of dispersal between the Near East and Arabia, most likely via the Gulf corridor. Although some relict lineages survive in Arabia from the time of the out-of-Africa dispersal, 60 ka, the major episodes in the peopling of the Peninsula took place from north to south in the Late Glacial and, to a lesser extent, the immediate post-glacial/Neolithic. Exchanges across the Red Sea were mainly due to the Arab slave trade and maritime dominance (from ∼2.5 ka to very recent times), but had already begun by the early Holocene, fuelled by the establishment of maritime networks since ∼8 ka. The main “back-to-Africa” migrations, again undetected by genome-wide dating analyses, occurred in the Late Glacial period for introductions into eastern Africa, whilst the Neolithic was more significant for migrations towards North Africa.

Bayesian analyses of Yemeni mitochondrial genomes suggest multiple migration events with Africa and Western Eurasia

“…This is consistent with previous findings of multiple migration events between the Horn of Africa and the Near East (Pagani et al, 2012;Boattini et al, 2013;Hodgson et al, 2014;Fernandes et al, 2015). Of the twelve subhaplogroups that we studied because they are well-represented in Yemen, five of them reflect Type 2 migration.…”

“…Thus, the age and distribution of sequences in the Yemeni L3x1 clade indicates Yemeni in situ evolution and migration going back to the latest Pleistocene. This result is consistent with the mitochondrial founder analyses of Fernandes et al (2015), which also detected migrations from Africa to Arabia dating to 13 kya.…”

“…Furthermore, use of the program ALDER does not properly identify older admixture events since the program is biased towards more recent events (Hodgson et al, 2014). Additionally, Fernandes et al (2015) note their autosomal admixture dating analyses (using a similar program, ROLLOFF) could not detect the Pleistocene back-migrations into Africa that their mitochondrial founder analyses had detected. Notably, nuclear genome sequence data from Yemen could prove informative for investigating the question of the SDR vs. NDR.…”

“…Richards et al (2003) and others (e.g., Kivisild et al, 2004;Cern y et al, 2008;Badro et al, 2013;Fernandes et al, 2015) suggest that Arab slave trade over the last 2,500 years resulted in significant female-mediated migration from sub-Saharan Africa to Yemen as evidenced by a high frequency (relative to other Near Eastern regions) of African L haplogroups. Specifically, analyses of mitochondrial hypervariable region (HVR) sequences suggest there has been ongoing migration between Yemen and the Horn-of-Africa over the last 3,000-4,000 years as well as broader migration into Yemen from sub-Saharan Africa, South Asia, and western Eurasia Cern y et al, 2008).…”

“…Furthermore, analyses using whole mitogenome sequences from haplogroups frequently found in Yemen (e.g., HV1, R0a, and R2) indicate Yemeni mitogenome variation may be the product of multiple population expansions within the last 10-15 kya Musilov a et al, 2011;Al-Abri et al, 2012). Additionally, Fernandes et al (2015) argue based on founder analyses that Arabia mitogenomic diversity was influenced by multiple stages of migrations from the Levant, subSaharan Africa, and South Asia over the last 16 kya. Interestingly, there is a noticeable gap between these largely Holocene (11.7 kya to present) migration events involving Yemen and the ancient OOA migration.…”