|
Post by Admin on Oct 25, 2020 23:24:08 GMT
Materials and Methods Population Samples We have scanned for LP variants in five population samples from Arabia (Yemen) and five Arabic-speaking populations in Africa. Of these populations, all but two (the Rashaayda from Sudan and the Sudanese Arabs) correspond with samples whose mtDNA data have been previously published (Černý et al. 2007, 2008, 2009; Kujanová et al. 2009). The results and geographical locations of all the populations examined here are presented in Table 1. While all Yemeni groups and two Arabic populations in Africa maintain a sedentary lifestyle, the ancestors of the three remaining Arabicspeaking African populations sampled in Sudan, Chad, and Nigeria entered Africa as nomadic pastoralists, and their present-day descendants still continue in this tradition. However, while the Rashaayda in Sudan and Baggara in Chad continue to rely on nomadism fully, the Shuwa in Nigeria today already lead a more sedentary, seminomadic lifestyle. These Shuwa Arabs are also, according to our previous mtDNA analyses, more admixed with sub-Saharan populations (Černý et al. 2007). For some comparative purposes, previously published data sets characterizing other populations from the Arabian Peninsula and its neighborhood were included in some of the analyses (see the Appendix). Laboratory and Statistical Analyzes In total, 920 chromosomes (460 individuals) were screened for the variants associated with LP. We sequenced the 359-bp fragment located in intron 13 of the MCM6 gene where the four main LP-associated mutations can be detected. We used the same primers as reported in a previous study (Coelho et al. 2009). PCR products were sequenced with forward primers, and in cases of ambiguity, reverse complements were generated. The variants were identified by means of BioEdit software (Hall 1999), and allele frequencies were then calculated. Chisquare tests for the Hardy-Weinberg equilibrium, heterozygosity, fixation indexes (F), analysis of molecular variance (AMOVA), and principal coordinate analysis (PCoA) were undertaken by means of the GenAlEx statistical package (Peakall and Smouse 2012). The spatial frequency distributions of the main “Arabian” LP variant –13,915*G found in Arabia and neighboring regions was visualized by constructing interpolation maps using the “Spatial Analyst Extension” of ArcView version 3.2 (www. esri.com/software/arcgis/arcview/). The clinal pattern of –13,915*G within the Arabian Peninsula was further tested by correlogram autocorrelation analysis (Moran 1950) as implemented in PASSaGE software (Rosenberg and Anderson 2001); these spatial analyses also included other previously published data sets (see Appendix).
Table 1. Population Samples Used in This Study Population Abbreviation N Sampling location Country Longitude Latitude Lifestyle Yemeni 1 YAC 43 Al-Akhkum Yemen 44.18 13.32 Sedentary Yemeni 2 YTI 66 Around Hudeida Yemen 43.00 14.82 Sedentary Yemeni 3 YHG 34 Around Hajja Yemen 43.60 15.70 Sedentary Yemeni 4 YHA 40 Wadi Hadramawt Yemen 48.74 15.93 Sedentary Yemeni 5 YSO 65 Soqotra Yemen 53.85 12.50 Sedentary Arabs Rashaayda RAS 52 Abu Talha Sudan 36.30 15.35 Nomadic Arabs Sudan ARS 46 Along the Nile Sudan 30.55 19.88 Sedentary Arabs Baggara ACH 27 Around Mao Chad 15.31 14.12 Nomadic Arabs Shuwa ASW 53 Ngala and around Nigeria 14.19 12.34 Seminomadic Arabs Egypt ELH 34 El Hayz Egypt 28.66 28.02 Sedentary
Table 2. Intrapopulation Variation of –13,915*G in the Yemeni Populations and African Arabs Population TT TG GG T G Chi squared p-Value Ho He FI YAC 6 18 19 0.349 0.651 0.265 0.606 0.419 0.454 0.079 YTI 14 26 26 0.409 0.591 2.263 0.132 0.394 0.483 0.185 YHG 5 16 13 0.382 0.618 0.000 0.983 0.471 0.472 0.004 YHA 20 19 1 0.738 0.263 2.057 0.151 0.475 0.387 –0.227 YSO 15 24 26 0.415 0.585 3.737 0.053 0.369 0.486 0.240 RAS 4 16 32 0.231 0.769 0.924 0.336 0.308 0.355 0.133 ARS 39 7 0 0.924 0.076 0.312 0.576 0.152 0.141 –0.082 ACH 8 13 6 0.537 0.463 0.027 0.869 0.481 0.497 0.032 ASW 47 5 1 0.934 0.066 2.932 0.087 0.094 0.123 0.235 ELH 33 1 0 0.985 0.015 0.008 0.931 0.029 0.029 –0.015 abbreviations: Ho, observed heterozygosity; He, expected heterozygosity; FI, fixation index. Chi-square tests are for Hardy-Weinberg equilibrium. The p-values indicate probability for the chi-square tests—all are statistically nonsignificant at the 5% level.
|
|
|
Post by Admin on Oct 26, 2020 20:46:59 GMT
FIGURE 1. PCoA plot based on the frequencies of –13,915*G, –13,910*T, and –14,009*G. Squares, Arabian samples; circles, African samples. For population abbreviations, see Table 1. Results Genotypes and allelic frequencies of –13,915*G, the most frequently occurring LP variant in our populations, are reported in Table 2, showing that it appears in all analyzed Arabian (Arabian Peninsula) and Arabic-speaking (Africa) groups, albeit with varying frequency. While the Yemeni populations display more or less similar frequencies (except for the one from Hadramawt), Arabic-speaking populations from Africa fall into one of two distinct groups: one having a high and another having a low population frequency of this variant. Interestingly, this division correlates with lifestyle, supporting the hypothesis of the recent introduction of this variant from Arabia; both nomadic groups from Sudan (the Rashaayda) and Chad (the Baggara) attain relatively high –13,915*G frequencies, while the seminomadic Shuwa from Nigeria and sedentary Arabs from Sudan and Egypt show a much lower frequency of this variant than do the Yemenis and the Arabian nomads from Africa. AMOVA provides an interpopulation insight into the structure of the LP variants in our samples; both –13,915*G and other, more frequently occurring variants, such as –13,910*T and –14,009*G, were considered in this analysis. The percentage of whole molecular variance for all variants combined was 60% within populations and 40% among populations. When considering the –13,915*G variant alone, 56% of the molecular variance was within populations and 44% among populations; for –13,910*T alone, 93% of the molecular variance was within populations and 7% among populations; and for –14,009*G alone, 87% of the molecular variance was within populations and 13% among populations. Most of the overall variance between the combined populations is therefore explained by variant –13,915*G. The construction of the PCoA plots was carried out on the PhiPT matrix values calculated by the AMOVA. This analysis clearly separates the analyzed samples into two different groups (see Figure 1): one composed of populations with higher frequencies of the 13,915*G variant (left), and the other composed of populations with lower frequencies of this variant (right). The populations of the Hadramawt (YHA) and the Baggara Arabs (ACH), both having intermediate distributions of –13,915*G, lie between these groups The chi-square test for the Hardy-Weinberg equilibrium shows that the frequency of 13,915T/G genotypes is in equilibrium in all 10 analyzed populations (Table 2). This result is confirmed by comparisons of observed versus expected heterozygosities measured by fixation indexes; only the Shuwa Arabs from Nigeria (ASW) and the population from Soqotra Island (YSO) are characterized by higher positive values, indicating a very slight deviation toward an excess of homozygotes, even if levels of significance were not exceeded.
FIGURE 2. Map for –13,915*G frequencies in Arabia and neighboring regions. For population data, see the Appendix.
|
|
|
Post by Admin on Oct 26, 2020 22:37:25 GMT
FIGURE 3. Spatial autocorrelation analyses for –13,915*G. Circles are statistically significant at the 5% level; x, statistically nonsignificant at the 5% level. The geographical distribution of –13,915*G is presented in Figure 2, where its geographical dominance in Arabia is clearly visible. Only the Hadramawt and most of the northern Omani populations (ODA, ODH, OMU, and OBA; see Appendix) have a lower frequency of this mutation. The diagram clearly shows that there is a virtual absence of –13,915*G eastward of Arabia in the Middle East and India. On the other hand, Africa has a much higher frequency of this variant. It occurs in the Baggara and especially in the Rashaayda, both of which, as evidenced by the mtDNA data, have a relatively low African admixture; both groups are also highly dependent on nomadic pastoralism and milking. The correlogram analyses of the –13,915*G population frequencies (Indian, Chad, and Nigerian data excluded) show that a clinal pattern cannot be rejected (Figure 3). In fact, there is a depression signal of a cline that afffects only a part of the study area (see significant autocorrelation coeffficients present for both the highest and lowest Moran’s I-values in Figure 3). Aside from the –13,915*G variant, we have detected other mutations within the analyzed segment of the MCM6 gene (see Table 3). The “European” –13,910*T was found in two eastern Yemeni samples in Hadramawt and Soqotra, and in the Baggara and Shuwa. On the other hand, in Yemeni samples the “African” LP-associated variants –13,907*G and –14,010*C were revealed to come from the western part of the country, which is geographically close to Africa. In Africa, –13,907*G was found only in sedentary Arabs from Sudan. The Shuwa Arabs show two additional mutations: –13,965*G and –14,107*A. Moreover, –14,009*G was observed quite frequently in all analyzed samples of the African Arabs (except in the Rashaayda), and in one case from the western part of Yemen. Last but not least, one Egyptian from el Hayz carries an additional –14,042*G mutation. All these mutations were detected in heterozygote state only (Table 3). Interestingly, some of the variants just described were present together in the same individual, showing three combinations in total. The combination of –13,915*G and –14,009*G was seen twice in the population of the Baggara Arabs and once in the population of the Sudanese Arabs. A combination of the –13,910*T and –13,915*G variants was seen in three populations: once in the sample of Baggara Arabs, once in the Shuwa Arabs, and once in the Yemeni 4 (YHA) group. One individual from the Yemeni 1 (YAC) population also revealed a case of the combination of –14,010*C and –13,915*G.
Table 3. Counts (in Parentheses) and Frequencies of Mutations Found in the Analyzed Segment of MCM6 Gene Population –13,915*G –13,907*G –13,910*T –13,965*G –14,009*G –14,010*C –14,042*G –14,107*A YAC (56) 0.651 (1) 0.012 (1) 0.012 YTI (78) 0.591 (1) 0.015 (1) 0.008 YHG (42) 0.618 YHA (21) 0.263 (3) 0.038 YSO (76) 0.585 (5) 0.046 RAS (80) 0.769 ARS (7) 0.076 (1) 0.022 (11) 0.130 ACH (25) 0.463 (1) 0.019 (2) 0.037 ASW (7) 0.066 (9) 0.085 (1) 0.009 (1) 0.009 (1) 0.009 ELH (1) 0.015 (4) 0.059 (1) 0.015
|
|
|
Post by Admin on Oct 27, 2020 4:26:21 GMT
Discussion It has been believed that the ancestors of the Yemeni people were (in contrast to the Bedouins of Saudi Arabia) sedentary cultivators (Chelhod 1984). However, we have shown here that LP, which both enables fresh milk consumption in larger quantity and supports the hypothesis of the coevolution of genes and culture (Gerbault et al. 2009; Holden and Mace 2002), occurs with a relatively high frequency throughout Yemeni territory (perhaps except in the Hadramawt region). Although the original food production system in southern Arabia is not yet understood in full detail, there are clear archaeological indications of the existence of cattle keepers in Yemen at the beginning of the Neolithic. In fact, a cattle-keeping population has been discovered (Fedele 2009) at the Early Neolithic Yemeni settlement at Chawlan, in the at-Tiyal area to the east of Sana’a (de Maigret 2003). In the eastern part of Yemen (Hadramawt), the first traces of pastoralism are even more pronounced, as documented by 6,400-year-old cattle sacrifices found in Kheshiya (McCorriston et al. 2012). Such findings underline the economic and social importance of cattle, and likely also of milking practices. Archaeological evidence also reveals that, due to climatic deterioration in the Middle Holocene and its associated degradation of natural pastures, the original Hadramawt herders later started to concentrate in a more limited area and established the first irrigation systems to maintain pastures, enabling and with time leading to the cultivation of some plants (Harrower 2008). These changes subsequently led to a division within the southern Arabian population into nomadic pastoralists and settled farmers. In fact, such archaeological findings match our genetic data showing that LP-associated variants are found throughout all Arabian populations, with the only exception being those found in the Hadramawt. This discrepancy can be explained by the repeated waves of out-migration and back-immigration experienced by this specific region of southern Arabia during the last 500 years (Manger 2010); thus, it is not surprising that the region today harbors an uncommonly high level of sub-Saharan mtDNA haplogroups (Černý et al. 2008). The demographic history of the Arabian pastoralists is closely linked with the rise of South Arabian caravan kingdoms and the flourishing trade with Mediterranean civilizations. Pastoralists took an active part in this business, as domestication of the camel (Breton 1999; Retsö 1991) allowed them to export frankincense and other items across the barren desert of central Arabia. After the collapse of South Arabian civilization in the sixth century AD, some of the Arabian herders lost their jobs and were forced look for new opportunities outside of Arabia. Their dispersal to Africa is linked with the spread of Islam and is carefully recorded in archaeological and historical sources (Levy and Holl 2002; Zeltner 2002). The most important migration of the Arabian pastoralist tribes to Africa was associated with the conquest of Egypt in seventh century AD (Zeltner 2002), but deeper penetration into the continent was prevented for several centuries by the Christian kingdoms of Nubia. It was only after the Mamluks had conquered the Nubian Dongola in the beginning of fourteenth century that the Arabian pastoralists could continue with their further migration into the African interior. By the beginning of the sixteenth century they had finally dispersed along the Blue and White Nile, as well as westward to the Lake Chad Basin. It is also known that Arabic tribes appeared between Lake Fitri in Chad and Bahr el Ghazal in Sudan in that century, in the area once called Shuwa (Seignobos 2000). Some tribes reached Baguirmi in the mid-sixteenth century and the eastern shores of Lake Chad at the beginning of the eighteenth century, crossing the Shari River shortly thereafter. Immigration of Arabic-speaking pastoralists to Africa was, however, likely a continuous process, as is evidenced by the Rashaayda Bedouins, who entered Africa only very recently in the 1860s (Young 1996). The above-described wandering of the Arabic tribes through Africa has been revealed by both archaeological findings and historical documents and is consistent with our observed frequencies of –13,915*G: its prevalence and partly clinal pattern within the Arabian Peninsula and closer-lying parts of Africa and patchy distribution within broader Africa. Interestingly, while the recent Arabic immigrants (the Rashaayda) bear only the –13,915*G variant, groups arriving earlier have gained a more varied repertoire of LP-associated mutations that might have been introduced by other African nomads. This is especially the case for the –14,009*G variant, which might have been introduced into the Arabic pastoralist population in Africa by Somali camel herders (Ingram et al. 2009). The Shuwa, who reached as far as the Lake Chad Basin, bear –13,910*T; this variant might have been introduced into their population by possible contacts with the Fulani, the only sub-Saharan African population where this “European” variant has been detected in higher frequency (Lokki et al. 2011; Ranciaro et al. 2014). In conclusion, our data contribute to the hypothesis that –13,915*G’s potential place of origin is in Arabia (Enattah et al. 2008). The high frequency of LP in Arabia, and even in the southern locations where plant cultivation is more significant than pastoralism, is consistent with archeological records describing a Yemeni Early Neolithic mode of subsistence (Fedele 2009; McCorriston and Martin 2009). We show that the non-Arabic-speaking people on Soqotra bear a frequency of the –13,915*G variant similar to that found on the Arabian mainland, which in fact suggests that this island had likely been colonized by people already bearing this variant; indeed, a relatively recent colonization of this specific island around 6 kya has been suggested by mtDNA analyses (Černý et al. 2009). Furthermore, the decreasing gradient of the frequency of –13,915*G among Arabic-speaking groups in African Sahel and its virtual absence in South Africa (Breton et al. 2014; Ranciaro et al. 2014) supports a model of migration, sedentarization, and admixture with local African populations, as is evidenced in the Shuwa Arabs (Levy and Holl 2002). Our study also presents the highest frequency of the –13,915*G variant so far recorded in the genetic literature, attaining 76.9% in the Rashaayda Bedouins, who, while today living in Eastern Sudan, have descended from Arabian ancestors that arrived in Africa only some 150 years ago. The ancestors of the African Rashaayda originally probably lived in the region of Hejaz lying in western Saudi Arabia close to Mecca. In fact, there are many tribes in Arabia bearing the name Rashaayda, but all claim to have originated in Hejaz. Interestingly, further support of the Hejaz origin of the African Rashaayda is provided by the similar features of their female costumes (Young 1996).
|
|