Post by near on Feb 22, 2011 19:02:18 GMT -5
Mitochondrial DNA reveals distinct evolutionary histories for Jewish populations in Yemen and Ethiopia
1. Amy L. Non1,*,
2. Ali Al-Meeri2,
3. Ryan L. Raaum3,
4. Luisa F. Sanchez1,
5. Connie J. Mulligan1
Article first published online: 7 JUL 2010
DOI: 10.1002/ajpa.21360
Southern Arabia and the Horn of Africa are important geographic centers for the study of human population history because a great deal of migration has characterized these regions since the first emergence of humans out of Africa. Analysis of Jewish groups provides a unique opportunity to investigate more recent population histories in this area. Mitochondrial DNA is used to investigate the maternal evolutionary history and can be combined with historical and linguistic data to test various population histories. In this study, we assay mitochondrial control region DNA sequence and diagnostic coding variants in Yemenite (n = 45) and Ethiopian (n = 41) Jewish populations, as well as in neighboring non-Jewish Yemeni (n = 50) and Ethiopian (previously published Semitic speakers) populations. We investigate their population histories through a comparison of haplogroup distributions and phylogenetic networks. A high frequency of sub-Saharan African L haplogroups was found in both Jewish populations, indicating a significant African maternal contribution unlike other Jewish Diaspora populations. However, no identical haplotypes were shared between the Yemenite and Ethiopian Jewish populations, suggesting very little gene flow between the populations and potentially distinct maternal population histories. These new data are also used to investigate alternate population histories in the context of historical and linguistic data. Specifically, Yemenite Jewish mitochondrial diversity reflects potential descent from ancient Israeli exiles and shared African and Middle Eastern ancestry with little evidence for large-scale conversion of local Yemeni. In contrast, the Ethiopian Jewish population appears to be a subset of the larger Ethiopian population suggesting descent primarily through conversion of local women. Am J Phys Anthropol, 2010. © 2010 Wiley-Liss, Inc
Comparison between Jewish populations and non-Jewish neighbors
Haplogroup distributions and haplotype networks: Yemenite Jews and non-Jews
The mitochondrial haplogroup distributions of Yemenite Jews were compared to those of their non-Jewish neighbors (Fig. 2, Supporting Information Fig. S1). The frequency of sub-Saharan African L(xM,N) haplotypes is higher in the non-Jewish population (30%) than in the Jews (20%) (see Fig. 2). A previous study of 115 non-Jewish Yemeni found even higher levels of L(xM,N) haplotypes (48%, Kivisild et al., 2004) although those samples were collected in Kuwait, which likely produced an over-sampling of individuals of African ancestry as Yemeni living in Kuwait are predominantly Akhdam (a low status Yemeni population of African ancestry) or from Hadrawmat (a region known to have higher African ancestry (Cerny et al., 2008). We believe that our expansive sampling of Yemen provides a better representation of non-Jewish Yemeni genetic diversity than any previously collected sample and, thus, we focus subsequent analyses on our Yemeni samples. The diversity of L(xM,N) haplotypes is also reduced in the Yemenite Jewish population as compared with non-Jewish Yemeni, with a complete lack of deeper L0-L2 haplotypes that are common in the non-Jewish Yemeni population. A previous study of Yemenite Jews found a similar frequency of L(xM,N) (18%) and similar lack of L0-L2 haplotypes (Behar et al., 2008). Though reduced in comparison to Yemenis, the frequency and diversity of L(xM,N) haplotypes in Yemenite Jews is still much higher than typically seen in other Jewish populations.
The Yemenite Jewish population also contains a high frequency of west Eurasian haplogroups, including most of the major subclades of R (∼67%) as well as non-R haplogroups N and W (∼9%), in addition to haplogroup M (∼4%) that is found both in and out of Africa. Some of the Eurasian haplogroups in the Yemenite Jews are entirely absent in the non-Jewish Yemeni population, i.e. HV1, I1, J2a1a, U, U1 and W, suggesting some non-Yemeni Eurasian maternal contribution to the Jewish population. The R subclades, e.g., HV, J, U, are found at high frequencies further north in the Middle East/Caucasus region; e.g. Saudi Arabia (70.5%), Jordan (77%), Palestine (74%), Syria (87%), Bedouin (69%), Iraq (87%), and Iran (80%) (compiled in (Abu-Amero et al., 2007), indicating a potential source for R variation in the Yemenite Jews.
A network analysis was used to represent the exact sequence differences between haplotypes and illustrate the evolutionary relationship of Jewish and non-Jewish Yemeni (Supporting Information Fig. S1). The two populations are clearly different, as only two of the haplotypes in the Jewish population are identical to haplotypes in the non-Jewish population (2/30, 6.7%). On average, each Jewish individual's haplotype is ∼2.1 mutational steps away from a non-Jewish Yemeni haplotype, with a few individuals as many as four or six mutations away. When comparisons are made that include the larger Yemeni dataset (n = 115) collected by Kivisild et al. ( 2004) in Kuwait and the larger Yemenite Jewish sample (n = 119) collected by Behar et al. ( 2008), 19.3% (11/57) of Jewish haplotypes are identical to a haplotype in the non-Jewish population.
Haplogroup distributions and haplotype networks: Ethiopian Jews and non-Jews
Next, Ethiopian Jewish individuals were compared with non-Jewish Semitic-speaking Ethiopians (Kivisild et al., 2004) (Fig. 2, Supporting Information Fig. S2). Examination of the haplogroup distribution reveals Ethiopian Jews to be a potential subset of the greater Ethiopian population (see Fig. 2). With one exception (L3b2), no unique haplogroup was identified in the Ethiopian Jewish population that was not also present in the non-Jewish population. The Ethiopian Jewish proportion of L(xM,N) haplotypes is nearly identical to that found in Semitic-speaking Ethiopians, i.e., just over 50% in both populations. Ethiopian Jewish and non-Jewish haplotypes show a great deal of overlap, with a high frequency of Jewish haplotypes found in the non-Jewish population (9/27, 33.3%). Furthermore, each Ethiopian Jewish haplotype is on average only 0.88 mutations away from a non-Jewish Ethiopian haplotype, suggesting a close and recent relationship between the two populations. In contrast, the Ethiopian Jewish population shows very low mitochondrial similarity to other Jewish populations, with no haplotypes identical to any found in other Jewish populations (Thomas et al., 2002; Behar et al., 2008).
Jewish Diaspora populations are some of the most studied populations in the world due to their successful migration and colonization of multiple global regions, their unique history of population isolation, and the wealth of archaeological, linguistic, and genetic data available. We focus on two Jewish groups at the geographic crossroads of some of the most significant migrations in human history, including the first migration of modern humans out of Africa. We compare mitochondrial diversity between these Jewish populations and surrounding groups in order to better understand their maternal population histories and the broader history of the region itself. We integrate our data with other anthropological data to investigate alternate origin and migration histories.
Our data show that Yemenite and Ethiopian Jewish populations differ from other Jewish Diaspora populations due to a high frequency of sub-Saharan African L(xM,N) mitochondrial haplotypes, i.e., ∼20% in Yemenite Jews and ∼50% in Ethiopian Jews. A higher frequency of L(xM,N) haplogroups could be explained by proximity to Africa except for the fact that all North African Jewish populations (n = 33 Moroccan, Tunisian, and Libyan) carry only a single L haplotype and other Middle Eastern Jewish populations (n = 23 Iranian and Iraqi) have no L haplotypes (Picornell et al., 2006). This high frequency of L(xM,N) haplogroups in both Yemenite and Ethiopian Jews suggests a relatively large number of African founders and/or a fair amount of admixture with African populations. A very high frequency of the Eurasian haplogroup R0a was also found in both Yemenite Jews (11%) and Ethiopian Jews (22%). However, despite the high frequency of L(xM,N) and R0a haplogroups in both Yemenite and Ethiopian Jewish populations, no exact haplotype matches were identified between the two Jewish populations. This low level of haplotype sharing implies very little gene flow between the two populations, despite their geographic proximity, and potentially unique population histories for each population.
Ethiopian Jewish population history
In contrast to Yemenite Jews, our analysis of Ethiopian Jewish mitochondrial diversity indicates strong similarity with the neighboring non-Jewish population, e.g. both populations carry ∼50% L(xM,N) haplotypes and 38% of the Jewish haplotypes have exact matches with non-Jewish Ethiopian haplotypes. Ethiopian Jewish mitochondrial diversity is, effectively, a subset of Ethiopian non-Jewish diversity. This result supports previous studies of Ethiopian Jewish mtDNA, which also show a high similarity between Ethiopian Jewish and non-Jewish sequences (Zoossmann-Diskin et al., 1991; Ritte et al., 1993; Thomas et al., 2002; Behar et al., 2008).
The main alternative origin theories for the Ethiopian Jewish population consist of (1) descent from ancient Israeli exiles ∼2,500–2,700 years ago, (2) descent from Ethiopian converts to Judaism in the fourth century, and (3) descent from Ethiopian Christians who adopted Jewish Biblical practices in 14th–16th centuries. The fact that Ethiopian Jewish mitochondrial diversity represents a subset of non-Jewish Ethiopian diversity suggests that the Jewish community may have formed from a local subsampling of the Ethiopian population. Additionally, the lack of shared haplotypes with other Jewish populations (Thomas et al., 2002; Behar et al., 2006) further weakens support for descent from ancient (female) Israeli exiles. Finally, the age of ∼1,000–2,160 years for the 16305(T) variant implies a minimal age for the Ethiopian Jewish population that is too recent to be consistent with the first theory.
The second and third origin theories both suggest that the Jewish population derived directly from converted Ethiopians, either from Agaw-speakers in the forth century CE or Ethiopian Orthodox Christians in the 14th–15th centuries CE. Our simulation analyses suggest that the number of converted female individuals, i.e., the founding maternal population, might have been as small as 224 effective individuals. The timeframes for both origin theories are within the range of our simulation estimates for the minimum age of the Jewish population, i.e. ∼420–1,690YBP (although the age estimated using the method of Tavare et al. ( 1997) is older than both origin theories). Linguistic and historic data support the idea that Judaism was present in Ethiopia prior to Ethiopian Orthodox Christianity, thus accounting for the similarity between the two religions, i.e., contra the third origin theory. Specifically, there are numerous Jewish-Aramaic or Hebrew loanwords for Jewish religious concepts (meswat [“commandment” or “charity”], tabot [“Noah's Ark” or “Ark of the Covenant], ta'ot [idol]) in the Ge'ez version of the Ethiopic Old Testament, attesting to a Jewish presence in Ethiopia at or prior to the time of the translation of the Christian Bible from Greek (Polotsky, 1964; Kaplan, 1992). Additionally, an historical account by Venetian explorer Marco Polo describes Jews in Abyssinia (Ethiopia) as far back as the 13th century CE (Quirin, 1992), thus supporting an early presence of Jews in Ethiopia. However, if the second origin theory is correct, Ethiopian Jews would have spoken Agaw dialects prior to the adoption of Amharic (Henze, 1977; Quirin, 1992; Kessler, 1996) leading to the prediction that Ethiopian Jews might be genetically closer to Cushitic-speakers than Semitic-speakers (Henze, 1977). This prediction is not supported by our data, i.e., there are fewer haplotypes in the Ethiopian Jewish population found in Cushitic Ethiopians than Semitic Ethiopians (15.4% vs 34.6%). Thus, our genetic data provide support for a theory of large-scale conversion of local Ethiopians, but cannot distinguish between the two primary conversion theories.
Maternal and paternal histories may vary for Ethiopian Jews as for Yemenite Jews. For instance, we assayed nine Y microsatellite markers in the male R0a1b individuals and found all males exhibited a different Y haplotype (data not shown), in contrast to their identical mitochondrial genomes. A previous study of Y variation showed a uniquely close relationship between Ethiopian Jews and Africans, similar to our mitochondrial data, but also demonstrated a high level of shared Jewish ancestry in Ethiopian Jews in contrast to our mitochondrial data (Shen et al., 2004). Shared paternal, but not maternal, Jewish ancestry would be consistent with the popular Ethiopian legend of the Queen of Sheba, in which Prince Menelik, the purported son of Israeli King Solomon and Ethiopian Queen of Sheba, may have introduced as many as 10,000 Israeli men into the local Ethiopian population, converting many Ethiopian women to the Jewish religion (Quirin, 1992).
Our mitochondrial DNA study of Jewish populations in the HOA and Arabia sheds light on the history of two religiously defined groups formed relatively recently in these geographic regions. We found no identical haplotypes shared between Yemenite and Ethiopian Jewish populations despite their geographic proximity. Our data on Yemenite Jews suggest possible maternal descent from ancient Israeli exiles, and also demonstrate shared African and Middle Eastern ancestry with little evidence for large-scale conversion of local Yemeni. In contrast, our data on Ethiopian Jews suggest maternal descent primarily from the local Ethiopian population.
onlinelibrary.wiley.com/doi/10.1002/ajpa.21360/full
1. Amy L. Non1,*,
2. Ali Al-Meeri2,
3. Ryan L. Raaum3,
4. Luisa F. Sanchez1,
5. Connie J. Mulligan1
Article first published online: 7 JUL 2010
DOI: 10.1002/ajpa.21360
Southern Arabia and the Horn of Africa are important geographic centers for the study of human population history because a great deal of migration has characterized these regions since the first emergence of humans out of Africa. Analysis of Jewish groups provides a unique opportunity to investigate more recent population histories in this area. Mitochondrial DNA is used to investigate the maternal evolutionary history and can be combined with historical and linguistic data to test various population histories. In this study, we assay mitochondrial control region DNA sequence and diagnostic coding variants in Yemenite (n = 45) and Ethiopian (n = 41) Jewish populations, as well as in neighboring non-Jewish Yemeni (n = 50) and Ethiopian (previously published Semitic speakers) populations. We investigate their population histories through a comparison of haplogroup distributions and phylogenetic networks. A high frequency of sub-Saharan African L haplogroups was found in both Jewish populations, indicating a significant African maternal contribution unlike other Jewish Diaspora populations. However, no identical haplotypes were shared between the Yemenite and Ethiopian Jewish populations, suggesting very little gene flow between the populations and potentially distinct maternal population histories. These new data are also used to investigate alternate population histories in the context of historical and linguistic data. Specifically, Yemenite Jewish mitochondrial diversity reflects potential descent from ancient Israeli exiles and shared African and Middle Eastern ancestry with little evidence for large-scale conversion of local Yemeni. In contrast, the Ethiopian Jewish population appears to be a subset of the larger Ethiopian population suggesting descent primarily through conversion of local women. Am J Phys Anthropol, 2010. © 2010 Wiley-Liss, Inc
Comparison between Jewish populations and non-Jewish neighbors
Haplogroup distributions and haplotype networks: Yemenite Jews and non-Jews
The mitochondrial haplogroup distributions of Yemenite Jews were compared to those of their non-Jewish neighbors (Fig. 2, Supporting Information Fig. S1). The frequency of sub-Saharan African L(xM,N) haplotypes is higher in the non-Jewish population (30%) than in the Jews (20%) (see Fig. 2). A previous study of 115 non-Jewish Yemeni found even higher levels of L(xM,N) haplotypes (48%, Kivisild et al., 2004) although those samples were collected in Kuwait, which likely produced an over-sampling of individuals of African ancestry as Yemeni living in Kuwait are predominantly Akhdam (a low status Yemeni population of African ancestry) or from Hadrawmat (a region known to have higher African ancestry (Cerny et al., 2008). We believe that our expansive sampling of Yemen provides a better representation of non-Jewish Yemeni genetic diversity than any previously collected sample and, thus, we focus subsequent analyses on our Yemeni samples. The diversity of L(xM,N) haplotypes is also reduced in the Yemenite Jewish population as compared with non-Jewish Yemeni, with a complete lack of deeper L0-L2 haplotypes that are common in the non-Jewish Yemeni population. A previous study of Yemenite Jews found a similar frequency of L(xM,N) (18%) and similar lack of L0-L2 haplotypes (Behar et al., 2008). Though reduced in comparison to Yemenis, the frequency and diversity of L(xM,N) haplotypes in Yemenite Jews is still much higher than typically seen in other Jewish populations.
The Yemenite Jewish population also contains a high frequency of west Eurasian haplogroups, including most of the major subclades of R (∼67%) as well as non-R haplogroups N and W (∼9%), in addition to haplogroup M (∼4%) that is found both in and out of Africa. Some of the Eurasian haplogroups in the Yemenite Jews are entirely absent in the non-Jewish Yemeni population, i.e. HV1, I1, J2a1a, U, U1 and W, suggesting some non-Yemeni Eurasian maternal contribution to the Jewish population. The R subclades, e.g., HV, J, U, are found at high frequencies further north in the Middle East/Caucasus region; e.g. Saudi Arabia (70.5%), Jordan (77%), Palestine (74%), Syria (87%), Bedouin (69%), Iraq (87%), and Iran (80%) (compiled in (Abu-Amero et al., 2007), indicating a potential source for R variation in the Yemenite Jews.
A network analysis was used to represent the exact sequence differences between haplotypes and illustrate the evolutionary relationship of Jewish and non-Jewish Yemeni (Supporting Information Fig. S1). The two populations are clearly different, as only two of the haplotypes in the Jewish population are identical to haplotypes in the non-Jewish population (2/30, 6.7%). On average, each Jewish individual's haplotype is ∼2.1 mutational steps away from a non-Jewish Yemeni haplotype, with a few individuals as many as four or six mutations away. When comparisons are made that include the larger Yemeni dataset (n = 115) collected by Kivisild et al. ( 2004) in Kuwait and the larger Yemenite Jewish sample (n = 119) collected by Behar et al. ( 2008), 19.3% (11/57) of Jewish haplotypes are identical to a haplotype in the non-Jewish population.
Haplogroup distributions and haplotype networks: Ethiopian Jews and non-Jews
Next, Ethiopian Jewish individuals were compared with non-Jewish Semitic-speaking Ethiopians (Kivisild et al., 2004) (Fig. 2, Supporting Information Fig. S2). Examination of the haplogroup distribution reveals Ethiopian Jews to be a potential subset of the greater Ethiopian population (see Fig. 2). With one exception (L3b2), no unique haplogroup was identified in the Ethiopian Jewish population that was not also present in the non-Jewish population. The Ethiopian Jewish proportion of L(xM,N) haplotypes is nearly identical to that found in Semitic-speaking Ethiopians, i.e., just over 50% in both populations. Ethiopian Jewish and non-Jewish haplotypes show a great deal of overlap, with a high frequency of Jewish haplotypes found in the non-Jewish population (9/27, 33.3%). Furthermore, each Ethiopian Jewish haplotype is on average only 0.88 mutations away from a non-Jewish Ethiopian haplotype, suggesting a close and recent relationship between the two populations. In contrast, the Ethiopian Jewish population shows very low mitochondrial similarity to other Jewish populations, with no haplotypes identical to any found in other Jewish populations (Thomas et al., 2002; Behar et al., 2008).
Jewish Diaspora populations are some of the most studied populations in the world due to their successful migration and colonization of multiple global regions, their unique history of population isolation, and the wealth of archaeological, linguistic, and genetic data available. We focus on two Jewish groups at the geographic crossroads of some of the most significant migrations in human history, including the first migration of modern humans out of Africa. We compare mitochondrial diversity between these Jewish populations and surrounding groups in order to better understand their maternal population histories and the broader history of the region itself. We integrate our data with other anthropological data to investigate alternate origin and migration histories.
Our data show that Yemenite and Ethiopian Jewish populations differ from other Jewish Diaspora populations due to a high frequency of sub-Saharan African L(xM,N) mitochondrial haplotypes, i.e., ∼20% in Yemenite Jews and ∼50% in Ethiopian Jews. A higher frequency of L(xM,N) haplogroups could be explained by proximity to Africa except for the fact that all North African Jewish populations (n = 33 Moroccan, Tunisian, and Libyan) carry only a single L haplotype and other Middle Eastern Jewish populations (n = 23 Iranian and Iraqi) have no L haplotypes (Picornell et al., 2006). This high frequency of L(xM,N) haplogroups in both Yemenite and Ethiopian Jews suggests a relatively large number of African founders and/or a fair amount of admixture with African populations. A very high frequency of the Eurasian haplogroup R0a was also found in both Yemenite Jews (11%) and Ethiopian Jews (22%). However, despite the high frequency of L(xM,N) and R0a haplogroups in both Yemenite and Ethiopian Jewish populations, no exact haplotype matches were identified between the two Jewish populations. This low level of haplotype sharing implies very little gene flow between the two populations, despite their geographic proximity, and potentially unique population histories for each population.
Ethiopian Jewish population history
In contrast to Yemenite Jews, our analysis of Ethiopian Jewish mitochondrial diversity indicates strong similarity with the neighboring non-Jewish population, e.g. both populations carry ∼50% L(xM,N) haplotypes and 38% of the Jewish haplotypes have exact matches with non-Jewish Ethiopian haplotypes. Ethiopian Jewish mitochondrial diversity is, effectively, a subset of Ethiopian non-Jewish diversity. This result supports previous studies of Ethiopian Jewish mtDNA, which also show a high similarity between Ethiopian Jewish and non-Jewish sequences (Zoossmann-Diskin et al., 1991; Ritte et al., 1993; Thomas et al., 2002; Behar et al., 2008).
The main alternative origin theories for the Ethiopian Jewish population consist of (1) descent from ancient Israeli exiles ∼2,500–2,700 years ago, (2) descent from Ethiopian converts to Judaism in the fourth century, and (3) descent from Ethiopian Christians who adopted Jewish Biblical practices in 14th–16th centuries. The fact that Ethiopian Jewish mitochondrial diversity represents a subset of non-Jewish Ethiopian diversity suggests that the Jewish community may have formed from a local subsampling of the Ethiopian population. Additionally, the lack of shared haplotypes with other Jewish populations (Thomas et al., 2002; Behar et al., 2006) further weakens support for descent from ancient (female) Israeli exiles. Finally, the age of ∼1,000–2,160 years for the 16305(T) variant implies a minimal age for the Ethiopian Jewish population that is too recent to be consistent with the first theory.
The second and third origin theories both suggest that the Jewish population derived directly from converted Ethiopians, either from Agaw-speakers in the forth century CE or Ethiopian Orthodox Christians in the 14th–15th centuries CE. Our simulation analyses suggest that the number of converted female individuals, i.e., the founding maternal population, might have been as small as 224 effective individuals. The timeframes for both origin theories are within the range of our simulation estimates for the minimum age of the Jewish population, i.e. ∼420–1,690YBP (although the age estimated using the method of Tavare et al. ( 1997) is older than both origin theories). Linguistic and historic data support the idea that Judaism was present in Ethiopia prior to Ethiopian Orthodox Christianity, thus accounting for the similarity between the two religions, i.e., contra the third origin theory. Specifically, there are numerous Jewish-Aramaic or Hebrew loanwords for Jewish religious concepts (meswat [“commandment” or “charity”], tabot [“Noah's Ark” or “Ark of the Covenant], ta'ot [idol]) in the Ge'ez version of the Ethiopic Old Testament, attesting to a Jewish presence in Ethiopia at or prior to the time of the translation of the Christian Bible from Greek (Polotsky, 1964; Kaplan, 1992). Additionally, an historical account by Venetian explorer Marco Polo describes Jews in Abyssinia (Ethiopia) as far back as the 13th century CE (Quirin, 1992), thus supporting an early presence of Jews in Ethiopia. However, if the second origin theory is correct, Ethiopian Jews would have spoken Agaw dialects prior to the adoption of Amharic (Henze, 1977; Quirin, 1992; Kessler, 1996) leading to the prediction that Ethiopian Jews might be genetically closer to Cushitic-speakers than Semitic-speakers (Henze, 1977). This prediction is not supported by our data, i.e., there are fewer haplotypes in the Ethiopian Jewish population found in Cushitic Ethiopians than Semitic Ethiopians (15.4% vs 34.6%). Thus, our genetic data provide support for a theory of large-scale conversion of local Ethiopians, but cannot distinguish between the two primary conversion theories.
Maternal and paternal histories may vary for Ethiopian Jews as for Yemenite Jews. For instance, we assayed nine Y microsatellite markers in the male R0a1b individuals and found all males exhibited a different Y haplotype (data not shown), in contrast to their identical mitochondrial genomes. A previous study of Y variation showed a uniquely close relationship between Ethiopian Jews and Africans, similar to our mitochondrial data, but also demonstrated a high level of shared Jewish ancestry in Ethiopian Jews in contrast to our mitochondrial data (Shen et al., 2004). Shared paternal, but not maternal, Jewish ancestry would be consistent with the popular Ethiopian legend of the Queen of Sheba, in which Prince Menelik, the purported son of Israeli King Solomon and Ethiopian Queen of Sheba, may have introduced as many as 10,000 Israeli men into the local Ethiopian population, converting many Ethiopian women to the Jewish religion (Quirin, 1992).
Our mitochondrial DNA study of Jewish populations in the HOA and Arabia sheds light on the history of two religiously defined groups formed relatively recently in these geographic regions. We found no identical haplotypes shared between Yemenite and Ethiopian Jewish populations despite their geographic proximity. Our data on Yemenite Jews suggest possible maternal descent from ancient Israeli exiles, and also demonstrate shared African and Middle Eastern ancestry with little evidence for large-scale conversion of local Yemeni. In contrast, our data on Ethiopian Jews suggest maternal descent primarily from the local Ethiopian population.
onlinelibrary.wiley.com/doi/10.1002/ajpa.21360/full