Genetic Roots of the Ashkenazi Jews Aug 14, 2019 17:50:10 GMT
Post by Admin on Aug 14, 2019 17:50:10 GMT
Geographic variation of ancestry sub-components in Latin American individuals. a Native American, b European, and c East/South Mediterranean. Each pie represents an individual, with pie location corresponding to birthplace. Since many individuals share birthplace, jittering has been performed based on pie size and how crowded an area is. Pie size is proportional to total ancestry from all sources depicted in that specific figure, and only individuals with >5% of such total ancestry are shown. Coloring of pies represents the proportion of each sub-continental component estimated for each individual (color-coded as in Fig. 1; Chaco2 does not contribute >5% to any individual and was excluded). Pies in (c) have been enlarged to facilitate visualization
Patterns of European ancestry in the CANDELA dataset
Importantly, SOURCEFIND distinguishes between closely-related ancestry components from the Iberian Peninsula, as well as from the East and South Mediterranean (including individuals self-identified as Sephardic; i.e. Iberian Jews; Supplementary Note 2). The distribution of European ancestry in the CANDELA sample shows a sharp differentiation between Brazil and the Spanish American countries (Fig. 1c). In Brazil the predominant European sub-component matches mostly the Portugal/West-Spain reference group while in Mexico, Colombia, Peru and Chile it is mostly Central/South-Spanish ancestry that is inferred (Figs. 1c and 2b). This differentiation closely matches colonial history. The European settlement of what is now Latin America involved two main areas of colonial expansion, as agreed in the Tordesillas treaty of 1494. This treaty established that territories west of a meridian somewhat east of the Amazon river-mouth were ascribed to Spain, while territories east of this meridian were attributed to Portugal3,30. Portuguese migration thus concentrated in Eastern South America, gradually expanding beyond the Tordesillas meridian until achieving Brazil’s current political borders, which at independence, remained a single political entity. Conversely, Spanish immigrants settled mainly in territories of Central America and Western South America, which at independence fragmented into separate countries3. The relatively small contribution inferred here for the Basque and Catalan agrees with historical information documenting that Spanish migrants to the Americas originated mainly in Southern and Central Spain31. In addition to Portugal/West-Spain ancestry the Brazilian sample also shows substantial genetic components most closely related to the Italian and German reference groups, and these concentrate in the South of the country (Fig. 2b). This pattern is consistent with the documented migration to Southern Brazil of large numbers of Germans and Italians starting in the late 19th century30.
Dating admixture from different sources
To assess the time-frame of admixture between the ancestry components described above we used the program GLOBETROTTER27. Since admixture proportions in Latin Americans vary greatly, we analyzed each individual separately; simulations confirmed the accuracy of GLOBETROTTER in this setting (Supplementary Note 2). Inferred dates for events involving an Iberian source (the initial settlers arriving from Europe and allegedly the first to admix with the Natives) had a median of ten generations (IQR = 7–13), consistent with other estimates for admixture in Latin America6,10,15. Noticeably, individuals with more recent inferred dates of admixture have greater Native ancestry (Fig. 3a, Supplementary Table 4), with simulations suggesting this is consistent with continuing admixture between admixed Latin Americans and unadmixed Natives (Supplementary Note 2), possibly as a result of the decline in Iberian immigration after the mid-17th century, concomitant with the demographic recovery of neighboring Native American populations1,32. Compared to inferred dates related to Iberian admixture, admixture events involving non-Iberian European sources (Northwest Europe, Italy) have a significant skew towards more recent dates (Fig. 3b; Wilcoxon rank-sum test one-sided p-value = 3 × 10−8), consistent with the relatively recent arrival of Germans and Italians30.
Times since admixture estimated using GLOBETROTTER. a Top: frequency distribution of admixture times for individuals in which a single admixture event between Native and European sources was inferred (dashed line indicates the mean). Bottom: mean continental ancestry (%) as a function of time since admixture among these individuals. Only time bins including >20 individuals are shown. (NAM Native American, EUR European, ESM East/South Mediterranean, SSA Sub-Saharan African, EAS East Asian). b–e show contrasts of the distribution of admixture times involving Iberian versus other sources: b North-West Europe and Italy, c East Mediterranean and Sephardic, d Sub-Saharan Africa and e East Asia. p-values for comparing the mean date of Iberian versus each other ancestry source are from a one-sided Mann–Whitney U test, and numbers of inferred admixture events are given in parenthesis
East/South Mediterranean ancestry in the CANDELA dataset
SOURCEFIND finds that Sephardic/East/South Mediterranean ancestry is detectable in each country’s samples: Brazil (1%), Chile (4%), Colombia (3%), Mexico (3%) and Peru (2%). Altogether, ~23% of the CANDELA individuals show >5% of such ancestry (an average of 12.2%) (Fig. 1d) and in these individuals SOURCEFIND infers this ancestry to be mostly Sephardic (7.3%), with smaller non-Sephardic East Mediterranean (3.9%) and non-Sephardic South Mediterranean (1%) contributions. Individuals with Sephardic/East/South Mediterranean ancestry were detected across Latin America (Fig. 2c). It is possible that outliers with particularly high values of Sephardic/East/South Mediterranean ancestry are descendants from recent non-European immigrants. For 19 of 42 individuals with >25% Sephardic/East/South Mediterranean ancestry, genealogical information (up to grandparents) identified ancestors born in the Eastern Mediterranean (thus validating the SOURCEFIND inference). However, no recent immigration was documented for other individuals, including all Colombians with >5% Sephardic ancestry (despite these Colombians showing the highest estimated Sephardic ancestry across countries; ~10% on average, Fig. 1d). Furthermore, GLOBETROTTER estimates for the time since East/South Mediterranean admixture were not significantly different from those involving Iberian sources (Fig. 3c; Wilcoxon rank-sum test one-sided p-value > 0.1), consistent with most of this ancestry component being contributed simultaneously with the initial colonial immigrants. Jewish communities existed in Iberia (Sepharad) since roman times and much of the peninsula was ruled by Arabs and Berbers for most of the Middle Ages, by the end of which large Sephardic communities had developed33. Genetic studies have detected South and East Mediterranean ancestry in the current Spanish population, as well European admixture in the Sephardim34,35,36. The estimates of South/East Mediterranean ancestry in Latin Americans obtained here represent values over and above those present in the Iberian individuals we examined, suggesting colonial migration to Latin America involved people with relatively higher levels of South/East Mediterranean ancestry. Columbus’ arrival to the New World in the late 15th century coincided with the expulsion and forced conversion of Spanish Jews, with similar measures subsequently affecting Spanish Muslims. Although Christian converts were legally forbidden from migrating to the colonies, historical records (often from the Inquisition) document that some individuals made the journey33. Since this migration was mostly a clandestine process, its magnitude has been difficult to assess. Genetic studies have occasionally provided evidence that certain Latin American populations could have some Converso ancestry and this is at times supported by some historical evidence3,37,38. Our findings indicate that the signature of a colonial migration to Latin America of people with relatively high South/East Mediterranean ancestry is much more prevalent than suggested by these special cases, or by historical records.
Sub-Saharan African ancestry in the CANDELA dataset
It has been estimated that Brazil received about 4.2 million African slaves (about half of those brought to the Americas) while Spanish America altogether received about 1.5 million3. However, the average Sub-Saharan ancestry in the full CANDELA sample is relatively low (<4%), probably reflecting the fact that regions which historically received large numbers of slaves are under-represented in this sample (particularly for Brazil, which was sampled mainly in the South of the country)19. Altogether, ~22% of the individuals studied show more than 5% sub-Saharan African ancestry. SOURCEFIND infers a marked predominance of the West African sub-component, particularly in the Spanish American countries (Supplementary Figures 5 and 6), consistent with previous genetic analyses, and with historical information1,39. The distribution of dates involving Sub-Saharan African admixture mostly overlaps with that for Iberian admixture, although a substantial proportion of recent dates were also inferred (Fig. 3d), possibly reflecting continuing African admixture in the regions sampled.
East Asian ancestry in the CANDELA dataset
Other than the major Native American, European/Mediterranean and sub-Saharan African ancestry components, historical information indicates some East Asian migration to Latin America, particularly after independence in the 19th century30. SOURCEFIND estimates East Asian ancestry in the CANDELA sample to be, on average, very low (<1%) in Brazil, Chile, Colombia, Mexico, and slightly higher in Peru (1.4%). In individuals with >5% East Asian ancestry, this component is inferred to be most closely related to the Chinese and to a lesser extent the Japanese, except in Brazil where the opposite is found (Supplementary Fig. 7). These results match historical records documenting the arrival of Chinese laborers to Peru since the middle 19th century40 and Japanese laborers to Brazil since the early 20th century41. Reflecting the relatively recent nature of these events, GLOBETROTTER estimated dates for admixture involving an East Asian source were significantly more recent than those involving Iberian sources (median = 3, IQR 2–5 generations ago, Wilcoxon rank-sum test one-sided p-value < 1 × 10–15; Fig. 3e).