DNA and the Sephardic Diaspora:
Spanish and Portuguese Jews in Europe
By Abraham D. Lavender Ph.D.
from HaLapid Winter, 2003
The study of genetics, and especially the concept of DNA testing, has gained much interest in recent years in the medical and forensic areas. This interest also is beginning to increase in the area of genealogy and in several areas of academia. Since the mid-1990s, I have introduced my classes in ethnicity to the research of L. Luca Cavalli-Sforza, Paolo Menozzi, and Alberto Piazza, The History and Geography of Human Genes (Princeton University, 1994). These authors compiled charts (similar to road mileage charts showing the distance between cities) showing the genetic distance between numerous ethnic and nationality groups throughout the world. Using somewhat abstract numbers, one can see how Italians, French, Spanish, Portuguese, English, Polish and twenty other European nationality groups are related to each other genetically. They also have charts for Africans, Asians, Native Americans, and other nationality groups.
The use of genetics (DNA) to help understand one’s own ethnic, religious, nationality, and/or racial ancestry, also is increasing. This article first discusses some major concepts of DNA testing, and then discusses the use of genetics as it applies to descendants of the Jews of Spain and Portugal (Sephardim) and the tremendous Sephardi diaspora that developed as a result of the Inquisition in Spain and Portugal. Because the male Y-chromosome mutates more rapidly than the comparable female mtDNA, the male Y-chromosone is used for analyzing more specific genealogical patterns, but significant mtDNA research also is being conducted, and the future holds tremendous potentials for females and males (e.g., Mark G. Thomas et al, “Founding Mothers of Jewish Communities: Geographically Separated Jewish Groups Were Independently Founded by Very Few Female Ancestors,” American Journal of Human Genetics, Volume 70, 2002, pages 1411-1420).
Sophisticated DNA tests today give results for 25 markers, that is, 25 numbers representing specific points tested on a male’s genetic chain. Each marker or point has one number, which is from a possible range of about ten numbers (markers vary in the range of possible numbers). Other tests are based on a smaller number of markers tested, frequently 12 or 9. The more markers two males match on, the more recently they had an exact common male ancestor, and the more closely they are related. If two males match, for example, on nine out of nine markers, then they probably had a common male ancestor, but one has to go back a good number of centuries to reach the common male ancestor. If two males match exactly on 25 markers, then they are very closely related, and the common male ancestor is probably within the last several generations.With nine markers, one probably goes back to shortly before the Inquisition to have a 50% chance of having a common male ancestor, although the frequency of mutations is still debated (see Lev A. Zhivotovsky et al, “On the ‘Evolutionary’ Mutation Rate at Y Chromosome STRs,” paper presented at the 2002 Meetings on Human Origins and Disease, November 3, 2002, Cold Spring Harbor Laboratory, New York).
Whatever the number of markers used, if two males match on all the markers except one, and they are only slightly off on that one, they are said to be a one step mutation match. Results also frequently are given for two step mutation matches. If several males share a great-great-great grandfather, for example, then most of them will have the exact same numbers on all 25 markers, but one or more might differ on one marker (or, rarely, two) because of a mutation that occurred in one line of descendants. The further the distance from the common male ancestor, the more likely there are to be mutations on several markers, and therefore the less exact matches the two males will have. Males frequently will begin with nine or twelve markers, and if a relationship is found there, then they will upgrade to a more expensive 25 marker test to see how more closely/recently they are related.
When a male is tested for any number of markers, for example, nine markers, his nine numbers (representing the exact number for each of the nine markers), is referred to as his haplotype. A 9-marker haplotype could be, for example, 14 13 29 23 11 13 13 11 14. Many males, of course, can share a specific haplotype, especially when the number of marker numbers in a haplotype is small. Geneticists then group together a number of haplotypes which share certain genetic patterns, and this grouping or cluster of haplotypes is referred to as a haplogroup.
When a male is tested by a major testing company, such as FamilyTreeDNA, for example, he is given his haplotype, and the number of other men who match with him exactly on all markers tested, the number who match with a one step mutation, and the number who match with a two step mutation. If other men who matched have given permission for their names and contact addresses to be given, then the males can contact each other and pursue more information. FamilyTreeDNA, for example, with a database of over six thousand people and growing, also asks men to give their paternal country of origin if known. In some cases, people are identified as Ashkenazic or Sephardic. However, in a family which was Jewish but left Judaism several centuries ago, many descendants will not be aware of an earlier Jewish origin going back fifteen or so generations, so the number of “Jewish distant cousins” is likely to be understated. But, FamilyTreeDNA makes a major contribution to Jewish research by giving results for people who have known Jewish ancestry.
There also are several databases on the web where a person can type in his haplotype, and find out how many exact matches he has in different categories. In the United States Database, he is given the number (no names) of exact matches he has among European-Americans, African-Americans, and Hispanics as self-defined by men tested (see http://ystr.org/usa). There also is an Asian database, which includes Turkey, Egypt, Syria, and eleven other countries which might be of interest to Sephardim (http://ystr.org/asia). Unfortunately for Sephardim, there is not a database for North Africa or for sites of heavy Sephardic settlement in the Caribbean and the Americas. But, to get an idea of one’s “distant cousins” in Europe, the European Database is most helpful (http://ystr.charite.de). Click on “Help” if you need help inserting numbers; if you see a box of comments when trying to type in a number, simply click on o.k.).
The European Database is compiled by receiving results for nine markers from labs which, after meeting stringent testing requirements, send their results to be added to the European Database. It is important to note that these site results are nearly always based on random samples taken from the geographical area, and do not come from criminal institutions. Even if an occasional “criminal” gets in, he might easily represent scores of close and distant noncriminal “cousins.” with the same haplotype. Results for nearly all of the sites have been published in scientific journals (see http://ystr.charite.de), then click on “The Forensic Y-User Groups” to see names of researchers and e-mails to contact them. I have e-mailed a number of researchers for further details, and most have responded promptly.
The European Database grows as more lab results are received. In June 2002, the database had results from 72 sites (a city, region of a country, or small country) with a total of 9,607 men. By mid-November, there were 82 sites with a total of 12,259 men. This actually includes three sites in Latin America with European-based populations: Sao Paulo (Brazil), the Antioquia area of Colombia, and Argentina. Unfortunately, data from the Recife area of Brazil, where many Sephardim settled and stayed as Catholics after an expulsion in 1654, can not be a part of the European Database because a different method of testing was used. There is a “Portuguese” database which includes sites in Iberia and Latin American, but the numbers are not comparable to those in the European Database.
The sites currently averages 50 people per site (12,259 divided by 82), but the numbers currently vary from 44 for Friesland to 510 for Sweden (I have combined six small sites in Norway into one site with a total of 300). The smaller the sample size, the more one should view the results as suggestive rather than descriptive. One can compute a percentage for each site, and see where he has the most matches adjusted for sample size. Roewer et al note that the European Database “should be equally useful in forensic analysis and anthropological or archaeological research” (“Online Reference Database of European Y-chromosomal Short Tandem Repeat (STR) Haplotyes,” Forensic Science International, Volume 118, 2001, pages 106-113). The authors also give a very good history and description of the database.
The haplotype I gave above as an example (14 13 29 23 11 13 13 11 14) is the second or third largest haplotype in western Europe. Therefore, it is not surprising that there will be many matches in Europe for this haplotype. Most men will not have the number of matches this haplotype has. In fact, for a number of haplotypes, there will be only a few or even zero matches in the European Database with “only” 12,259 men so far in the database. For examples of haplotype frequencies, see, for example “Distribution of Y-chromosome STR Defined Haplotypes in Iberia” (Annabel Gonzelez-Neira et al, Forensic Science International, Volume 110, 2000, pages 117-126; Mercedes Aler et al, “Y-Chromosome STR Haplotypes From a Western Mediterranean Population Sample,” Volume 119, 2001, pages 254-257).
The percentage of matches for the above haplotype was 1.35 for Spain, from which many Jews were allowed to leave from 1391 on. Although moderate or small, there were matches in seven out of eight sites in Spain (in descending order, Madrid, Asturias, Valencia, Barcelona, Galicia, Zaragoza, Andalusia, and, with no matches, Cantabria). Percentages were 1.79 for southern Portugal, and 3.82 for central and northern Portugal where most Jews were trapped from 1497 on when they were not allowed to leave without secretly escaping.
The percentages for other European areas are: 4.80 for Belgium and 4.48 for southern Netherlands (Holland and Zeeland) where major Sephardic communities existed (especially Amsterdam, referred to as “The New Jerusalem”); 2.18 for the northern Netherlands (Friesland and Groningen) where smaller Sephardic communities existed; 0.87 for Hamburg (where a Sephardic community was begun, but later lost numbers to Copenhagen), 2.41 for five sites in western Germany (Dusseldorf, Cologne, Limburg, Mainz, Freiburg) on or very close to the Rhine River whose mouth is in the Netherlands, 6.06 for Strasbourg, on the Rhine River, on the French-German border, which was referred to as “The Jerusalem of Germany,”1.45 for seven sites elsewhere in Germany (Munster, Magdeburg, Rostock, Berlin, Leipzig, Stuttgart, Munich), 1.35 for Rome, 1.22 for four sites in northern Italy, and 0.27 for five other sites in Italy. London, England and Dublin, Ireland, both of which had Sephardic communities, had respectively 2.83 and 1.87.
There was a 6.35% match (to be interpreted cautiously because of a small sample size) for Copenhagen, Denmark, which had a Sephardic community, but for Scandinavian countries without Sephardic communities, there was 0.67% for Norway and 0.57% for Sweden. Finland was 0%. The percentage was 0.62 for eight sites in Poland, and 0.16 for four sites in Eastern Europe (Lithuania, Estonia, Latvia, and Moscow). Budapest, Hungary, earlier part of the Ottoman Empire with a Sephardic community, had 1.71, Croatia, with a Sephardic community, had 2.00, and Krakow, Poland, also with a Sephardic community, had 0.93 (see Chaim Raphael, The Sephardi Story, Valentine Mitchell, 1991, pages 128-129, for a map of some communities).
Large numbers of Sephardim went to Greece and Turkey, but there were no matches for this haplotype in Greece and a small match in Turkey.. But, we must remember that we are dealing with one case study, and so we cannot assume that “distant cousins from one very large extended family” went to every Sephardi area of settlement. Even a large “family” will not necessarily spread to every settlement; it is the overall pattern that is important. Even if this haplotype did go into exile in Greece or Turkey, however, we must remember that about 95% of the Spanish-Portuguese Jews in Greece (mostly in Salonika) were killed by the Nazis, and that after 1948 large numbers of Sephardim left Turkey for Israel and other places, its Jewish population (about 95% Sephardic) decreasing from 79,765 in 1945 to about 20,000 today (Daniel J. Elazar et al, The Balkan Jewish Communities, University Press of America, 1984; and Esther Benbassa and Aron Rodrigue, The Jews of the Balkans, Blackwell, 1995). There were also a few situations difficult to explain. Vienna and Graz in Austria had no matches, which was no surprise (although Vienna had a Sephardic community), but the Tyrol area of Austria, the Austrian panhandle sandwiched between southern Germany, Northern Italy, and eastern Switzerland, including Innsbruck, had a match of 2.62. I have been unable so far to find references to Sephardic settlements in that area. Lausanne, Switzerland, with 4.43%, is in a similar situation. It is possible that Huguenots from France, previously Sephardim, might account for a flow into Switzerland from France, but I do not know about these two sites, and welcome facts or suggestions. We also must know whether the community has changed significantly in genetic makeup in the last few centuries because of migration or other factors.
In general, the regions where I had large matches are known to have had Sephardi settlements after the exiles from Spain or Portugal. For references to some sites, in addition to Raphael above, see; Miriam Bodian, Hebrews of the Portuguese Nation, Indiana University Press, 1997; John Edwards, The Jews in Christian Europe 1400-1700, Routledge, 1998; Jonathan I. Israel, European Jewry in the Age of Mercantilism 1550-1750, Clarendon Press Oxford, 1985; and Haim Beinart, editor, The Sephardi Legacy, Magnes Press, 1992.
It is possible that a bottleneck, a number of gene patterns being greatly reduced because of a specific event, or genetic drift (random changing of genes over time), can lead to apparent matches that are not real or not as close as they appear. But, there is one major reason that causes hesitation in concluding that all of this data represents a Sephardic migration pattern after the Inquisition era. This can be referred to as the “ancient history” thesis. It is possible that these DNA patterns go back thousands of years to very early migrations to western Europe, and long predate a Jewish presence in western Europe (for a discussion of early genetic patterns in Iberia, see J. Bertranpetit and L.L. Cavalli-Sforza, “A Genetic Reconstruction of the History of the Population of the Iberian Peninsula,” Annals of Human Genetics, Volume 55, 1991, pages 51-67). This haplotype does have matches in areas of Europe other than areas of major Sephardic settlements. Similar patterns of varying degrees have been found by this author in analyses of twenty other haplotypes from Iberia.
The ancient history approach could be used to argue that the pattern discussed in this paper pre-dates the Inquisition-based Sephardic exile. Another suggestion is that Sephardic descendants of the exiles, in the 600 years since significant migrations began from Spain (migrations started in 1391, following a wave of persecution), have spread to many places within Europe, and that many of the descendants are today Christians who have no memory of their Jewish ancestry. We know from a number of writings that a significant number of Jewish descendants have been lost to the community in this manner. Certainly in Germany there have been major Jewish conversions to Christianity for several hundred years, and there are many descendants of Jewish families who have lost all traces of their Jewish ancestry. The pattern shown in this article, with the specific relationship between this DNA pattern and known Sephardic areas of settlement, is too strong to support an ancient history pattern of genetic distribution. Even taking into account the matches in areas with unknown Sephardic settlements, as in parts of Germany, the overall pattern is too strong to have happened by chance. Although I approach this conclusion with double caution because I share the haplotype analyzed, I do think that the pattern indicates that this haplotype is part of the Sephardic diaspora.
Other findings that deserve more attention come from an analysis which I have done using DNA patterns from a recent Israeli study which used six markers (Almut Nebel et al, “The Y Chromosome Pool of Jews as Part of the Genetic Landscape of the Middle East,” American Journal of Human Genetics, Volume 69, 2001, pages 1095-1112). Five of these markers are also in the European Database, and I used these five for analysis. Out of a total of 526 men, the Israeli study had 78 men who identified themselves as Sephardic (over 70% identified as North African), and 79 men who identified themselves as Ashkenazic. Although these findings are preliminary, using samples from North African Sephardim and Ashkenazim, the Sephardi matches in Europe were, in descending order, in Albania, Bulgaria, Umbria (Italy), Budapest (Hungary), Latium (Italy), Marche (Italy), Freiburg (Germany), Mainz (Germany), Friesland (Netherlands), Lombardy (Italy), Tyrol (area of Innsbruck, Austria), Cologne (Germany), Denmark, Stuttgart (Germany), and Holland (Netherlands). On the other hand, the matches for those who identified as Ashkenazi were Novgorod (Russia), Wroclaw (Poland), Moscow (Russia), Warsaw (Poland), Graz (Austria), Groningen (Netherlands), Ljubljana (Slovenia), Eastern Norway, Bydgoszcz (Poland), Magdeburg (Germany), Lithuania, Romani of Bulgaria, Romani of Baranya area in Hungary, Dusseldorf (Germany), Munster (Germany), and Krakow (Poland). Although these five markers go back a number of centuries, the Sephardic pattern bears a striking similarity to that of the haplotype used in this study, and a marked contrast to the Ashkenazic pattern, despite conclusions that both Ashkenazi and Sephardi share Near Eastern origins. It seems to this researcher that several Ashkenazim had a Sephardi genetic pattern, and that several Sephardim had an Ashkenazi pattern, probably due to people being absorbed over time into the other culture through migration and/or marriage. Without these cases, the Sephardi-Ashkenazi distinction would be even greater.
There have been conversions to Judaism (and other forms of gene input) throughout history among both Ashkenazim and Sephardim, and a maternal based Jewish status has brought at least some “new” male Y-chromosomes into Jewish communities going back centuries. Much more research is needed for both Ashkenazim and Sephardim, and DNA research must be approached with an open mind and integrated with knowledge of a group’s specific history, culture, and migration patterns. Why Spanish-Portuguese Jews went into exile, how many went, where they went, how many were left in Spain or Portugal, and how many immediately or eventually left Judaism must be integrated into DNA research. SCJS can play a major role in this area.
Abe Lavender PhD, a sociologist, specializes in ethnicity and the Jewish community, with a special interest in Sephardim. One of his courses is World Jewish Communities.
Dr. Lavender expresses appreciation to Dr. Rene Herrera, geneticist, Department of Biological Sciences, Florida International University, and to Reese Mozingo of Raeford, North Carolina, for their assistance and takes full responsibility for all interpretations and positions taken on still controversial issues.
Society For Crypto Judaic Studies