Jewish Genetics: Abstracts and Summaries

Part 5: Karaites

Introduction to Karaism and Karaites

       The Karaites represent an autonomous stream within Judaism. Karaite Judaism and Rabbinical Judaism have differences in their dietary laws, their calendar systems, their methods of observing the holidays of Rosh Hashanah and Sukkot, and their wedding ceremonies and contracts. Karaites don't observe the Rabbinical holiday of Hanukkah. The most significant difference between the two forms of Judaism is that Karaism rejects the legal authority of Rabbinical writings like the Talmud and Mishnah, preferring to use their own scholars' books of biblical interpretation. Over the centuries, Karaite followers have lived in places as varied as Iraq, Egypt, the Byzantine Empire, Spain, and Eastern Europe.

       The origins of Karaism are still debated by historians and there have also been many theories about the ethnic backgrounds of its followers. The earliest appearance of East European Karaites was documented in the travelog of Rabbi Petakhya from Regensburg (1177-1187 C.E.) who met them in "the land of Qedar". Some of the pathways of Karaite people's migrations are documented, including the movement of Karaites from Constantinople/Istanbul in modern-day Turkey to the Crimean peninsula in modern-day southern Ukraine.$1$ The Crimean Karaite community, in turn, has well-documented connections with the smaller Karaite communities in western Ukraine, Poland, and Lithuania.

       Karaites have lived in such East European towns and cities as Chufut-Kale, Sulkhat, Bakhcheserai, Halicz, Lwów, Kukizów, Troki, Warsaw, Odessa, and Kiev under such governments as the Golden Horde, the Lithuanian Grand Duchy, the Ottoman Empire, the Austrian Empire, and the Russian Empire.

       Traditionally, the East European Karaites viewed themselves to be descended from Jewish people living in ancient Israel and this belief was reinforced by the Karaites' main religious text (the Torah) and their liturgy. However, in the 20th century an alternative point of view emerged suggesting that East European Karaites descend from one or more Turkic-speaking tribes that allegedly converted to Karaite Judaism, such as Khazars or Kipchaks. This latter theory is sometimes used to explain the Turkic language that became a native language for East European Karaites.

       Rabbinical Judaism is the predominant form of Judaism in modern times. Practicioners of Rabbinical Judaism who descend from the Israelite people are culturally divided into several groups, including the Ashkenazi Jews, the Sephardi Jews, and the Mizrahi Jews (though the distinction between the last two is sometimes blurred). Ashkenazi Jews are those Jews who settled in Central Europe and Eastern Europe. Sephardi Jews developed their culture in Spain and Portugal but later were scattered to countries like Syria, Bulgaria, and Morocco. Mizrahi is the Hebrew term for "eastern" Jews such as the Jews of Iraq, Iran, and Uzbekistan.

       Intermarriage between Karaites and Ashkenazi Jews seldom occurred. No such mixed marriages occurred in the Austrian Empire despite long-standing settlements of both communities in its Galician region, but some mixed marriages happened in the Russian Empire and the Soviet Union$2$ and examples are known from history where marriage contracts contained special clauses to accommodate Karaite-Rabbanite marriages.$3$ A small number of Rabbanites converted to Karaism.$4$

       Karaite Jews mainly use descent through the male line as the way to determine a person's automatic Jewish or non-Jewish religious status. The same used to be true of Rabbinical Judaism until about the 4th century B.C.E. to the 2nd century C.E. when matrilineal descent became the Rabbinical standard.

       Some men in both the Karaite and Rabbinical communities carry a special title inherited from father to son. Cohen designates a descendant of the priests who served in the temple in Jerusalem in ancient times. Levi is the name of an ancient Israelite tribe whose members once served as assistant priests.

Purpose of the Study

       In our research, we are trying to use DNA as an independent source of information about the origins of East European Karaites. As there is evidence that some beliefs about Karaite ancestry were constructed and perpetuated for purely political reasons, the goal of the historian and scientist must be to examine the documents and DNA of this people on a non-partisan basis.

Methodology

       Our genetic study began in March 2005. It was conducted using the services of Family Tree DNA, a business division of Gene by Gene, Ltd., formerly known as Genealogy by Genetics, Ltd., in Houston, Texas. Samples were obtained from Karaites living in the United States, Canada, Israel, and the Crimean peninsula of Ukraine (now part of Russia). All participants were asked about their paternal and maternal ancestries both in terms of ethno-religious identifications and geographic origins.

       We sampled 22 East European Karaite men and 5 East European Karaite women: a total of 27 individuals with deep roots in the East European Karaite communities. We also sampled one Karaite man whose family had immigrated and integrated into the Crimean Karaite community but was not originally from Europe. Except for this man's paternal line, only East European Karaite lineages were tested. Some of our participants were children of mixed marriages between a Karaite and a non-Karaite. The women could only be tested on their maternal lines since the Y chromosome is not transmitted from fathers to their daughters. As a result of these circumstances, 13 participants were sampled only on their paternal lineages (Y-DNA), 7 only on their maternal lineages (mtDNA), and the remainder (8) on both. We evaluated 21 Y-DNA samples (19 from Crimean Karaite lines, one from a Lithuanian Karaite line, and one from an immigrant family to the Crimea) and 15 mtDNA samples (14 from Crimean Karaite lines and one from a Galician or Halych Karaite line) altogether. One of the Crimean Karaite Y-DNA lines is Levite and another is Cohen.

       The samples were processed at the Arizona Research Labs of the University of Arizona in Tucson, Arizona.

Results

       The following summarizes the results obtained to date as of September 28, 2014.

       A Y haplogroup is a group of haplotypes with the same set of single nucleotide polymorphism (SNP) mutations. Its members may share a common ancestor within several generations or more than a thousand years ago. Y haplogroups are only transmitted in unbroken lines of male descent from fathers to sons. Names of Y-DNA haplogroups change from time to time. We are using the names that were most current at the time of this writing as defined by the April 25, 2014 revision by the Y Chromosome Consortium (YCC), which is used by Family Tree DNA.$5$ We also consulted the August 11, 2011 revision of the Y-DNA Haplogroup Tree by the International Society of Genetic Genealogy (ISOGG),$6$

       Different members tested their Y chromosomes to different short tandem repeat (STR) levels. The more STRs there are, the more precision in relative-matching is available. Whereas our first 8 Karaites were initially tested only at the 12-marker level (Y-DNA12) in 2005-2006, all of them along with subsequent Karaite samples have since been upgraded to the 37-marker level (Y-DNA37) and some have been further upgraded to the 67-marker level (Y-DNA67).

       An mtDNA haplogroup is defined by a common set of mutations in reference to the revised Cambridge Reference Sequence. mtDNA haplogroups are typically older than their male counterparts. mtDNA haplogroups are transmitted in unbroken lines from mothers to daughters, although the final recipient in the chain of descent can be either a daughter or a son. For the phylogenetic tree of mitochondrial haplogroups, we use Mannis van Oven and Manfred Kayser's Phylotree.

       Family Tree DNA's extensive database enabled us to find many matches for our samples.

       We begin our accounting with the Y-DNA results. Sample K05, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J1 (J-M267). At the 25-marker level he is one step mutation from an Ashkenazi from Romania, an Ashkenazi from Belarus, and a man from Ukraine and two step mutations from an Ashkenazi from Moldova. At the 12-marker level he is one step mutation from an Ashkenazi from Poland as well as from an Ashkenazi Cohen from Belarus and other men mostly from Europe.

       K04, whose paternal line comes from the Crimea, identifies as Levite, and has the surname Levy, belongs to Y-DNA haplogroup L2a (L-M349). At the 25-marker level he is two step mutations from an Egyptian Karaite, two Ashkenazim from Moldova, and one Ashkenazi from the Russian Federation. The Y-DNA TiP Report told us based on the 25-marker data that the probability that our Crimean Karaite shares a common ancestor with one of the Moldovan Ashkenazim is about 90% within 24 generations and about 83% within 20 generations and the same probabilities hold for his relationship to the other Moldovan Ashkenazi, whereas his relationship to the Egyptian Karaite is about 89% within 24 generations and about 80% within 20 generations (rounding to the nearest whole percents). At the 12-marker level he is an exact match with a Sephardi from Turkey and one step mutation from a different Sephardi from Turkey plus some other men.

       K24, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M172) and is an exact match with an Ashkenazi from the Russian Federation at the 12-marker level.

       K16, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup E1b1b1 (E-L117). He is one step mutation from a man from France and an Italian from Sicily and two step mutations from one man each from Spain and the United Kingdom at the 25-marker level. At the 12-marker level he is an exact match with two Ashkenazim from Poland, an Ashkenazi from the Russian Federation, two Ashkenazim from Ukraine, two other people from Russia, two others from Poland, one other from Ukraine, eleven from Germany, one from the Netherlands, one from Italy, one from Cyprus, two from England, one from Scotland, two others from the United Kingdom in general, one from the United States generically, and one from Switzerland.

       K02, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M172). He is two step mutations from another Crimean Karaite in our study (sample K01) at the 37-marker level. He is four step mutations from a Frenchman at the 37-marker level. At the 25-marker level he is two step mutations from an Italian. He matches all 12 markers with a Syrian Arab. He is one step mutation from an Armenian, two Hungarians, two Saudi Arabians, and one man from Turkey, among others, at the 12-marker level.

       K01, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M172). He is two step mutations from our sample K02, as stated above. He is four step mutations from a confirmed Frenchman as well as from a man with the same surname as the confirmed Frenchman but with one less letter. At the 12-marker level he is one step mutation from a Syrian Arab and a British man.

       K06, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup G2a (G-P15). He is three step mutations from another Crimean Karaite in our study (sample K18) at the 37-marker level and exactly matches him at the 25-marker level. At the 25-marker level he is two step mutations from sample K12, another Crimean Karaite.

       K18, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup G (G-P15). He is three step mutations away from Crimean Karaite sample K06 at the 37-marker level, as stated above, and they are an exact match at the 25-marker level. At the 25-marker level he is two step mutations from Crimean Karaite sample K12.

       K12, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup G2a3b1 (G-P303). At the 25-marker level he is two step mutations from Crimean Karaite samples K06 and K18 as well as one Ashkenazi from Hungary, one Ashkenazi from Germany, one man from the Russian Federation, and one man from Slovakia.

       K10, whose paternal line comes from the Crimea, identifies as Cohen, and has the surname Kogen, belongs to Y-DNA haplogroup J1 (J-CTS11741, but he was formerly listed as J-M267). He exactly matches an Ashkenazi Cohen from Belarus (surnamed Kagan) at the 12-marker level but does not match him at the 25-marker level. His one step mutation matches at the 12-marker level are overwhelmingly Ashkenazim and Sephardim, including some Ashkenazi Cohens and Ashkenazi Levites. His only match listed at the 25-marker level is a man with roots from Ukraine whose ethnic identity isn't specified, and this match is at a genetic distance of two step mutations.

       K15, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M172). He is an exact match with an Ashkenazi from Russia at the 12-marker level. At the 37-marker level he is one step mutation from a man from Ukraine.

       K21, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J1 (J-M267). At the 12-marker level he is one step mutation from Ashkenazim (including but not limited to Cohens), Sephardim (including those from Syria, Spain, Lebanon, and Italy, and some of these Sephardim are Cohens), two Mizrahim from Afghanistan, and a Mizrahi from Kurdistan, among others, with some of those matches being men from Armenia, Saudi Arabia, and Yemen. At the 25-marker level he is two step mutations from a man from Russia.

       K19, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup E1b1b1 (E-L795), which Family Tree DNA is currently reevaluating as he may belong more specifically to its branch E1b1b1b2a1a3, which is also found among multiple Arabs and one German Jew who have tested with that company. At the 37-marker level he is three step mutations from a Karaite from Egypt and according to the Y-DNA TiP Report the probability that they share an ancestor within 20 generations is about 96% and within 24 generations nearly 99% (rounding to the nearest whole percents). At the 12-marker level he exactly matches a different Karaite from Egypt as well as an Ashkenazi from Belarus and a man from the United Arab Emirates plus a man from Bulgaria and ten men from Switzerland. At the 25-marker level he is two step mutations from a man from Jordan and a man from Germany.

       K14, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup E1b1b1 (E-M78). His exact matches at the 12-marker level include an Ashkenazi from Germany as well as men from many other countries including one from Armenia, one from Austria, one from Hungary, two from Italy, two from Poland, and two from the Czech Republic, among others. He has further Ashkenazi matches one step mutation away at the 12-marker level (four Ashkenazim from Belarus, three Ashkenazim from Lithuania, and three Ashkenazim from the Russian Federation) but many more matches at that step who don't list Jewish origins including but not limited to 17 men from France, 22 from England, 8 from Italy, four from Greece, one from the Czech Republic, one from Kuwait, two from Saudi Arabia, and four from the United Arab Emirates. At the 25-marker level he is two step mutations from two men from Germany and one man from England as well as a man with a surname that stems from English and German origins, a man whose most distant ancestor in this lineage had English first, middle, and last names, and a man with Arabic first and last names who did not list his country of paternal origin and did not answer our email requesting this information.

       K07, whose paternal line comes from Lithuania,$7$ belongs to Y-DNA haplogroup G2a (G-P15). At the 12-marker level he is one step mutation from two Englishmen, two men from Italy, one man from Poland, one man from Ukraine, and one man from Germany.

       K09, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup E1b1b1a1c (E-V22). At the 12-marker level he is one step mutation from a Mizrahi from Iraq, a Saudi Arabian, an Albanian, and one man from Belgium.

       K20, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M67). At the 37-marker level his exact matches are two men from the Chechnya region of the Russian Empire whose surname is identical. One step mutation away he matches a man from the Republic of Ingushetia in the Russian Federation whose surname is found among ethnic Ingush people. Two step mutations away he matches another man from Ingushetia as well as one man from the Republic of Georgia. Three step mutations away he matches yet another man from Ingushetia. Four step mutations away he matches three men from Chechnya and two men from Ingushetia. The Y-DNA TiP Report told us based on the 37-marker data that the probability that our Karaite shares a common ancestor with the two Chechens who exactly match him at that level is about 83% going back 4 generations, about 97% going back 8 generations, and nearly 100% going back 12 generations (rounding to the nearest whole percents). The report's probabilities between him and the Ingush who is one step away from him are about 89% going back 8 generations, about 98% going back 12 generations, and nearly 100% going back 20 generations (rounding to the nearest whole percents). At the 25-marker level he is one step mutation from a man from Armenia who doesn't show as as a match at the 37-marker level and two step mutations from men from Georgia and Ingushetia who don't show as matches at the 37-marker level. One man from Slovenia shows as an exact match as the 12-marker level. Three Sephardim and two Ashkenazim show as very distant matches, at the 12-marker level with one step mutation, but so too do dozens of non-Jews from countries throughout Europe, the Caucasus, and the Middle East including but not limited to Iran, Kuwait, Lebanon, Saudi Arabia, Italy, Spain, England, Wales, France, and the Netherlands.

       K03, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J1 (J-M267). At the time of this writing, he does not match with any other man in the Y-DNA database. He belongs to a genetic cluster that also includes two men from Italy, one man from Germany, and one man from the United Kingdom, but he is related to them only extremely distantly, hence why none of them show as matches for him even at the 12-marker level with one step mutation.

       K26, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup R1b1a2 (R-M269). At the time of this writing, he does not match with any other man in the Y-DNA database.

       K27, whose paternal line comes from the Crimea, belongs to Y-DNA haplogroup J2 (J-M172). At the time of this writing, he does not match with any other man in the Y-DNA database.

       Maternal haplogroups are distinct from paternal haplogroups, so, for example, maternal haplogroup H has nothing to do with paternal haplogroup H, even though they share the same nomenclature. HVR1 stands for hypervariable region 1 and is the low resolution level. HVR2 stands for hypervariable region 2. The combination of HVR1 and HVR2 provides an intermediate resolution level. CR stands for the Coding Region; the combination of HVR1, HVR2, and CR provides complete coverage of mitochondrial DNA. Full matches using HVR1+HVR2 tests and HVR1+HVR2+CR tests give ranges of hundreds or thousands of years for the TMRCA (Time to the Most Recent Common Ancestor).

       K22, a woman whose maternal line comes from the Crimea, belongs to mtDNA haplogroup I. Her HVR1 mutation values are C16148T, T16187C, C16189T, G16230A, T16278C, C16311T, G16391A (they were formerly designated as 16129A, 16148T, 16223T, 16391A, 16519C) and her HVR2 mutation values are C146T, C152T, C195T, T199C, A247G, T250C, C494a, 522.1A, 522.2C, 309.1C, 315.1C, 573.1C, 573.2C (they were formerly designated as 73G, 199C, 250C, 263G, 309.1C, 315.1C, 494A, 573.1C, 573.2C). Her HVR1 matches include a Sephardi from Algeria, three Sephardim from Morocco, one Bedouin from Israel, and one Lebanese, but also non-Jewish people from Europe: one from Belarus, one from Poland, one from Hungary, one from the Czech Republic, one from Austria, three from France, three from Germany, four from Ireland, three from England, three from Scotland, three other people from the United Kingdom in general, and one from Canada generically. At the HVR1+HVR2 level she matches two Sephardim from Portugal, one Sephardi from Spain, and one Crimean Karaite (sample K13); these Portuguese and Spanish matches added their Sephardi designations to the database between January and September of 2014.

       K13, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup I. His HVR1 mutation values are C16148T, T16187C, C16189T, G16230A, T16278C, C16311T, G16391A (they were formerly designated as 16129A, 16148T, 16223T, 16391A, 16519C) and his HVR2 mutation values are C146T, C152T, C195T, T199C, A247G, T250C, C494a, 522.1A, 522.2C, 309.1C, 315.1C, 573.1C, 573.2C (they were formerly designated as 73G, 199C, 250C, 263G, 309.1C, 315.1C, 494A, 573.1C, 573.2C). His HVR1 and HVR2 matches are the same ones that sample K22 has and, as stated, samples K13 and K22 match each other at the HVR1+HVR2 level.

       K05, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, C16169T, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16320T (they were formerly designated as 16169T, 16320T, 16519C) and his HVR2 mutation values are G73A, A93G, C146T, C152T, C195T, A247G, 315.1C (they were formerly designated as 93G, 263G, 315.1C, 522-, 523-). His matches at the HVR1 level include one Mizrahi from Georgia (Georgian Jew) and one Ashkenazi from Ukraine. His matches at the HVR1+HVR2 level include a Crimean Karaite (sample K02), an Azerbaijani, one person from Belarus with a Jewish middle name and an Ashkenazic last name who did not respond to our query to confirm whether his maternal line is Ashkenazic, and one person from Lithuania with a surname found among Ashkenazim and Germans who did not respond to our query to confirm whether his maternal line is Ashkenazic.

       K02, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, C16169T, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16320T (they were formerly designated as 16169T, 16320T, 16519C) and his HVR2 mutation values are G73A, A93G, C146T, C152T, C195T, A247G, 315.1C (they were formerly designated as 93G, 263G, 315.1C, 522-, 523-). His HVR1 and HVR2 matches are the same ones that sample K05 has and, as stated, samples K02 and K05 match each other at the HVR1+HVR2 level.

       K03, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H9a. His HVR1 mutation values are A16129G, C16168T, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T (they were formerly designated as 16168T, 16519C) and his HVR2 mutation values are G73A, C146T, C195T, A247G, 522.1A, 522.2C, 315.1C (they were formerly designated as 152C, 263G, 315.1C). His CR mutation values are A769G, A825t, A1018G, G2706A, A2758G, C2885T, G3591A, T3594C, G4104A, A4310G, T4312C, T4553Y, T7028C, G7146A, T7256C, A7521G, T8468C, T8655C, G8701A, C9540T, C10334T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, T13020C, G13105A, G13276A, T13506C, T13650C, T14766C. Most of his matches at the HVR1 level are Europeans but he also matches six people from Lebanon, one from Syria, and one from Turkey. His matches on HVR1+HVR2 all come from European countries: a Crimean Karaite (sample K25), three people from England, one from France, one from Germany, two from Greece, one from Hungary, five from Italy (including a Sicilian), two from the Netherlands, one from Poland, one from Scotland, one from Slovakia, one from Switzerland, and two from Ukraine. His closest matches on his full genomic sequence are a person from England and a person from Spain who match him at the genetic distance of -2. His other reported matches on his full genomic sequence are two people from Germany, one person from Sweden, one person from Greece, and one person from the United Kingdom, all reported at a genetic distance of -3 from him.

       K25, a woman whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. Her HVR1 mutation values are A16129G, C16168T, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T (they were formerly designated as 16168T, 16519C) and her HVR2 mutation values are G73A, C146T, C195T, A247G, 522.1A, 522.2C, 315.1C (they were formerly designated as 152C, 263G, 315.1C). Her HVR1 and HVR2 matches are the same ones that sample K03 has and, as stated, samples K25 and K03 match each other at the HVR1+HVR2 level.

       K26, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, T16304C, C16311T, C16519T (his HVR1 value was formerly designated as 16304C) and his HVR2 mutation values are G73A, C146T, C152T, C195T, A247G, C456T, 522.1A, 522.2C, 309.1C, 309.2C, 315.1C (they were formerly designted as 263G, 309.1C, 309.2C, 315.1C, 456T). He has numerous HVR1+HVR2 matches including two Ashkenazim from Austria, an Ashkenazi from Hungary, an Ashkenazi from Lithuania, an Ashkenazi from Poland, two Ashkenazim from the Russian Federation, and five Ashkenazim from Ukraine, but also many non-Jews from countries like Greece, Poland, Sweden, and Germany, among others.

       K18, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, T16187C, C16188T, C16189T, T16223C, G16230A, T16278C, C16311T (they were formerly designated as 16188T, 16519C) and his HVR2 mutation values are G73A, C146T, C152T, C195T, A247G, 522.1A, 522.2C, 315.1C (they values were formerly designated as 263G, 315.1C). He has abundant HVR1 matches including one Sephardi from Bulgaria, one Ashkenazi from the Netherlands, and numerous non-Jews from European countries such as seven Italians. His HVR1+HVR2 matches include one person from the Czech Republic, one from England, four from Germany, one from Italy, one from Ukraine, and four named people without geographic indicators as well as another Crimean Karaite in our study (sample K12). In conjunction with data reported above, this means samples K18 and K12 match both maternally and paternally.

       K12, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, T16187C, C16188T, C16189T, T16223C, G16230A, T16278C, C16311T (they were formerly designated as 16188T, 16519C) and his HVR2 mutation values are G73A, C146T, C152T, C195T, A247G, 522.1A, 522.2C, 315.1C (they were formerly designated as 263G, 315.1C). He has the very same HVR1+HVR2 matches as sample K18 does.

       K23, a woman whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. Her HVR1 mutation values are A16129G, T16187C, C16188T, C16189T, T16223C, G16230A, T16278C, C16311T (they were formerly designated as 16188T, 16519C) and her HVR2 mutation values are G73A, C146T, C152T, C195T, T199Y, A247G, 522.1A, 522.2C, 315.1C (they were formerly designated as 199Y, 263G, 315.1C). Her CR mutation values are A769G, A825t, A1018G, G2706A, A2758G, C2885T, T3594C, G3849A, G4104A, T4312C, A4976G, T7028C, G7146A, T7256C, A7521G, T8468C, C8628T, T8655C, G8701A, G8863A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, G13105A, G13276A, T13506C, T13581C, T13650C, T14766C. Her HVR1+HVR2 matches are one person from England, three from Germany, and one from Ukraine. At the HVR1 level she matches numerous people including a Sephardi from Bulgaria, an Ashkenazi from the Netherlands, and others without a Jewish designation. At the time of this writing, she does not match with any other person in the mtDNA database on her full genomic sequence.

       K11, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, T16362C, A16482G, C16519T (they were formerly designated as 16362C, 16482G) and his HVR2 mutation values are G73A, C146T, C152T, C195T, T239C, A247G, 522.1A, 522.2C, 309.1C, 315.1C (they were formerly designated as 239C, 263G, 309.1C, 315.1C). He has numerous HVR1+HVR2 matches including an Ashkenazi from France, an Ashkenazi from Poland, two Ashkenazim from Ukraine, two Ashkenazim from Hungary, three Sephardim from Morocco, two Druze from Israel, and a Palestinian Arab from Israel, but also many non-Jews from Europe such as 41 people from Ireland and 48 from Germany.

       K16, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup H. His HVR1 mutation values are A16129G, G16145A, T16187C, C16189T, T16223C, G16230A, T16278C, C16311T, C16354T, T16356C (they were formerly designated as 16145A, 16354T, 16356C, 16519C) and his HVR2 mutation values are G73A, C113T, C146T, C152T, C195T, A247G, C256T, 522.1A, 522.2C, 315.1C (they were formerly designated as 113T, 256T, 263G, 315.1C). At the time of this writing, he does not match with any other person in the mtDNA database.

       K28, a woman whose maternal line comes from the Galician region of western Ukraine, belongs to mtDNA haplogroup HV-T16311C!. Her HVR1 mutation values are A16129G, T16187C, C16189T, T16223C, G16230A, T16278C, C16519T (her HVR1 was previously listed as 16311C) and her HVR2 mutation values are G73A, C146T, C152T, C195T, A247G, 522.1A, 522.2C, 309.1C, 315.1C (they were formerly designated as 263G, 309.1C, 315.1C). Her CR mutation values are G622R, A769G, A825t, A1018G, A2758G, C2885T, T3027C, T3594C, G4104A, T4312C, G7146A, T7256C, A7521G, T7621Y, T8468C, T8655C, G8701A, C9540T, G10398A, T10664C, A10688G, C10810T, C10873T, C10915T, A11719G, A11914G, T12705C, T12757C, T12879C, G13105A, G13276A, C13305T, T13506C, T13650C, T14766C. She has numerous HVR1+HVR2 matches including a Sephardi from Bulgaria, seven people from Ireland, 29 from Germany, one from Croatia, two from Belgium, one from Austria, one from Ukraine, and four from Scotland, among others. She does not match with any other person in the mtDNA database in the Full Coding Region screen. Because her test was conducted by National Geographic's Genographic Project and she has 4 extra mutations that other GenBank members of her haplogroup don't have, the mtDNA expert Dr. Ian Logan and I were wondering whether they gave her a mitochondrial genome sequence that fell slightly short of being complete or one which was inaccurate, but her daughter subsequently tested with Family Tree DNA's in-house "mtFull Sequence" test and came up with an identical sequence, proving that the mother's test was complete and accurate.

       K17, a woman whose maternal line comes from the Crimea, belongs to mtDNA haplogroup N1c. Her HVR1 mutation values are A16051G, T16086C, A16129G, T16187C, C16189T, C16201T, G16230A, A16265G, T16278C (they were formerly designated as 16051G, 16086C, 16201T, 16223T, 16265G, 16311C, 16519C) and her HVR2 mutation values are G143A, C146T, C152T, A189G, T204C, G207A, A210G, A247G, 522.1A, 522.2C, 315.1C (they were formerly designated as 73G, 143A, 189G, 195C, 204C, 207A, 210G, 263G, 315.1C). At the time of this writing, she does not match with any other person in the mtDNA database.

       K24, a man whose maternal line comes from the Crimea, belongs to mtDNA haplogroup T1. His HVR1 mutation values are T16126C, A16129G, A16163G, C16186T, T16187C, T16223C, G16230A, T16278C, C16294T, C16311T, T16325C (they were formerly designated as 16126C, 16163G, 16186T, 16189C, 16294T, 16325C, 16519C) and his HVR2 mutation values are C146T, C152T, C195T, A247G, 522.1A, 522.2C, 309.1C, 315.1C (they were formerly designated as 73G, 263G, 309.1C, 315.1C). At the time of this writing, he does not match with any other person in the mtDNA database.

Analysis

       We found that East European Karaites are often related to Ashkenazi Jews. Sixteen of our samples (K02, K04, K05, K10, K11, K12, K14, K15, K16, K18, K19, K20, K21, K23, K24, K26) match with Ashkenazim. Within this group, our East European Karaite Cohen matches an Ashkenazi Cohen.

       Also seemingly significant is how two of our East European Karaites (K04, K19) match Egyptian Karaites.

       Eleven of our East European Karaites (K04, K10, K11, K12, K13, K18, K20, K21, K22, K23, K28) match Sephardi Jews. One (K09) matches an Iraqi Jew. One (K21) matches a Kurdish Jew. Two (K02, K05) match Georgian Jews. One (K21) matches Afghan Jews. Many of these relationships, however, are distant.

       Ten of our East European Karaites (K01, K02, K03, K09, K11, K19, K20, K21, K22, K25) match non-Jewish Arabs, and another of our East European Karaites (K14) matches a person with a fully Arabic name who is likely to be a non-Jewish Arab. Two (K13, K22) match Bedouin Arabs. One (K11) matches Druze. Four (K02, K14, K20, K21) match Armenians. Four (K02, K03, K20, K25) match Anatolian Turks. Three of these East European Karaites who match some Eastern Mediterranean and Middle Eastern peoples (K01, K03, K25) did not match any rabbinical Jews in the database so far. (By "Middle East" we refer to the region formerly called the "Near East" in popular Western culture that's also geographically known as "West Asia" and "Southwest Asia", to the exclusion of Central Asia.)

       One of our East European Karaites who has no matches in the database (K17) is a member of the mtDNA haplogroup N1c, which is common in the Middle East, especially Saudi Arabia and Turkey. Another of our East European Karaites with no matches (K27) is a member of the Y-DNA haplogroup J2, which is especially frequent among the inhabitants of Turkey, Armenia, Azerbaijan, Iraq, and Lebanon, and this reveals Middle Eastern origins for him as well. Scientists have determined that the Y-DNA haplogroup family J, to which both J1 and J2 are subclades, originated in the Middle East, and this tells us something important about the paternal ancestry of our many Karaites who carry J1 and J2. Significant, too, is the presence of the Y-DNA haplogroup E1b1b1 in several of our Karaite samples as it also originates in the Middle East.

       Sample K04, an East European Karaite who matches an Egyptian Karaite, also matches Ashkenazi Jews fairly closely. This means the three groups (East European Karaites, Egyptian Karaites, Ashkenazi Jews) have a common origin. This impression is strengthened by the presence of Turkish Sephardi Jews among their more distant matches. This sample's haplogroup, L, originated in South Asia before it migrated to the Middle East.

       Not only East European Karaites and Egyptian Karaites share origins with Rabbinical Jews. By way of comparison, we also had an Azerbaijani Karaite give us access to his Y-DNA results and we saw that he exactly matches 2 Ashkenazi Jews at the 12-marker level, has a large number of Ashkenazi matches at the genetic distance of 1 step mutation from the 12 markers, and matches a man from Iran at the 37-marker level with 4 step mutations.

       The match between sample K10 and an Ashkenazi Cohen from the Kagan family shows that at least some Karaite Cohens are related to Rabbinical Cohens.

       Sample K26 is in the Y-DNA haplogroup R1b1a2 (also called R-M269; formerly called R1b1b2 and before that R1b1c and before that R1b3) which is found in western Europe and at a high frequency among the Bashkirs (who live in the Bashkortostan and Perm regions of the Russian Federation).

       The mtDNA haplogroup H is very commonly encountered among the peoples of Europe, though it is also found (in lesser frequencies) in the northern regions of the Middle East, the southern portions of the Caucasus region, and North Africa. H's parent is haplogroup HV. Haplogroup H9 is a branch of H that's also found among some ethnic groups in Europe and the Middle East but only in low frequencies. Our East European Karaites who carry HV, H, and H9a match Rabbinical Jews (Ashkenazim, Sephardim, and Mizrahim) as well as many non-Jewish European peoples and some non-Jewish Middle Eastern peoples, so it is difficult to assess whether they obtained these haplogroups from recent Middle Eastern ancestry or not. Scientists believe the H family originated in the Middle East prior to its migration into Europe. It is known that Ashkenazic maternal lineages sometimes originated among the peoples of Europe from marriages between Israelite Jews and European women who converted to Judaism, so it could be the case that European women are the source of our Karaites' H family haplogroups.

       Interpretation of the mtDNA haplogroup I could also go in several directions, as our East European Karaites who carry it match with Sephardim and non-Jewish Middle Easterners as well as non-Jewish Europeans. Haplogroup I is present among ethnic groups of Europe, the Middle East, and South Asia. As an example, the Slavic-speaking Lemko Rusyns of Slovakia, Poland, and Ukraine have it in a high frequency.

       We observed that several of our East European Karaite samples had matches, or near matches, in terms of their HVR1 and HVR2 mutation values, with Rabbinical Jews who were tested in a comprehensive mtDNA study published in 2008 by Behar et al.$8$ Our Karaites K02 and K05 have identical control regions (16169 16320 16519 and 93 263) within mtDNA haplogroup H as were found among some Georgian Jews in Behar et al. 2008. This wasn't a surprise to us since we already saw that our Karaites match a Georgian Jew in Family Tree DNA's database. Also of note, the control regions 16129 16148 16223 16391 16519 and 73 199 250 263 within mtDNA haplogroup I were found among some Moroccan Jews in Behar et al. 2008, just as they were in our Karaites K13 and K22 whom we found matched multiple Moroccan Jews in Family Tree DNA's database; however, the Moroccan Jews in Behar et al. 2008 also had the control region 204 unlike our Karaites.

       Our East European Karaites do not belong to any of the four main founding lineages of Ashkenazic mtDNA (haplogroups K1a1b1a, K1a9, K2a2, N1b2) that were reported by Behar et al. 2008.$9$ Nor do any of our Karaites have some of the clearly Middle Eastern mtDNA haplogroups that are shared among Ashkenazim, Sephardim, and Mizrahim, like R0a2 or HV1b2. The Karaites' maternal roots, therefore, do not look as though they're necessarily Middle Eastern, in contrast to many of their paternal roots that are shared with Ashkenazim. In this way, Karaite Jews fit the general pattern of Rabbinical Jewish communities across the Diaspora from Morocco to Georgia to India that were founded by Israelite men who often married non-Israelite (Iberian, Italian, Indian, etc.) women.

       Overall, East European Karaites are largely a Middle Eastern people descended from the Israelites, but like other Jewish populations they are a mosaic - the descendants of several ethnic groups that joined this specific stream of Judaism during different periods. Aside from their Israelite component, our DNA study has led us to conclude that they also descend in part from ethnic groups that lived in the Byzantine Empire and in Asia. They may have small amounts of Western European and Caucasus region ancestries (the Byzantine Empire at times included portions of those regions).

       In some cases, we could not recognize the ethno-ancestral characteristic of a specific haplotype; further comprehensive testing would accomplish this task. The detection of additional SNPs, testing for extended STR markers, and full mtDNA testing would help us to get a better understanding of the subjects of our investigation. Analysis of the autosomal DNA is another technique that would provide valuable inputs on this topic. Therefore, our study should be viewed as only a starting point for the analysis of this people.

Acknowledgements

We are thankful for the guidance of Bennett Greenspan, president of Family Tree DNA, and for the assistance of Rebekah Adele Canada, a volunteer administrator at Family Tree DNA. We also wish to thank all the Karaites who volunteered to participate in our study and to those who assisted in funding the tests. Most samples were gathered by Kevin Brook, Leon Kull, and Abraham Kefeli. One sample came to us from National Geographic's Genographic Project.

Footnotes

1. Shapira 2003, pp. 1-3. There were also back-migrations of Crimean Karaites to Istanbul (Shapira 2003, p. 4 and Kefeli 2005, pp. 3, 5), including as late as the late 19th century, and due to the migrations of Karaites in both directions many surnames were shared in common between the Karaites living in both communities (Kefeli 2005, p. 6).
2. Kizilov 2009, pp. 219-221.
3. Gammer 2001, p. 71.
4. Kizilov 2009, p. 219.
5. ytree.ftdna.com has the complete YCC tree and the nomenclature conversion between it and ISOGG's.
6. ISOGG's Y-DNA Haplogroup Tree 2011 v.6.67, which has subsequently been updated.
7. Sample K07's ancestry was reported to be Polish Karaite on p. 231 of Brook 2006 since his father was born there, but he later confirmed that his paternal grandfather had previously lived in Lithuania.
8. Behar, et al., 2008, "Table S1. Control-region haplotype and haplogroup labeling"
9. Behar, et al., 2008, "Table 2. A list of complete mtDNA based lineages established for all frequently present variants in the Jewish communities studied"
10. Gurianov, et al., 2015, pp. 92, 93.
11. Gurianov, et al., 2015, pp. 91, 93, 102.

Selected Bibliography