This podcast is brought to you by the Albany public library main branch and the generosity of listeners like you. God daddy these people talk as much as you do! Razib Khan’s Unsupervised Learning.
Hey, everybody, this is Razib Khan with the Unsupervised Learning podcast and today, you know, usually you say I have a very special episode. But today I do have a very special episode because we have multiple guests, which I have not done, I think on this podcast yet. And so here we are. And since I've got 1, 2, 3 - I think, guests I will let them introduce themselves. How about we go in order of alphabetic first name? So Aric, then Josh, then Leo?
Sure, no problem. So hi, Razib. Hi everybody who’s listening. My name is Aric. I'm actually a data scientist in real life. Great hobbyist of genetics been following Razib for quite a while obviously. And recently also took as part or as a result of my involvement, mine and Josh and Leo's involvement in the Erfurt study also started part time as a research assistant for a Carmi for Shai Carmi. I'm in Israel. So I'm the one out of country out of time here. And great to finally meet you Razib and Well, Josh, and Leo are all acquaintances of mine already. Actually even got to sit for a beer with Josh few months ago when he traveled overseas to Israel, and other than that, excited to be here.
Great to have you all right, Josh, you go.
Hi Razib, good to be back. I'm a returning Unsupervised Learning guest. I'm a psychologist in training. But as you know, longtime hobbyist of human population genetic and historical inquiry, specifically, when it comes to Jewish genetics and history, and have gotten formally involved in a study for the first time over the last half year with the Carmi team with Aric and Leo, and honored to know all three of you in person.
All right, Leo, go.
Hello, everyone. My name is Leo, thank you Razib for having us three on today. We very much appreciate it. I am a recent, undergrad graduate. I live in Michigan. And similar to Josh and Aric, I am a sort of obsessive hobbyist about population genetics have been a active project admin on Family Tree DNA for some years now, and was lucky enough to help out with the Erfurt study. So thanks again for having us.
Yeah, no, it's my pleasure. So you know, you guys have reference reference, the Erfurt study. So let me mention it real quickly for the listeners in the viewers that are woefully ignorant of Jewish population genetics and don't know what the Erfurt study is. So the paper is in cell now it was a preprint earlier. So if you Google the title, you should be able to get the preprint. But I think the paper itself is open access. So you should be able to get that too. So I recommend everyone read the paper. It's actually pretty short and succinct. You guys added some stuff in the supplement. So we'll get into that. The papers genome wide data from medieval German Jews show that Ashkenazi founder event predated the 14th century, I believe these samples probably are dated to around 1359 or so. They were associated with a pogrom, I believe or a killing or something like that, right? Am I correct? You can so that's why we can probably get the date pretty precise. So when they say 14th century, they're saying before 1359. This is important, because the origin of the Ashkenazi Jews is historically delimited. Probably you should not go much earlier than much earlier than 1000 ad. You know Ashkenazim as a liturgical tradition as a cultural tradition. You probably go back to these Rhineland, Jews and Rashi and those rabbinical authorities and that's how they can distinguish them from the Sephardim, the Jews of Spain. Were obviously a big deal at around the same period, you know, before the Reconquista during the Reconquista. So you have this like medieval, you know, epic where the different, you know, dominant European Jewish traditions emerge. And so getting the genomic data from this particular time is a snapshot of a very, very important period. And they got genome wide data from 33 Ashkenazi Jews. Some of them were related. So they had families in there, they had a lot of children and, you know, older people, they got to a good range there, they found a bunch of different they just covered a bunch of different things. So, for the listener and the viewer, you know, there's plenty of out there and I'll link to my earlier pieces on Ashkenazi Jews. You know, I've written about this topic, too, basically, you know, we all know, I mean, for almost a century now, we know there's been a bottleneck due to the elevation of recessive diseases in the Ashkenazi Jewish community. So for example, Tay-Sachs is probably the most well known, but there are other ones out there too. And, you know, some of the BRCA mutations associated with breast cancer and cancer in general, are also found in the Ashkenazi Jewish community main hypothesis here as they must have gone through a bottleneck. This paper, along with many previous papers in the last 20 years confirmed that the Ashkenazi Jews did go through a bottleneck. Those of you who are Ashkenazi Jewish, who get on platforms like ancestry and 23andme, are overwhelmed by relatives. And the reason you're overwhelmed by relatives is because you know, 500 to 1000 years ago, you share all the same common ancestors because the population went through a bottleneck, probably between 100 to 1000 people, and your pedigrees are all interrelated that way, - it’s similar to some Indian caste groups and whatnot. This is not an exceptional thing. I do want to clarify here. You know, it's common to say Ashkenazi Jews are inbred. But, you know, if you look at how inbred they are on a worldwide scale, it's not really that bad. They're in the range of Finns. There's a lot of groups that are isolates that are considerably more inbred. So you have to be a little careful when you use that term, because you're using a specific semantic term for something that's on a continuum. Second, I guess we have to talk about in this paper, and you guys are gonna talk about it, you know, where did the Ashkenazim come from genetically, as opposed to culturally so culturally, you know, Ashkenazim, you know, means like Jews in Northern Europe, Germans, right. the term Ashkenazi might actually though come from the Bible and originally referred to the Scythians - which those of you who read my substack, which you should, will know that right, so like, that's a weird circumlocution. But in any case, setting that aside, obviously, physically, Ashkenazi Jews look different than, to some extent, Sephardic Jews, but definitely Mizrahi Jews, Mizrahi Jews, the Jews in the Near East, these categories are somewhat artificial, and they're somewhat new, and they use the Ashkenazim as a reference, right, but whatever that is, what it is. So you know, there's hypotheses, whether they're indigenous converts, or whether they're not, and these are evaluated by until recently, by looking at like, physical features. There are some Ashkenazi Jews that look very, very similar to the non Jewish neighbors, and many who do not. Well, it turns out, they're an admixed population, but the proportions is, proportions is what we're gonna argue about, I think we now understand, you know, three primary components. And you know, the weighting of these vectors or these, you know, parameters is the key here. So they have indigenous, Northern European, probably Slavic, or East German ancestry, the smallest component, and then the two other components, probably the largest component is something like, Italian, Southern European. And I'll let you guys go into details on that. And then you have, of course, the Middle Eastern component, which is evident in the Y chromosomes for sure, with the Cohen Modal Haplotype. But also just in the genome wide. Now, the original first pass estimates, were basically about half Middle Eastern and half European, the Carmi paper that you guys are on, has some alternative models, I'm going to say ahead of time, and we'll we'll dig into this in detail, but I'm gonna say ahead of time, that it's quite clear to me, and they are included in Carmi et al. , I mean, you know, the paper is quite clear as well. The proportions that you estimate, are extremely sensitive to how you specify the model. And there are many alternative models right now that seem to fit the data. And probably the way this is gonna get resolved is with more ancient DNA, or it's not going to get resolved to be honest. I mean, it could get resolved with, you know, new novel methods that are going to emerge in the last in the next like 10 years. I don't know, but right now, we have alternative readings that fit the data. And you know, it's kind of up in the air like we know what the mix we To the populations that are mixing into create Ashkenazi Jews, we do not know the proportions very well. So that is what I'm going to say, to introduce the paper to interest to kind of contextualize the conversation. Now, I just want to get into it with what you guys contributed first, and then we can loop back to the general questions. I don't know, Josh, if you want to start and then Eric or Leo can go, but you know, it's up to you guys.
Yeah, I'd be happy to I think this is a good opportunity to call back to our conversation last fall, you and me Razib on the podcast. And, you know, we were talking in a more general way about different lines of evidence for how Ashkenazi Jews came to be as a population. And I made some references to some of the investigations, mostly relying on patterns in modern Ashkenazi DNA, that some of my friends and I from the internet had found that gave us some sense of correspondence to historical and linguistic data. And specifically, we were talking about how I think that Ashkenazi Jews primarily are a fusion of two early medieval Jewish groups, some who lived in Western Germany and France, kind of known as Ashkenazim proper, from the Rhineland area, and then another group that existed at the same time, it's a little bit less well known that lived in, in Bohemia in Bavaria, in some of the surrounding Central European lands that seem to have spoken an old dialect of Czech. But anyway, what is remarkable is we had this conversation, before the Erfurt data came out even in preprint form. And in many different senses, we could not have asked for better confirmation of better testability of and confirmation of this hypothesis, because Erfurt is one of the places that Jewish historians in linguistics had already believed was a site of mixing between these two communities. Max Weinreich, the Yiddish linguist wrote about this in the early to mid 20th century. And lo and behold, Carmi and the team found really good proof of this. And then Leo, Aric and I reached out and said, Actually, we think this is an even bigger bombshell than you have reported. And here is all of this inquiry that we've done. That brings in other lines of evidence, we think this is an even more complete story than you've reported so far. And Shai Carmi, very gracefully brought us in from there. So I'll, I'll hand it over to whoever wants to speak on things next.
I will let you go. But I will say, Josh, you were right. Because I did I don't know this area, you were right about a lot of things. So I'm gonna give that to you give that to you, all you guys, your whole community. You know, I know you guys are being gracious and saying, Oh, we're just amateurs. And we're hobbyists, and all this stuff. You were right, you knew the truth, before the scientists stumbled on the truth. So let's not like soft pedal what's going on here? There's public data, you guys had an interest and you reach for the truth. And you stumbled on the truth before? You know? I mean, you knew this before. Shai Carmi did, you know so let's just keep it real here. It's great. I'm really impressed. And obviously, this is why we're in it. This is why we're having this discussion is why we spend our free time or our paid time doing this sort of stuff, the truth, right? Like this is great. Like we are like as the gods like we know now, you know, like, can you imagine like five years ago, getting ancient DNA from Erfurt and confirming these things? Right. So I'm super excited. I know you guys are super excited. I hope the listeners and the viewers, I hope they understand why we're excited. And I hope some of them also understand this could be them, like they could, you know, follow their passions and their interest. The data is out there. There's not enough people looking at the data. But anyway, Leo, you go, man.
Thank you. Thank you. I did just want to shout out one more time to Family Tree DNA, which is one of the only places that's where something like this is actually possible, the amount of samples they have on that, like website in that company. No academic study could ever come close. And if you know how to use it, and you have to basically volunteer to have access to a lot of this stuff. But once you do you understand that You know, the power that comes from big sample sizes is possible. And it's and it can be put into the hands of everyday people who are just willing to put up some of their time. And it is it is really encouraging, I think that we can start to bring together the the bigger names and academics with this sort of on the ground community research stuff, which has the, again, the actual statistical power to make claims like we did, before there was ancient DNA. It is definitely a path towards the discovery of so much of the actual structure of populations over time throughout the world. And it's not surprising that it's happening first with with AJs. Considering how well tested they are, but it is really good news for the future of the field.
Yeah, um, I just say, first of all, these are exciting times, I agree with all of you. But I would just like to emphasize, first of all, about the openness of Shai to ask, this is really not a trivial thing. I mean, we've all like had our hypotheses. And yeah, we were right, in some of the things you're right, resemble, like humble, humble enough or not. But yeah, we were write on some things. But the fact that I mean, we had like, dozens, if not, like, almost 100 emails, exchanges with Carmi, who has tons of other things to do. And his mind was really busy pushing this preprint. And it was really open to hear what we have to say. And I think the fact that we could, in this collaboration, kind of like, reach out and connect the massive, that pretty massive compared to this kind of interest community that we have with all the data that we have, from both Family Tree DNA. And all the different papers that we read, maybe not all of them, like or not all, not every academic just can’t sit and read all the papers all the time. They have other things to do, and other things to focus on in the research, and analysis and stuff like that. But in general, the fact that we could collaborate like that, is really a shout out to and I think generally every person which has an interest in any scientific stuff, or any scientific subject, to get such collaboration from someone who is really big deal in the academic industry, or in the academia, is basically a dream come true in some way in some form. And this could not have happened without his openness. And I mean, he basically, we were conversation, we're actually we're discussing about basically everything from the European southern - from the Southern European admixture, which not all of it was included in the paper for peer review reasons. And as you actually Razib correctly observed, we might have to wait for more additional ancient samples to be able to say a more clear answer to give a more clear answer on that in future papers. But it was also extremely open to Family Tree DNA, even though it's not a peer reviewed, not a peer reviewed database at all. I mean, these are not samples which were academically sampled, and these are people but he was really opened, he actually also David Reich himself, basically, he met with him and showed him all the data, some of it got into the paper, some of it not because, you know, peer review, which is understandable. Some of it actually is is not there yet, but when the Norwich paper came out, which is another paper which also talks on the shows and discusses ancient Ashkenazi or not Ashkenazi, but Sephardic northern French, medieval Jewish remains. He, Shai was really open to also discuss the samples when they were out. And these not that was not included in the final paper. But I can say that for instance, one of the things which was not included, like for instance, the same pretty rare, like redhead mutation, which was found in the Norwich paper was also found in in one of the Erfurt M E group, which is not in the final paper, by the way, because it was discovered after we already submitted everything, but I just wanted to say that the dynamicity the the kind of natural attitude that all of us had, including Shai was what made this possible and guys, all of you that are hobbyists, like hobbyists and professionals like this attitude to go this is the way to actually push science forward. In my opinion, my humble opinion.
Yeah, so let's, um, let's actually, uh, Josh, do you want to just like talk about, you know, the main results in the abstract and then maybe you three can dig into what you contributed. I know there was a lot of stuff related to Y and mtDNA. I think it's mostly uni parentals. But other things. So can you can we go Can we go in that direction? Because, you know, we've been talking about the Erfurt paper. I've read it. I know you guys read it. You guys probably read it a bunch of times because you know, revisions. But I give give the listener the viewer, the high level, I mean, not the high level, given the medium level, like we're about the medium level here.
Okay, well, we'll give it a shot. So basically, mid to late 14th century Erfurt cemetery, which was found as part of excavations and renovations and construction in the city in central Germany, although it was part of East Germany during divided German times. And they were able to obtain pretty good quality genotypes for these 33 Jews who they realized pretty early on, resembled Ashkenazi Jews genetically pretty closely, no surprise, considering the mainstream position is that all Ashkenazi Jews, at least by virtue of their heritage language, Yiddish, which is a Germanic language, have roots in medieval Germany. However, what was interesting was they found that there were two distinct clusters of Jews buried side by side, in the same cemetery, no distinguishing features in terms of the graves, often, you know that they weren't buried in different sections. And one of these groups was very genetically close to Sephardi Jews, Italian Jews, also, not incidentally, to modern Western Ashkenazi Jews from France and Germany. And then the other group was similar to them, except they had 20 to 40% Slavic admixture in addition to that, and it was really curious that on top of that, they were able to demonstrate that this wasn't just people from the same community, some of whom had recent Slavic admixture. The two groups were distinct in terms of their isotope ratios, which strongly suggested that they were originally from two distinct places. And that lines up with a lot of documentary evidence, I actually ended up accessing the primary source this spring when the preprint first came out, where you can find the list of the names of all the Jews who were registered as living in Erfurt in the late - late 14th century, and all of their places of origin. Because again, remember, this is a community that was presumably wiped out during a pogrom in the 1340s 1350s and then repopulated from a million different places. So they found that two distinct groups not just distinct in terms of their level of Slavic admixture, but also distinct in terms of literally where they were from. And that basically, if you average these two groups together, kind of on a 60-40 basis, you get modern Eastern Ashkenazi Jews, which is pretty astonishing. And in addition to that, you know, they were able to run a bunch of different models, which Aric can talk a lot more about that show that Erfurt Jews if you average them all together, and also modern Ashkenazi Jews can be pretty effectively modeled as a three way mix between in descending order of size Southern European component, a Middle Eastern or Levantine component, and an Eastern European component. And in addition to this, many major Ashkenazi uni parental lineages were found in Erfurt And at least in the case of the Y chromosomes, this was what was really astonishing. The lineages that we have long believed, were associated with Rhineland Jews, showed up in the Rhineland type group, and the lineages including my own, that we had hypothesized for the last three years were associated with medieval bohemian and Bavarian Jews. Lo and behold, showed up in that much more Slavic ad mixed group.
So I, you know, let me
jump in real quick here. And I think you said Eric should should take over at this point, I want to say so we have these two groups that just to review we got these two groups of Jews. I think the nomenclature here is really not clear to me sometimes but I think of him as like Western proto Ashkenazi, and Eastern proto Ashkenazi, but ultimately the modern Ashkenazi, and he said 60/40 They're a combination of these two in general. And these are the ancestral groups and the modern Ashkenazi are quite homogenous, actually. So they're panmictic over these two groups. Is that a correct interpretation?
Broadly, at least Eastern Ashkenazi, if you find Jews who are and there are fewer and fewer these days who are descended from people who have been in Alsace exclusively, for the last few 100 years are from a lot of parts of Germany, you will find that there are still Western Ashkenazi Jews today, fewer and fewer of full Western Ashkenazi ancestry, who can be modeled as 90%, or more descended from that Western proto Ashkenazi community? But that's a drop in the bucket compared to the global Ashkenazi population.
All right. All right. Go on, Aric
Yeah, so as Josh very eloquently explained, the paper. The models are quite interesting, because Razib, again, I need to stress it out, we're still in somewhat of the dark here regarding what is exactly the proportions of each ancestry- each ancestry component in Ashkenazi Jews. But what we can all agree on is the Southern European one. And the part which is quite revolutionary in this paper, is that until quite recently, there used to be a consensus in academics of something that we and the hobbyists, kind of called it half and half, which is pretty much a mixture of Northern Italians and medieval Italian and Levantine population. As you said, the consensus is that most of the Levintine were thé men and they converted and married with Northern Italian, non Jews women mostly, and of course, it's not one hundred percent only the men, not 100% of the women because as we look at the subjects, we see some non eleventeen subjects and some non Middle Eastern subgroups, although these are the minority among the men, and among the paternal lineages and some near Eastern lineages among the maternal side. So it's not, it's not like a 100%, that and 100% that, but this paper actually open up to all sorts of other models, which is something we discussed for the past four years. Also, there used to be a lot of talks about it in, in some of the websites I don’t know if you can mention them - at family tree dna, etc, etc, the more famous ones. And what this paper actually first did is to create. So Shai was actually really open to test other groups. And they find yourself discussing and testing and basically, experimenting with with him and Shamam also all sorts of other populations. And he found also that Greeks can also play the same proxy, as Italians in the preprint. Before we even got involved. He first, their group found the southern Italians to be a more adequate proxy for the South and European component, which actually makes sense historically. Because if you look at the historical progress, where Jews used to live even during Roman times, which was the heydays of conversions of Romans, to Judaism, after that, you know, after the adoption of Christianity, that chance or that window of opportunity, if you can call it that, pretty much shuts down, or dramatically shuts down. But basically, most of the communities lived either in Rome or
suffering that we have catacombs mostly in Venosa, in Rome. But the communities in North Italy were never that substantial, especially during Roman times. They were there, Jews lived all over the Roman Empire. But what is interesting is that if you look at Antonio et al. and all the papers, which came out on Italy, in Rome and Italy, in the past few years, you see that this stuff in Italian profiler wasn't more profound, or we can call it pretty much penetrated all the way almost to where Tuscany is today to some degree, of course, not in the same proportions. Get the cline, South to North cline. But it first I know, when they told me that l came out, people thought it might be only Rome because it was the capital city, then all bunch of other papers came out as well, which we can Ponce et al, in other papers as well. And they pretty much showed that. I mean, we have -
I want to I want to jump in here, Aric, because Antonio et al. I know what that paper is. I've read it at some of the listeners will. But let me just like summarize it really quickly so that they have context as you go on. Basically, what we're and I wrote a substack about this in March 2021. But basically, there's a fair amount of evidence now that during the height of the Roman Empire, during the Roman Empire, during the imperial period, there was a shift if especially in the cities that we have samples from, towards ancestry, let's just say more to the eastern Mediterranean. I mean, it was more cosmopolitan overall. But you know, Juvenal’s critique is that basically Rome was dominated by, you know, Greeks and Syrians not totally off. And then what we see after the end of the Empire is a shift back to, you know, let's say indigenous Italian people from the hunter hinterlands, probably this is just due to the fact that cities are, you know, fertility shredders, and unless you have continuous migration, you're not going to maintain the size of a city. And so you know, there are people in Rome north into Tuscany, etc, etc. Actually, into England, even to some extent, that have very Eastern Mediterranean genetic profiles in the ancient DNA. So the historical records are correct. And it's not shocking that there's Greek inscriptions all over the catacombs in Rome, because a lot of those people were there, why modern Italians, not as genetically Eastern as that would indicate. And a lot of that just has to do with probably differential demographics between rural and urban areas. I just want to set that context go on Aric.
Right. So yeah, but then we have southern Italians, which modern Southern Italians are still pretty much Eastern shifted, which probably is the result of that demographic shift. But in any case, as you very eloquently pointed out, the Greeks on the catacombs, also what it's actually writing, which dominated the early Jewish catacombs, which makes sense because you know, Koine, Greek was the lingua franca of the Mediterranean back then, especially in the Eastern Mediterranean. And what is brought up is that perhaps not that all of the Southern European component is specifically Italian, again, we cannot say that 100% it's pending future samples, maybe from farther to the east of Italy. But what I think we can now say is that instead of focusing only on northern Italy, or maybe only on southern Italy, and by the way, this is also what the peer review committee pretty much said, Guys, you can just say specific location, we now need to talk about something I can only term as meta Eastern Mediterranean component can be southern Italy can be Anatolia, can be Greece can be a mixture of all of that, which makes sense because this is basically the movement of Jews from the Levant to Italy. So this is this is something which is quite revolutionary, as this is the first peer reviewed paper on Jewish on Ashkenazi genetics, which actually claims that it says guys, maybe we're not only talking about Italy.
Yeah. And yeah, and I want to jump, I want to jump in Aric and say one thing that I noticed, and maybe you can, like, elaborate or clarify. So the percentages, you know, we've been talking about for years, like about 40%, Middle Eastern, and 60%, European, and all of these things. And what I noticed is, okay, if you use a northern Italian proxy, the Middle Eastern percentage increases. And if you use a southern Italian proxy, the Middle Eastern percentage decreases, which naturally makes sense, because the southern Italians naturally will have allele frequencies that are already shifted to the Middle East. So bringing into this discussion, alternative models, you get all these different outcomes, and, you know, if you're going to build a narrative, this is actually not optimal, but it is what it is, I'm assuming with ancient DNA and IBD segment analysis, at some point, we will be able to narrow it down. It could have been, as you said, a metapopulation, where, you know, these are Greek speaking Jews, right? Probably the Alexandria community, etc. You know, they might have had like, very diverse antecedents. And, you know, we try to when we try to communicate this, whether in scientific papers or just in general, you try to like go as parsimonious as possible, but we might have gone too far. I do wonder about that now.
I’d like to jump in right there.
I totally agree with you,
Exactly. Yeah. The parsimony is, is definitely a guiding philosophy for me, and helps so much but something as as a young person myself that I was surprised about in the process of the paper and talking about the autosomal models with Carmi and other team members was just how small they wanted the the number of sources to be in the model. Whereas from uniparentals, which is definitely my area of expertise, it is clear that we have evidence for some ancestry from basically every modern country in the Mediterranean in some sort of broad sense. And so I'm, I think it'll be difficult in the field forward coming forward to sort of reconcile, like, parsimony in these models, getting results that you can actually talk about and make sense. Whereas they'll always be oversimplifications. They'll always be ignoring some truth that other evidences already imply. And I'd also like to bring this conversation a little bit back closer to Erfurt about what these samples ensembles actually imply for the differences between these two communities and their founding. So we have this classic idea about these western Jews that we say are kind of from the Rhineland, or France and Germany. And we have these understandings about them, whether it's like the Kalonymos family that implies ancestry from northern Italy, or it's the a lot of evidence in favor of some deep ancestral connection with the Sephardium which are the Jews of Northern France. And so the Western Jewish population there in Central Europe is already sort of understood, like we know, it ties into France and Italy in some way. And from that, also to Spain. But the Eastern group, it's still such a big question where they came from, as opposed to the Western group. And I'd like to just talk about how the, the paper found that the Erfurt ME group, which is what they call it, the Western group, that ancestry could be used in formal or informal methods as the main base of ancestry to which Slavic components could be added on to. And so that leaves the picture of like, okay, well, is this group just a, sort of was seeded from the Western group in some time, and then mixed with Slavs and then they're just distinct? Well, we actually know from especially onomastic, which is the study of like naming data, that the names that the Eastern Jews, Jews carried, were biblical in many cases, but were not the same biblical sort of names that the Western Jews were using. Similarly, we have uniparentals, that we can use as evidence for the distinctness distinctness between these two groups. And Josh's lineage that we've talked about before is the most perfect example, it's seemingly going to be one of the major y chromosomal lineages for that Eastern group. And we can say with pretty high confidence that it does come from Roman era Italian Jews. And so this sort of evidence from Erfurt actually gives us some of the first evidence for where the Jews of Bohemia, Bavaria, and other neighboring regions that weren't Ashkenazi Jews, they certainly led to moderns. But they, whether they were called Canaaniem or other names. These Jews, at least in some part came from Roman era Jews from Italy in the same way that the Western Jews did, and Jews from Spain did.
All right, I'm gonna I'm gonna let you guys, you know, this is great, fascinating. I want to set like a little bit of a context again, for the listener and viewer because we're getting, you know, we're getting medium level, we're getting fine grained here, and this is great. Obviously, I love it. But I want to just again, reiterating set some context for the listener or viewer. So we're talking about Ashkenazi Jews. And again, I mentioned Rashi you know, Shlomo Yitzchaki. I guess that's his real name. I don't know. But so you know, this is a A French rabbi in the Rhineland, I mean, French, like French didn't exist in the way we understand but, you know, in like what became France and Rashi is like 11th century, you know, so 1000s and, you know, he's kind of, I mean, okay, this is the touchstone I think for a lot of people, from a cultural perspective, understand the Ashkenazim, right. And so, what is France? What is Gaul, like this area has been under the Roman Empire, for I mean, you know, since the time of Julius Caesar 50 BC, and as Rome changed and evolved, post Roman world came, all of that happened, like we know that. Now, the second issue is you're mentioning these Moravian bohemian Jews, this area is quite different because it was outside of Rome. Okay. And, you know, it was pagan until the nine hundreds, though, do you know about good king Wenceslaus You know, like, this was like a piecemeal conversion that happened relatively late. So culturally, it's an extremely different area of Europe, then, you know, where Rashi was where the Ashkenazim I think we traditionally imagine they come out of this northern European post Roman world. Well, the Moravia is not post Roman because it was never Roman. In fact, it was pagan up until the 10th century. So if there were Jews there, and we know there were Jews in pagan Lithuania, Lithuania was pagan until 1387. AD, we know there were Jews in pagan Lithuania. So the Jews, outran Christianity and Christendom in some cases, just depending on you know, whether there were opportunities for them. So I don't know the historical context. But when we're talking about these eastern Jews, it's a very, very different cultural stream that they are entering into, as opposed to the Jews. That would be in Gaul. Francia, which is, if not totally civilized, at least, did not start from a state of total barbarism. And I'm not using those terms pejoratively. But I want to set the context here were Jews that come in, that are coming out of the Rhineland, these are Jews who always knew Rome and never knew anything different, fundamentally. Whereas if you have Jews that are in Moravia, yeah, they might have been Italian. And perhaps they were from Greece, I don't know where they were from. And maybe you guys can expand on that. But they live in a world that is fundamentally at the root, non Roman, and a world that is actually like barely Christian. As as Erfurt, as these Jews in Erfurt, were dying. There were still vast swathes of the pagan Baltic under Lithuanian rule. So it wasn't even a totally Christian world at this point. And one of the hypotheses I do I have presented, and I don't know if this is true, but you know, there was a ban, obviously, on Jews holding Christian slaves and whatnot. But this probably would not matter if the slaves were basically non Christian. So if you have areas of Eastern Europe, where there's a lot of non Christians, still pagans that are not catechized, that are not baptized, this is actually a potential opportunity for exogamy to happen without any social repercussions. And you can see a modern example of this. In California, in the early 20th century, when there were men from Punjab, who settled in the South Central Valley. And the miscegenation laws were such that Asiatics could not marry white women. But since they're Asiatic, they can't get naturalized. They can't own property, even though they wanted to buy this property. So they're married Mexican American women, because that was a loophole in the law, because they could the law was not applied. Like basically, if you're Asian, you could marry someone who was Mexican, because they'd never thought of that option. And so they married these women, the property was in the women's name. And so this Punjabi Mexican community emerged, just because of a loophole in the law. And it's still around actually, in the southern central valley. There's still traditional families that are like that. Anyway, I just kind of like a digression, but I wanted to set the historical context for the listener and Viewer. This is a really, really different world than the Europe obviously, of even three or 400 years later, like this is a liminal time. And the Jews are the East here, they come from a very liminal space. So go on, right, yeah.
So I would just like to add something, if I can. So you touch the spot here because usually we talk a lot about the Rhine River, the Rhineland in general. And Leo kinda mentioned that they arrived to Moravia, perhaps from Italy, perhaps from other places. But nearby, there is the Danube River, which goes from another part of the Roman Empire. Many people usually don't call it the Roman Empire, but the Byzantine Empire, which was basically Rome, just Greek speaking, as Rome was, by the way, the eastern parts anyhow. But basically, there is an hypothesis, which has some historical merits, that some of the Jews that have found themselves in Bohemia and Moravia there's even like a myth from one of the rabbis that telling that the earliest Jews have Prague arrived from the Byzantine Empire, but there is no historical evidence to support it, it just the lore. But anyhow, we can't really, we can't really say that they didn't come from there as well, either as merchants, you mentioned pagans and slavery? Well, one of the things in Judaism, at least, if you look at it from the ancient, it's not relevant anymore. But a slave has to be converted to Judaism, in order to be part of, of, you know, either a man has to be circumcised, or have to be, they have to agree to that it can be forcibly happened to them. But so this is one of the reasons why Christians were not allowed to be to be slaves to Jews, and all sorts of other you know, historical reasons, you know, for being the Christians were the rules or rulers in this country, they wouldn't like for Jews to own such Christian slaves. That's obvious, but also, because from the very early part, pagans, or whatever kind of religious affiliation the person had, would have to go through conversion to Judaism and become part of the part of the people. I'm not a religious authority. So don't quote me on the procedure, because it's not really relevant. Right. But it's, it's not first of all, but also one of the reasons why men wouldn't really hesitate to go because you have to be circumcised. Right? Which a process that isn't really something. Yeah, Josh wants to say something.
I just I just wanted to say and this is reflected so crisply, in the results from Erfurt because when I spoke to you a little over a year ago Razib you know, we had this vision of these two groups, I didn't really think they would be so genetically different. I think they have common roots and Mediterranean Jewish sources in classical antiquity. But what's very interesting is that despite those common roots, one of those groups ended up in Erfurt having been in Germany probably for hundreds of years. The the Western group in Erfurt looks so similar close to identical to Sephardi, and Italian Jews in the transit from Spain, southern France, Italy, to 14th century central Germany, they received very little admixture there is some uniparental evidence of Germanic and French admixture in Ashkenazi Jews, but very little, whereas this southern group came from whether it was Italy, the Byzantine Empire, and entered this barbarian world, not Christianized world and acquired a very large amount of local admixture. So the the structure of these two communities from the get go must have been very, very different to be so different by the 14th century at the time of their encounter.
And I also like to add something which we were actually tapping about in the chat between the three of us. And that's something which we were we didn't mention yet is the tiny bit of East Asian admixture which was found in variable amounts in the different samples. And the N at the N9a3 maternal lineage, which was found strictly in the Erfurt sample, which actually goes hand to hands with what Josh was mentioning, regarding the more Eastern, more Eastern origin of that group, and perhaps it ties to the either the silk road going through the Byzantine Empire or maybe some other location, which, you know, I don't want to say Khazars but here I am, I mentioned it. I don't think I mean, I think the evidence is enough that nobody thinks like no rational person in this in 2022 think that Jews are mostly descended from Khazar converts? This is like totally dead academically wise theory. But I think that this opens up also the slightly higher amount of very slightly higher amount of East Asian component, which I might which I must also emphasize, could not be adequately enough separated from the East European component as my discussions wity Shamam went on but I think it we can't really, we can't really disqualify and maybe a tiny amount of more elevated Khazar ancestry. Again, really minor on the low single digit to these east or Erfurt EU group, more Eastern group I would like to call them Konami for the sake of argument, or maybe Rhinelanders, mixed with Konaneem at that point of time, in the late 14th century, but yet this is something we need to take into account as well. Which, by the way, was not observed in the Erfurt ME group.
So let's, uh, can you just like, I've written about this, N9a3. Josh, do you want to talk about a N9a3 and like, why it's, I don't want to say the weird, but it's kind of weird.
I'm gonna field that for Josh
Yeah, that ones for Leo.
Yeah. So, the Ashkenazi subclade of N9a3 is called N9a3a1b1. Now that's entirely known only among Ashkenazi Jews. It has a particular geographic spread, with the sample we have from Family Tree DNA in eastern Ashkenazi Jews. It's barely known in western Ashkenazi Jews. And indeed, in Erfurt it was found in the eastern group. I think importantly, at N9a3a1b, so just one step up in the phylogeny. There's a few North Caucasian populations. There's Bashkirs, there's Ingush and there's North Ossetians and as well, there are also Chinese. And so it's hard to say for sure, exactly where it came from, in east Eurasia. Ultimately, ultimately, we know from ancient DNA that this variety of haplogroup N9 is from Far East Euasia. But proximately it does seem like the lineages closest relatives are formed are from North Caucasian and, and different Turkic speaking or non Turkic speaking, North North Caucasian populations. And it is a really unique connection and not something that I expected to see necessarily, in 14th century Central Europe
All right, so this is um, you're just you're talking about Khazars Now, come on, let's keep it real.
And what I what I should also none of us has been struck by lightning yet, so I think we're fine.
Yeah,
I think what this also just indicates I don't want to get too into the weeds, but there's also evidence, linguistic and and possibly genetic to and certainly historical, that there was a third group of Jews, you know, even farther east in East Slavic lands in Ruthenia. In early medieval times, who very possibly had links to Khazaria, they were probably not predominantly descended from Khazar, converts Jews emigrated to the Khazar empire. So so it's possible that that group already had ties to the central eastern European Jewish community in Bohemia, through migration and intermarriage. By the 14th century. It's not the same. But if you if you read, one of our favorite historical linguistic sources, who is also an author on the paper is the Yiddish linguist and also physicist, Alexandra Beider. Talks about Ashkenazi Jews probably being descended primarily from these, you know, Rhineland. And then on the other hand, Danube groups of Jews, but then in a tertiary sense from Jews from me even farther east. So that might be how and N9a3 eventually made its way from the North Caucasus to Erfurt.
All right. So, Aric , Leo, do you have any more to say on this particular topic?
I would say just very, very quickly, that I think we in just giving a try this, this sort of method of using high coverage y chromosomal sequences specifically, but also empty DNA sequences, just trying to push the envelope on what is possible with this data. With bigger sample sizes and comparing it to autosomal models and in particular using the uniparentals, the high coverage uniparentals as a way to rain the autosomal models in and to help researchers pick the right sources and pick like what are the necessary sources? I think we we did some good work here. And I'm excited to see what's possible in the future for combining both kinds of these data for better results for pop Gen overall.
So okay, um, just a real quick question. What is the autosomal, Middle Eastern estimate for y and mtDNA? Just give me a ballpark anyone.
So I would say the ranges for autosomal, like contribution from the Near East, for Eastern Ashkenazi Jews is probably and I guess you have to define what the Near East is, let's say not including North Africa. And let's say not including the like the sort of Dota conyza or Western Anatolia, region that's lump that in with Southern Europe for now. And say, just like classical Near East, I don't know, I would say probably a solid range of 20 to 35%. Around, add together, let's say 5%, North African, and a chunk, a big chunk for all of Southern Europe, everything from Western Anatolia, to the west Med, probably something like 30%, for that, up to 40%, perhaps, and then a total Northern European component of about 15 to 20%, most of which is Slavic, but some of which surely is going to be Western, European in origin. That kind of gives you the best sort of sense. For the autosomal contributions. I will say for uniparentals. It is unclear still, but certainly on the y's, there are like a disproportionate disproportionate amount of Near Eastern Y chromosomes. And on mtDNA is there are a disproportionate amount of ultimately European mitochondria. But at the end of the day, those sorts of confidence intervals are still so wide, because we're still trying to try this new kind of science out.
And one thing I just want to add, and I I'm not sure that I'm in full agreement with Leo's figures, but something that's a little paradoxical. And I know that this can happen due to drift and, and, you know, different additions to a population soup. Is that the Southern European component, which does seem to be plurality, autosomally seems to be underrepresented on both the Y and mitochondrial sides relative to its autosomal contribution. For example, y's are probably majority Near Eastern and, and the share of mitochondrial DNA that's northern or Central European. You know, rivals or might even outstripped the share that Southern European I know that's still very tentative. It's complicated with mitochondrial DNA. But there's a there is a weird disparity, you don't really, you can't really average the Y proportions with the mitochondrial proportions and end up with what seems to be the autosomal proportions. And I'm, you know,
I've noticed that.
Yeah,
it's strange.
I would just like to add that I think one of the reasons for that is because most of the conversions, and most of the, let's say, heterogeneity in origin, and different sources of population was mostly on the maternal side. We know that from historical records. We know that, you know, the patriarchy of Jewish culture, medieval Jewish culture, where men would sometimes go either to study in certain certain yeshivas or certain places, they might marry local women either converts in some very rare occasions as they go, you know, more upward in history. But, you know, as you go backward, I mean, the story might be multi layered. It's almost certainly multi layered. I don't know if I exactly I agree with Leo on his figures. I think that in terms of ancestry, I mean, we will basically say the same regarding the variety of Mediterranean origin terms of the Southern European continent. But the numbers I think are pending more ancient samples, I will be very careful to just throw numbers as it is, I mean, if anything, the Erfurt paper showed that you can use 85% southern Italian, or you can go again to the half and half model and have, you know, someone from, you know, population proxy from Tuscany, and Levantine and maybe the Middle Eastern is not only strictly Levantine, we have to remember that Judaism was also popular in the Middle East during Roman times. We know from, from Josephus, that a lot of Syrians converted in places like Antioch, which is almost near adjacent to Anatolia is long, you know, together with Alexandria, which I'm not sure that all of the converts, there would be Greeks, maybe some of them would be Egyptians as well. So it's, it's an extremely varied. And, again, we need to wait for and I know pending, more ancient samples, which hopefully will arrive soon. And you know, anxiously waiting for them.
So I want to before we move on, I want to say something real quick, about Sephardim. So I've done a little bit of exploratory analysis of the Sephardic samples, Family Tree DNA, by the way, where I did consult for a couple of years and I still have good relations. I think with them, you know, it's been a while but, you know, they've worked a lot in this Jewish space. And one thing with the Sephardic samples is Sephardim are culturally so Sephardic Jews Sephardim are defined by their liturgy, you know, by the Spanish tradition, it's a cultural definition. And they are genetically somewhat heterogeneous. So Moroccan, Sephardic Jews always seem to have some sort of Berber ancestry. And that's probably due to the fact that there were indigenous Jews, Berber Jews, or whatever in the area that merged with the Spaniards that came in. What I do remember specifically was the Turkish Sephardim, the Sephardic Jews from Turkey. Were very distinct in that they seem to be like no indigenous admixture, like I don't know, there were no Jews left in that part of Anatolia, or what? In the paper, it's noted that the Western Middle East, the Western, Ashkenazi Jews, you know, the ME group, you know, were very similar to those Turkish Jews. And to me, that wasn't surprising because I think those are, I think the Turkish Jews are like the original Sephardim. And those of you who've done ancestry analysis in earlier years know that Sephardic Jews often got a result that they were 50% Ashkenazim, and my hypothesis at the time, for that is always that, oh, well, you know, they don't have Sephardic Jews, in their reference sample. So they're finding the closest populations. And so they're a mix of like Southern Italian, Southern European and Ashkenazi, and that's what it comes out as well, what this result tells me this is tentative, but it does tell me, there's very little difference in 1000 ad, between Jews and the Rhineland and Jews in Spain, they might have already been linguistically somewhat different because the different versions of Latin they're speaking, but you know, this is very close to the beginning of the separation. And modern day Sephardic Jews, the reason that they have this affinity to the Ashkenazim is because of this, which we kind of already knew. But that's, I think, an important thing to point out.
Yeah, I'm so so. So true. And I mean, it's complicated. I think Eastern Sephardi Jews do have a significant we don't know how much component from Romaniote Greek speaking Jews who were there before, but we don't know how much. What is interesting is that there are a lot of uniparental lineages that confirm exactly what you're saying why, for example, Y lineages that are found in Sephardim and Ashkenazim that have a common ancestor sometime between 500 and 1000. C, and whether those people lived in southern France or in Italy, or maybe in northeastern Spain, but but the guests in many cases is probably southern France, where deep Jewish continuity is documented with links to both Ashkenaz and Sephard seems to be pretty good evidence that major components of both of these groups were are in the late first millennium CE are to be found in the same place. Basically, we're still in the process of splitting Until about 1000.
So I would like to also add something I'm really glad you brought up Sephardim because I want to touch two things about this. First of all, one of the things which is known in Costa et al. back in 2013, raised the possibility that one of the major maternal lineages among Ashkenazi which is K1a1b1a my own lineage, by the way, mine and like 40% Ashkenazi, so, not that rare, but, um, is also present at Sephardim in very tiny amounts. And he raised, of course, it's like 10 years ago, almost a decade ago, that maybe that's like a common Roman era kind of origin to that lineage. But what we already know from this our work and research, the were Ashkenazim that found their way into the Ottoman Empire. And even before that, it was a big exposure for France, back in the late 14th century. Some of them reached the Balkans. We know that they had separate cemeteries, and even the predates Sephard they were in small amounts. So this is one of the reasons why I mean, I'm not suggesting this is why Sephardim get like 40% Ashkenazi in the ancestry tests, that's that's not the amount of ancestry they have. But we know that Sephardim in general have a more heterogeneous origin. They also have Ashkenazi ancestry very minor one. And the second thing I would like to touch is that there is like a, there is a common law among a lot of Ashkenazim that Ashkenazim are the result of the explosion, the the Alhambra explosion, the explosion of Spanish Jews Sephardim from Spain, which settled all over Europe. I don't think this is the case. I mean, certainly their were some Sephardim who settled in Galicia in southern Poland, in other places, Netherlands, et cetera, et cetera. But I do agree with Josh, that the uniparental evidence, and I think that's because both communities, I mean, we need to see ancient Sephardim samples, hopefully, these will arrive someday. But I think both of these communities have a base in some sort of literal strip of North East Iberia, southern France, North Italy, early medieval late antiquity area, somewhere around at the eighth or like, ninth century CE I don't think we would be able to tell the difference or causality and even in some of the lineages, between North Eastern Syrian Jews or Sephardic or proto-Sephardim of that period in Sarfat itself and French Jews and North Italian Jews, because I think these communities really intermarried at that point at some point before even the the Moorish occupation or rule over Spain, and before the Gothic explosions of Jews from that area. And so in many ways, they have a very similar origin probably around the split from the FF and after the eighth century C. With the Jews with Ashkenazi Jews getting, you know, a more severe bottleneck, which caused you know, what we see today, but this is something which which which needs to be said. So, if I can summarize, Sephardim in Turkey, have very minimal Ashkenazi ancestry. Most Ashkenazim, of course, do not descend from Sephardim but they both have probably, again, we need to see ancient samples, probably a very tight origin. From what Josh said southern France, that littoral area, which connects Iberia, south France, and North Italy.
If I could jump in very quickly, the samples from Norwich, England also fit right into this pattern as well. Yeah. It is understood through historical sources that these Jews there probably came quite recently from Northern France. We've been calling that region Sarfat but these Jewish Y chromosomes that we got from that study, are the same clades that we've been looking at for a long time, that are evidence for the closeness in the, in the relationship, the ancestral like genealogical relationship between Ashkenazi Jews and Sephardi Jews. And we can say with some certainty that the Jews that lived in sort of high medieval Northern France also fit in with this community as well. And that sure there was a continue on. But by and large Western Ashkenazi Jews like the Rhineland, Jews of that high medieval time, the northern French Jews, the Southern French Jews, and in some big way the Jews of Spain all descend in a really big way from the same sort of early medieval ancestors. It's it's definitely not super surprising, but also is awesome to actually have, like, mounting evidence for.
But also, we know now that that only accounts for 60% of the ancestry of modern Eastern Ashkenazi Jews.
Yeah, and Leo I'm glad I'm glad, by the way, you mentioned the Norwich paper, because even though we're here to discuss the Erfurt one, because we also observed Norwich samples. While we were, I mean, because it came out straight after the preprint of Erfurt, we also got the chance to observe the samples there. And a lot of conclusions, including the redhead mutation, which I mentioned in the beginning, which is quite rare, rare, and was found both in SB 604, the sample in Norwich. And even though they and the number of samples there is very small, but it also was later on found in one of the Erfurt M E group. Not the Erfurt EU, which is also an interesting, and it's not a very common rate of mutation, it is quite rare. And also something which I like to bring up is the the issue of hereditary diseases and their frequency, which was discussed a lot in Norwich. But what is cool is that we had a chance in a very short amount of time to observe too tightly similar communities, basically, in I think there is a consensus among the three of us that perhaps that the Norwich samples represent the Norwich community, in some way represents something similar to the asset recommend, which gave rise to the Rhineland community later did Sarfati which mentioned, but because Jews back then spoke French, in England, but also that the level of I mean, some of the hereditary diseases we see today, the frequency in those samples, is very similar and akin to what you see in Ashkenazi Jews today. And this is quite I mean, for me, it's not really surprising, but it's astonishing because that what are the implications of the bottleneck? I mean, how far back does it go now? And a lot of questions for future studies arising from it's only two centuries apart, these two these two studies, and it's quite astonishing, in my opinion, to see that, let's say that the Norwich samples are from the 12th century, and we have 12th century samples, very similar to the late 14th century. Erfurt ME groups. And already with the same frequency of of diseases that we see in some cases, not all the diseases, the hereditary diseases that we see in, in modern Ashkenazi Jews. And this also gives question begs the question. I mean, this opens up a whole lot of I mean, a field for research. I mean, what what is if we can use that hereditary pattern that rather disease pattern and frequency to maybe pinpoint on exactly how tightly are modern Ashkenazi Jews descend from the Sarfatim not necessarily Norwich, but the ones who actually lived and stayed in northern France. I mean, this is quite something that, at least I'm looking forward to observing and studying and seeing being studied in the future in future research.
Well so you know, we let me let me talk about the bottleneck a little because it's been alluded to and kind of take it for granted, I guess, anyone who's interested in genetic, Jewish genetics, Jewish genealogy. So, you know, as I said earlier, in the podcast as when we started out, Ashkenazi Jews share all of these medium length segments. And that's a way that you date the period when the bottleneck occurred. The bottleneck seems to have occurred for more than one generation, there's not a shocking thing. I mean, out of Africa bottleneck occurred for 1000s of years, it seems, for example, so you see a pattern in the genome have certain length segments. So you initially start out with an inbred population that went through a bottleneck with really long segments, the segments get chopped up through recombination over time, and so you can estimate back when it happened. So it's on the order of 1000 years, like we can get more specific, there are specific point estimates that come out, like well, 50, etc, etc. These ancient DNA to me, suggested the bottleneck. Maybe it was a little older than some people would say. Also, they found Some of these populations, I think that the ME group, you know, the Western Ashkenazi in the Erfurt paper, and also some of the individuals in the Norwich, they're quite inbred, and they're affected population size is quite low, lower than modern Ashkenazi Jews. So one inferences you can make from that is you actually have an admixed population where the effective population size is larger, because there's multiple populations that are merging together, that inflates the number of genealogical ancestors. So the, the issue is like in the medieval Jewry of this period, you know, what I have hypothesized when I've suggested and other people have to, there are some sort of metapopulation, where there's these different fragmented groups that are going through bottlenecks. Now, if you merge them together, and they're genetically distinct, then their descendants will have less inbreeding or less, you know, of a bottleneck effect. So it's a little complicated here. And also, the historical records show actually a non trivial number of Jewish communities across Europe at this period. And so people are always like, Oh, but like, you know, it says there's like 1000 ancestors, what's going on there, aside from the fact is reproductive variants. So you have to multiply that 1000 by a lot. It could make sense if, candidly, a lot of these Jewish communities that are recorded historically are dead ends, you know, so there could have been a lot of Jewish communities, but only a small subset went through, you know, continuity into the massive demographic expansion that we see really in Poland, Lithuania in the 17th and 18th century that gave rise to most of the Ashkenazim today. So with that, I'll let you guys comment on what Ijust said there.
No, I actually really appreciate the fact that you added some complexity to what bottleneck actually means and what that might mean for us who like to study it. I would like to just start by saying that there's been a lot of papers this year, that have been using new methods for talking about bottlenecks, drift and founder events. And one paper in particular, that I love mentioning is Tournebize et al. , 2022. And it found that a lot of other Jewish groups have undergone bottlenecks or founder events, or whatever you want to call it. With under like similar intensity, and for similar like length of time, as Ashkenazi Jews. And so just want to shoot that out that a lot of Jewish communities that as we really think about them, had similar dynamics, as some communities, I'm sure were dead ends to later ones that would, would lead. But specifically, I'd like to say that, what when, when, like a researcher tries to calculate this Ashkenazi bottleneck, they're always going to be ignoring the fact that in reality, it is actually a bunch of like, small founder events in towns specific towns, that over the course of hundreds of years, and migrations and population expansions, and just demographics sort of average out to be statistically what they're detecting in this test. And I think it's it's really meaningful to just appreciate for a second that the Jews, let's say that survived the massacres during the first crusades. Like those, like that is a founder event right there. But also, like the Jews that were alive after the black death were that's another founder event, and the Jews that happened to make it that far north in the first place. That's a founder event. There's so many steps in which the sort of breeding population could have diminished or just an actual, like town's Jewish community was literally founded, all of these come together to be what we understand as this sort of Ashkenazi bottleneck. And I'd like to note real quick that in some PCAs that I've seen that I feel can actually look at this drift and visualize this drift, the Ashkenazi drift that peaks let's say in the Erfurt ME samples. The Norwich Jews are only about halfway down the cline between Western Jews on the on one end, which is like normal Sephardim like from Turkey or also Romaniotes on one end of the cline and then Erfurt ME and modern French and German Jews on the other end of the cline the Norwich Jews are right in the middle about together with modern Italian Jews. And that's Pretty significant to me and telling that like the bottleneck took hundreds of years to accrue over time. That's like the drift we're talking about, even if there were specific points of like, wow, that was a significant event. Wow, that was another significant event. It, we're oversimplifying it in a lot of ways.
I totally agree with Leo here. And I also like to comment that the, the nature of also Jewish culture and dynamics in terms of avuncular marital, marital loss, etc, what what is called keeping close in the family is also something which, which might have, in some way influenced that, I mean, many communities might have in some way been founded by small amount of people, even though there are many of them, and just the fact that they had no other people to join them. They had, they had significant amount of children and in some way, you know, cousin marriages and cetera, et cetera, that would also significantly account for some of the heterogeneity. This is essentially the heterogeneity that we're talking about the endogamy that we're talking about. But I think that something which we also need to remember about the bottleneck. I mean, no doubt it occurred, we see it in the - in the hereditary diseases, frequencies, we'll see it in the uniparental variety that we see, we see it in everything, even when we look at the basically the length and the ROH and all these other things. We see it happen. But the question is, which Leo brought here, how many times it's happened, like in how many waves if I can say it, it happened. And I think the consensus currently is that it happened once in a very significant way. But I don't think we can completely take off the table, the fact that there might have been, as Leo, adequately said that these these events happened in other locations. And if you look at the Norwich samples, you can actually see already a cline more adjustment in some of the PCAs, which we observed a cline from North Italian Jews or Italian Jews, actually northern Italian Jews, to Western European Jews, like modern German Jews are Erfurt ME so even in those medieval communities, despite the fact that they might post - be post bottleneck. You already see, by the way, as the Erfurt paper itself mentioned without Of course, referencing to the Norwich paper, there was a larger level of heterogeneity, despite the fact it was after the bottleneck. So there was farther pressure, be it explore explosions be it pilgrims be it massacres, further pressure, maybe small communities and cousin marriages, on genetic pressure on drift in intermarriage and endogamy which might have skewed, if I can say it. What you see in pretty much a shaped, modern Ashkenazi genetics, as we'll see today.
Yeah, okay. So, um, we've been talking about a lot of different things for a while. I think, at this point, like maybe, like, what do you guys? What do you guys know now that you didn't know before? Like, give me what was the biggest surprise the most interesting thing for you from this paper? I mean, I will say I'll start off, I'll say I wasn't entirely shocked that there was heterogeneity in 1359. Among what we now today call Ashkenazi Jews in terms of ancestral variation, but it was unknown, say gratifying, but it was it was good to know and start start to characterize it because modern Ashkenazi Jews are very, very homogenous. They don't show very much geographic structure. They do show a little little bit. I think people, you know, some of the training sets for some of the DTC companies underestimate the difference between say Litvaks and Galitzianers right but you know, whatever, it's all it's all like one big happy was one big family. Okay, let's put that way. But that was a surprise to me. Yeah, I guess I did. I did. I did change that last. So I mean, why don't why don't we go Aric Josh and Leo like tell me tell me what you find surprising about the paper.
So for me, there are several things which surprised me first of all, the fact that N9a3 was found so early on if If I can say about uniparental the East Asian uniparental, which was found the fact that it was found in what can be termed as the Caananic derived Erfurt, I’ll be careful here. But I mean, if I can be more, you know, speak freely, the Erfurt EU slash Caanaic, or more Caananim cananic kind of, of community. It's not that shocking. But it was shocking the fact that the level the amount of hereditary diseases, which were quite similar to what we've see in modern Ashkenazi, the fact that in many ways, they were so similar. I mean, I already expected that it would be similar to modern Ashkenazi, but just how similar to modern Western Jews. And also, a lot of the debates that the three of us had between Rhineland and Bavaria, Austria, Bohemia, the acronym BAB, which is loved by Josh, and Leo and take it from Beider. But you know, the split, the nice split between the different lineages was, was quite, quite nice to see. And, of course, if it's not for what we discuss and learn from Erfurt it is also I mean, the enjoyment of working with, with with, with Shai I mean, the openness that he was, to all our suggestions, this, this actually was very again, I need to stress and very, very enthusiastic and actually a very enlightening experience for me, please.
Yeah, I second, a lot of what Aric said. Shai Carmi, is a gentleman and a rock star. But what I would say is, I think what shocked me, it's not a matter of the results, it's just about our luck is that if you had had as an opportunity, a set of several dozen Jewish individuals from a medieval cemetery in Mainz in Western Germany, or one of the many other prominent, well characterized Jewish communities in Western Germany, at that time, you would not have had a bombshell paper like this, it would have been impressive, but Erfurt just by pure luck. And there's the whole story that you can read in the New York Times about the circumstances that made it possible to excavate and then do genetic analyses on these Jewish individuals remains. Efurt is one of those places where one would have hypothesized to find these two distinct communities coming into contact in the late Middle Ages. Lo and behold, you do and I think there was more that didn't surprise me. Once that was in the picture, then did we really, we did our homework we did our hypothesizing and a lot of what we would have expected and hoped to show up, did. What surprised me most is that one of these groups, despite being buried side by side with the rest of the Jews, despite by all accounts being fully accepted, as Jewish had so much more northern European admixture, and what it really speaks to is something that nobody and I think Razib explains how this could have happened, because of the differential pace of Christianization in different parts of Europe. For all of the things that have been hypothesized about these two founding, medieval Jewish communities, north of the Alps, nobody ever hypothesized that one would have been so much more admixed with Northern Europeans than the other. That's that's the kind of thing that you can only get from genetic evidence.
Amen to that. True.
Yes, exactly. And I would like to reiterate, reiterate, like Josh said, just how lucky we got at every step. Not only did we get, you know, a set of ancients that caught with with enough like high coverage, individuals on both sides. A really interesting example of diversity and like merger of community, but also with like the, the census that Josh also mentioned, mentioned earlier, that we know so many of these Jews that might Be the samples, we're talking about their names, their place of origin. We have that in the historical record to balance this with that they tested isotopes. And we were able to confirm that there's real difference on every step of the way. Everyone got so lucky. And I just want to shout out how incredible that is. But in terms of what's more surprising to me two things, one, that several of the lineages the unit, parental lineages that we found in the samples are either probably or definitely extinct. In modern Ashkenazi Jews, there are several Y chromosomal terminals, and some mtDNA as well, that are just not known, even in big samples of modern Ashkenazi Jews. I think that's pretty interesting and shows that this - that Erfurt also predated another sort of founding event, probably in some way that there were lineages there that, again, are not known. Also, I'd like to shout out one of the samples, which is an outlier was the only outlier in the isotopes tested and also is one of the biggest outliers in the in the PCAs. That focus on on the drift, like I was talking about earlier, it's the only possible sample in the set that you could claim wasn't born in Central Europe, and was actually a Jew from somewhere else. And it's still being tested Most certainly. But there's one sample in there that possibly wasn't wasn't from Central Europe. And even though they resembled Erfurt ME, might have been what I think probably an Italian Jew whether from Northern or central Italy.
All right. Yeah, so we've been going for a while we hit a lot of topics, I really appreciate how deep you guys dug into this. I mean, you know, again, for the listener viewer, the papers, genome wide data from medieval German Jews, show that the Ashkenazi founder event predated the 14th century. So the title I think, kind of undersells really what's in the paper, there's a lot of things that are in the paper. You know, we talked about the demographic history, the population origins, there's some stuff with the diseases of the functional genomics aspect, which is obviously a big reason, or a big aspect that's motivating Jewish genetics. So I guess, like, you know, to close out, is there anything that you guys would like to say, you know, to the listener to the viewer, just that you haven't said yet? And I think, you know, giving you some chances here to speak your Speak your mind. So it's fine if you guys want to pass but let's just go, Aric, Josh, and Leo in that order. How about that?
Right. So
Don’t give up if you have a hobby, a hobby that is extremely complicated, scientific wise, even though you're just an amateur, keep being keep involved, keep it through, you might get get lucky and get involved in an actual study as we did. And I even got, as as, as Josh said, in the email, I even got the chance to now work with Shai on future papers, and in more aspects of this samples and other future samples. But it's a shout out to everyone which is interesting, scientifically curious, keep going keep contributing. You can see that even as an amateur you still can make an influence you can you still can make an impact. And I mean, that's that's my that's my biggest shout out.
Yeah, I would say also, these hobbies are a great way to make unexpected friends I've spent time with Razib in San Francisco and with Leo in New York and with Aric most recently in Jerusalem. So that's just a beautiful thing. How many years of talking on line can can lead to real friendship, but also I want to say, for any Jews out there who are interested in their unique parental lineages why DNA mitochondrial DNA, we We continue kind of in a in a mostly non official way to be involved in our own projects surrounding that. Please reach out to us if you're curious to learn more about yours or want to find out what it is the state of knowledge, these days that we're trying to advance is lightyears ahead of what's been published in the academic data. And there are a lot of stories that have yet to be told properly.
Yeah, I would just like to say that my experience has been crazy as as someone who's on the younger side, and that truth is complex as hell, and I hope to learn more about it soon.
I would just like to point out something about Josh, for the we're actually in some sort sort of a race here to get what I would say, samples or Jewish samples, which might teach us on our peoples history, which might not be possible in a generation or two, for good reasons. This time, not any massacres or pogroms, But actually, you know, Jews all over the world. They intermarry, which is good, it's good for the health is good for whatever reasons for love. But the chances to actually get, what would what I would call an academically, passing samples, like four generations, four grandparents from the same community, that window of opportunity is narrowing and closing down. And if there is some Jews with I don’t know more unique communities, like as Josh said, Western German Jews, or even Romeni Jews, which I know Josh has been in some sort of relationship in the in New York, which are agreed to get sample they will learn something about their ancestry, and at the same time will contribute immensely to our peoples history.
Yeah, no, that's that's a point well taken. This is true for many endogamous communities. And it's one of those bittersweet things right. In terms of us history and genealogy nerds - these well attested and documents lineages are gold. But in terms of individual outcomes, um, there are some issues. I do want to say this is unrelated to this topic, but I get questions about this all the time. You know, just because your parents aren't officially cousins doesn't mean that they're not really cousins, if you come from an endogamous community, inbreeding is not some magical cultural thing. It's just a statistic that emerges from common genealogical ancestors. So if you're from an Indian caste, where they've only been marrying the same people for 2000 years, I don't really care if you guys have complex genealogies where you don't intermarry with your cousin, you you already are, you always will. So, you know, some of this has happened. You know, and I think the bottlenecks that you guys are talking about, you know, that's due to culture, that's due to common cultural patterns, where, you know, in Christian Europe and in Muslim in the Muslim world endogamy was enforced on Jews, what it wasn't during classical antiquity, when converts and proselytes came into the community. And I think maybe what we're seeing in these eastern Jews is a recapitulation, a little bit of the classical pattern, where, you know, pagan populations obviously don't have an ideological relationship to Jews that Muslims and Christians do. And so you have this sort of permeability. Now, in Kaifeng, China, that permeability was a little too permeable from the Jewish perspective. So it's a fine line. It's a fine line. And so again, you know, there's a lot of complexities and you know, I wish that there were more communities like you guys in various geological and genetic just endeavors because we do have a lot of data out there not everything is gonna get published. Some of the can inform the publishing people like David Reich. I don't know Shai personally, but he seems like a great person. I do know David, Nick Patterson, these guys are really really these guys and girls are really really open. Reach out to them, you know, they have not too much time but they're really interested in the truth. That's why they're in academia. And you know, if you're interested in the truth and you're not in academia, who cares? It's just like, you know, letters after your name institutions. You know, these guys are now on a paper that's it's Cell is it in Cell or ACH? Is it so yeah yeah, so they have a paper these guys have a paper now in Cell which is like, you know, Nature, Cell and Science are the three marquee publications really, especially in biosciences, but in general, you know, no offense, but you know, this A lot of physics people that just have a hard time getting into nature and science. And obviously so. So you know, they have other publications. But this is a big deal in the academic world. And so, you know, the love of learning and to try to find the truth like it yields results, right. So I really enjoyed this conversation. I hope that the listeners and viewers really enjoyed it. I hope you get a lot out from the nerd get a lot out of us nerding out on this, because there's just so much information. There's so much to learn. We have the world at our fingertips with the internet today. And I really do wish, you know, this was a Gen X, techno nerd dream that more of us would get involved more of us, our species would get involved in this stuff. Rather than watch the latest TikTok’er I guess is the term or Instagram reels. I'm just, I'm editorializing here. But you know, this is what the internet was great for, you know, I have met Josh, I hope to meet you, Aric and Leo someday, you know, knock on wood. I want to go to Israel someday and visit for sure. By the way, Aric. And, you know, in this way Internet brings people together. And more information out there we learn we contribute to the body of human knowledge. And so, you know, this is definitely a positive thing in this world. Even though we have a lot of negativity going on in various ways, I'm sure that we could talk about like, we focus on the positive today. And that was great. And I hope the listeners and the viewers really enjoyed this conversation. Again, Josh, Aric, and Leo, thank you for coming on. Thank you for giving of your time, obviously, on this podcast, but also thank you for giving us your time to contribute to this publication to the body of human knowledge like this is, this is what we should be doing more of as a species in my opinion. But that's those are my values and views. And I think a lot of people, if you're listening to this, you probably do agree with me, I am probably preaching to the choir, and I know I am with these guys. So I'm with that. Thank you for listening to Unsupervised Learning. And I think I'm going to be posting this before Christmas. So Happy Hanukkah and Merry Christmas to everybody out there. And check out my substack razib.substack.com I don't normally say this when it says one time because I think a lot of people are gonna be listening to this, they're gonna be listening to the end, please rate and review my podcast positively wherever you listen to it. And the reason I ask is that it gets more distribution. And I think, you know, people do want to hear and listen and view this sort of content. It's just can they find it? You know, I know I do. And then I stumble on something. And I'm like, Whoa, this is like a goldmine you know, and so that's why I'm asking you guys to do that. So please, if you could, and with that, I think that that's about it. And thank you again for your time and the publication ‘genome wide data from medieval German Jews shows that the Ashkenazi the founder of that predated the 14th century’ it is In Cell and obviously it is linked in the show notes for this episode thank you again
Is this podcast for kids. This is my favorite podcast!
