Why a large portion of E-V13 in the Iron Age might have been Dacian

[Archetype0ne]

The method of using TMRCAs of all the subclades to create the graph. Is it reliable? Does it work with other haplogroups?
I feel like since each tmrca on yfull has a really really high uncertainty, this method would be almost useless. I think it's a very good idea but we need better data for TMRCAs before we can apply it succesfully. I may be wrong though.

My 2 cents: I think this type of graph is broadly correct, but not so precise you can label a small uptick "Free Dacian recovery". The first peak in the EBA, followed by a bottleneck in MBA and another big Expansion in LBA is real for sure. So that bottleneck somewhere around AD must mean something, but I think it can be explained in multiple ways.
The data of Riverman's graph is pretty straightforward, but the data from Bruzmi's graph is dependent on the area included, the sensitivity selected, and the level shown, and is very sensitive to certain biases on Yfull. As a result it does not show what Bruzmi claims it shows.

So I would say, take both with a grain of salt.

Since I am kind of nauseated by typing the same thing over and over ill just leave this argument between me and Riverman from yesterday.

Riverman posted this graph making it a foregone conclusion with a wall of text that V13 and L283 have no historical correlation in regards to boom and bust cycles as far as phylogenic diversification in the Balkans. Claiming this was due to lack of shared history.

The first step is to trace down where the major modern Albanian lineages lived at which time, how they migrated and when they had major founder effects. Like we know for Proto-Germanics that I-M253 and R-L106 went in synchrony from a specific time frame onwards, concrete, after the Urnfieldisation/unified cremation horizon of the Nordic Bronze Age was established and confirmed by the time of the formation and first expansion of Jastorf. So we have two events, which likely can be associated with Proto- and early Germanics, like the Urnfieldisation of the NBA and the formatoin of Jastorf.

For the Albanian case we already have some dates for the moderns, and we can see related founder effects, but that's really late. The main Albanian lineages go in synchrony somewhere between 100 AD to 1.300 AD. A good position might be reached, but this is really vague and speculative, around 600 AD.
That's by going after the main haplogroups E-V13, J-L283, R-PF7563, R-CTS9219 (R-CTS1450).

It might look somewhat different already if just counting actual Albanian lineages.

Before 100 AD the general haplogroups (not Albanian specific) are very weakly to not at all correlated. In fact, they are even contradictory, like when J-L283 expanded E-V13 goes down and vice versa, same for R-PF7563, R-CTS9219.





If someone could do such a correlation analysis specifically for the Albanian subclades of these haplogroups, which is somewhat more tricky than for the general one, just going after the TMRCA dates, it could probably help. Because by going after the haplogroups in general, they could have been largely separated lineages up to at least 100 AD, probably even much later.

Because if lineages share a fate, if reaching a certain size, numbers, they usually go in synchrony, just like we can see it for the Germanic and Slavic lineages. If e.g. some E-V13 joined successfully the Slavs and expanded with them, its visible, because they go up with the Slavic main lineages. This is something we can see looking at YFull data alone.

For the Albanian lineages specifically, there were major founder effects, fairly late, these show up. How far back they go, that's open to debate - going by the general data, not before 100 AD, but probably the Albanian-specific data is different - though I doubt it.

Also note the size-numbers of E-V13. It can't have been confined to a small Eastern South Balkan area - if you compare it with the data from J-L283. Much to big and geographically even wider spread.

I see on this thread it he discarded those graphs, as at least in my opinion anyone with a statistical intuition could see the correlation ,be it in the form of correlation indicator, or lagging and leading correlation. Whatever its called, time, time lagged cross correlation?

You know whats funny to me Riverman?
After one conclusion based on the deployment of one method on a data fails, just make another conclusion, make sure that conclusion is in line in regards to negation of the first, and try again.
Last time I saw such methods use was from Ghurier and you over at Anthro, for different hypotheses, but both hypotheses negating the same thing. I am starting to think its not about proving something, its about denying something else. The negation the hypotheses inadvertebly brings here has more value to some than any positive statement coming out of the hypothesis itself.

But maybe its just me.

Ps: I am not saying anything about the hypothesis you bring, just found it funny looking at the method - data being recycled yet again, this time with different conclusions.

Edit: Also:

"Before 100 AD the general haplogroups (not Albanian specific) are very weakly to not at all correlated. In fact, they are even contradictory, like when J-L283 expanded E-V13 goes down and vice versa, same for R-PF7563, R-CTS9219."

Biggest cap.

Are we looking at the same data here? Like what?
When even one owns data contradicts their claims. C'mon are we supposed not to look at the data and just at the wall of text?

Funnily enough the bolded part is the very cap here.
I would give you the benefit of the doubt if we started talking about different time frames in reference to lagging and leading indicators vis a vis the two haplogroup diversification patterns. But even then lagging and leading still implies correlation.

Edit: Doubled checked. Stand by my words.
200 BC to 200 AD correlation stands. What you call "stable moderate Roman era expansion" is no expansion at all, no matter what label you put on it, its a retraction.
You can argue about magnitude of impact. But given the low number of phylogenies it is not out of the ordinary.

Also check 200-100 BC in relative % terms L283 dropped more than 80% while V13 55%? If anything the bigger impact on V13 100 years later was a lagging indicator. When you think what this data shows and how mutations happen 100 years is nothing.

Really this whole method is a mess.
The data is phylogenic diversification based on current clades on Yfull? With magnitudes of 1-14? Little room for statistical error, when you are dealing with random events such as Y mutations...
Even then... correlation among the two branches is staggering.
And you were claiming first there is no correlation? Then changed it upon scurrility with " oh during this arbitrary 200 years in history there was no correlation"... And even then I don't think the data backs you up.

There is formulas if you wanna test correlation. Go at it. And lets see the result.

The burden of proof is not on me here. Just run the data. And lets talk after.



Though I suspect based on the data set, this will be messy.

Edit: A time lagged cross correlation calculation would better fit the purpose here.

https://www.lexjansen.com/nesug/nesug12/fi/fi03.pdf

Something like this would set this straight.

Hint: I have no idea the results for either. But at least I am proposing ways to test it out and not just making statements out of thin air.

Aahmes, this is why I fully agree with you. Its not just that the data does not even support this particular interpretation. Even if it did and was in line with my current hypothesis, the method itself is unreliable.

For TMRCAs as well as random Y mutations 100-200 years are basically statistical errors. Then you add the uncertainty about biased sampling, And finally the low amount of data and potential for high variability when the range goes from 1-14 phylogenies, where one random mutation can impact the calculation at any given time frame by like 7%.... So yeah.

[Riverman]

Like I said before, there are potential in synchrony moments, but there are very significant out of sync moments as well. This doesn't mean however, that its not possible that some E-V13 lineages could have shared prehistory before the common era with some other J-L283 lineages.

We have to look at it region wise too, like for primarily Central-Southern Balkan E-V13 vs. primarily Central-Southern Balkan J-L283. It might be different than for the haplogroups in general. I don't think it will be fundamentally different, but it could be. Because I too said many times that I expect E-V13 to be e.g. in and near Albania latest since the Late Bronze Age.

An interesting but open question is how much of the lineages in e.g. modern Albania of both J-L283 and E-V13 were actually local in the Bronze, the Iron or the Roman period. This needs to be investigated, it can't be just assumed, even if we know that they bordered and mixed with each other.

For TMRCAs as well as random Y mutations 100-200 years are basically statistical errors. Then you add the uncertainty about biased sampling, And finally the low amount of data and potential for high variability when the range goes from 1-14 phylogenies, where one random mutation can impact the calculation at any given time frame by like 7%.... So yeah.

That's what I thought as well, before doing this. But then I looked at the graphs of all the major haplogroups I looked at, and they really correlate pretty well with each other and the major events. So it seems that at least if the numbers are big enough, there is some kind of automatic smoothing of statistical randomness of the data, which makes it fairly reliable. At least more reliable, than I myself initially thought.

[Riverman]

Here another quotation from Wikipedia which underscores the small uptick after Illyrians and Thracians proper being conquered, which I speculated was the "Dacian expansion":

Around 150 BC, La Tène material disappears from the area. This coincides with the ancient writings which mention the rise of Dacian authority. It ended the Celtic domination, and it is possible that Celts were driven out of Dacia. Alternatively, some scholars have proposed that the Transylvanian Celts remained, but merged into the local culture and thus ceased to be distinctive.[151][155]

Archaeological discoveries in the settlements and fortifications of the Dacians in the period of their kingdoms (1st century BC and 1st century AD) included imported Celtic vessels and others made by Dacian potters imitating Celtic prototypes, showing that relations between the Dacians and the Celts from the regions north and west of Dacia continued.[156] In present-day Slovakia, archaeology has revealed evidence for mixed Celtic-Dacian populations in the Nitra and Hron river basins.[157]

After the Dacians subdued the Celtic tribes, the remaining Cotini stayed in the mountains of Central Slovakia, where they took up mining and metalworking. Together with the original domestic population, they created the Puchov culture that spread into central and northern Slovakia, including Spis, and penetrated northeastern Moravia and southern Poland. Along the Bodrog River in Zemplin they created Celtic-Dacian settlements which were known for the production of painted ceramics.

https://en.wikipedia.org/wiki/Dacians#Language

So shortly before the Roman conquest, Dacians were on the move and expanded Westward, into Celtic controlled territories and other areas. Therefore, even the La Tene limited spread of E-V13, the late one, can be in part related to a growing Dacian influence on the Celtic world. What this also shows, this Wiki article, just like most sources I have read, is how closely related the Celtic and Dacian people in Pannonia became. Many Dacians started with Lateinisation with very little to no actual Celtic presence and being later assimilated or lived rather side by side with Celts later.

The Puchov or Puchauer culture of mixed character was a neighbour of the Sarmatian-Dacian mixed groups and the presumably Costobocci Dacians of the Lipitsa culture.

Here some maps, to put things into context for the later Roman era:

https://i.pinimg.com/564x/8e/82/7b/8...70a8101ccf.jpg

As one can see, the "Free Dacians" were still a force to be reckoned with, by that time, before the Germanic and later migrations.


https://upload.wikimedia.org/wikiped...y_Shchukin.png

Its also easy to explain, if looking at those maps, how some minority E-V13 lineages could end up fairly early in some Slavic tribes which might have directly bordered the Dacians and began to expand, again as a minority lineage, with the R-Z280 + I2a-din block. Also note the Puchov/Puchauer culture, especially in the area between the Danube bent and Tisza river.

Before the Roman attack, both sides of the Tisza river became Dacian, largely, because of this Dacian expansion shortly before the Roman campaigns started to destroy the base of the Dacian kingdoms.

The Puchauer-Zemplin culture also explains why we might find many R-U152/R-L2 (Celtic) beside E-V13 (Dacian) in that area, intermixed with incoming Sarmatians in the Roman and shortly after eras. Will be also interesting to see whether the sampled Gepids show some E-V13. In the Maslomecz Gothic group with obvious Dacian and Sarmatian influences, we still have the classical "cremation bias problem".

[Bruzmi]

Before the Roman attack, both sides of the Tisza river became Dacian, largely, because of this Dacian expansion shortly before the Roman campaigns started to destroy the base of the Dacian kingdoms.

The Puchauer-Zemplin culture also explains why we might find many R-U152/R-L2 (Celtic) beside E-V13 (Dacian) in that area, intermixed with incoming Sarmatians in the Roman and shortly after eras. Will be also interesting to see whether the sampled Gepids show some E-V13. In the Maslomecz Gothic group with obvious Dacian and Sarmatian influences, we still have the classical "cremation bias problem".

There is no "cremation bias problem". There are plenty of sites with plenty of samples. Where E-V13 isn't being found, this happens because E-V13 wasn't there.

"Dacian expansion" can't be used to explain the existence of samples in early medieval Tisza from a very different era. In general, the autosomal ancestry of these samples points to areas much more southern than Hungary. It's not a coincidence that in one of the studies in just 3 E-V13 samples, one was the southernmost individual of the entire study with an almost Sicilian-like profile. More importantly, there aren't any samples from Dacia (yet), hence we know nothing about Dacia itself.

The Maslomecz Goths don't have any "obvious" Dacian influence. You're claiming that there is "Dacian influence" in Maslomecz because there is one reported E-V13 sample among them. What you leave out of your narrative is that there are two J-L283 samples in Maslomecz (different research teams). There are more J-L283 than E-V13 samples in Maslomecz. Is this the result of "Dacian influence"?

The J-L283 individuals may have moved there from the northern Balkans or even from areas of Roman Transylvania, but they have nothing to do with "Dacians". The J-L283 sample examined by the Węgleński-Kokowski team belongs to J-Z38241+, an Iron Age Illyrian lineage of Croatia and his autosomal profile points to a southern European origin which is different from that of the other 17 samples (14 I1a, 4 R1a) of the study. A similar route from the northern Balkans is more likely than not for the E-V13 individual who moved to the north with J-L283 to live among the Goths.

[Aspar]

Since I am kind of nauseated by typing the same thing over and over ill just leave this argument between me and Riverman from yesterday.

Riverman posted this graph making it a foregone conclusion with a wall of text that V13 and L283 have no historical correlation in regards to boom and bust cycles as far as phylogenic diversification in the Balkans. Claiming this was due to lack of shared history.




I see on this thread it he discarded those graphs, as at least in my opinion anyone with a statistical intuition could see the correlation ,be it in the form of correlation indicator, or lagging and leading correlation. Whatever its called, time, time lagged cross correlation?







Aahmes, this is why I fully agree with you. Its not just that the data does not even support this particular interpretation. Even if it did and was in line with my current hypothesis, the method itself is unreliable.

For TMRCAs as well as random Y mutations 100-200 years are basically statistical errors. Then you add the uncertainty about biased sampling, And finally the low amount of data and potential for high variability when the range goes from 1-14 phylogenies, where one random mutation can impact the calculation at any given time frame by like 7%.... So yeah.

ArchetypeOne, I don't doubt your mathematical and statistical skills.

But the picture is really simple when we have a tone of aDna J-L283 found especially in Italy and the West Balkans and then little to none E-V13 in those places.
Then again, E-V13 popped up in the aDna record in an IA context further east in Moldova(no J-L283 there) and Slovakia(not sure if there was J-L283 among the Vekerzug samples).

Why do you need complex formulas when you have aDna which clearly speaks against shared history of J-L283 and E-V13?

[Riverman]

There is no "cremation bias problem". There are plenty of sites with plenty of samples. Where E-V13 isn't being found, this happens because E-V13 wasn't there.

"Dacian expansion" can't be used to explain the existence of samples in early medieval Tisza from a very different era. In general, the autosomal ancestry of these samples points to areas much more southern than Hungary. It's not a coincidence that in one of the studies in just 3 E-V13 samples, one was the southernmost individual of the entire study with an almost Sicilian-like profile. More importantly, there aren't any samples from Dacia (yet), hence we know nothing about .

The Maslomecz Goths don't have any "obvious" Dacian influence. You're claiming that there is "Dacian influence" in Maslomecz because there is one reported E-V13 sample among them. What you leave out of your narrative is that there are two J-L283 samples in Maslomecz (different research teams). There are more J-L283 than E-V13 samples in Maslomecz. Is this the result of "Dacian influence"?

It could be, because J-L283 people moved into Thracian territory, just like E-V13 moved into J-L283 territory. From very Late Urnfield onwards, they had many contact points, even a whole contact zone, and there were bidirectional movements for a lot of potential gene flow.

The reason why I write about Maslomecz having Dacian influence is primarily because many authors see this kind of influence in the whole Chernyakhov-Sântana de Mureș complex. In the beginning of the Gothic expansion, Dacian or Daco-Celtic people like the Costobocci and Carpi in particular were still among the leading elements of the Western steppe-North Carpathian coalitions the "Barbarians" formed. This is archaeologically attested, as is the Lipitsa culture and related formations. Even more important, we have such influences on the site of Maslomecz, especially the collective cremation burials, which are something specifically Dacian.

So the single E-V13 found there is just additional evidence, and I wrote before already that I think the results being seriously skewed, because if in sites like Maslomecz we would have all the data from those cremation burials, E-V13 wouldn't pop up way more than once.

I have to agree with you, that this is no proof, its just based on clues and logic, because the final and direct proof can only come from ancient DNA. But then again, how can we get it, if they burnt their dead?

And if you look at the Hassleben samples: The funerary rite matters, it matters a lot! If a people traditionally and consequently burnt their dead, it means nothing if you get samples from people which use a completely different rite, even if its from the same place, because chances remain high, that they were intruders and foreigners.

Like if you have cremation burials in a Jewish or Christian dominated region, its a clear warning signal that these burials might be from non-Jews, non-Christians! Same here.

You can deny all day long, but the cremation burials just mean we have this huge gap for a major people of Europe, actually for several in different periods, like early Jastorf Germanics and Iron to Roman era Slavs too, largely. This means we have to rely on mixed people, but usually the male determined the faith of the offspring (!) or irregular and special burials - as well as foreign influenced groups. But like I have shown to you: Even among the Scythianised Vekerzug people, in the core region of Dacian dominance, they regularly cremated still!

Then again, E-V13 popped up in the aDna record in an IA context further east in Moldova(no J-L283 there) and Slovakia(not sure if there was J-L283 among the Vekerzug samples).

The Vekerzug being split into the inhumation Western groups and the Eastern core group, which regularly cremated. Obviously, we can find J-L283 and all kind of haplogroups in the mixed West, where the Eastern complex had less of an impact and only occasional gene flow. So even for this group, if we would have samples from the core region, which is always East of the Tisza, E-V13 would sharply rise up.
Same with Viminacium: If we would have material from the cremated part of the population, E-V13 would just steeply rise up to even higher levels.

It's just a fact that the core zone of E-V13 cremated up to the Roman era, up to Christianisation, for the most part. This is why we get suddenly way more E-V13 when inhumation gets universal, even in areas in which we can assume the relative numbers went down actually.

[rafc]

As for the Roman era, E-V13 was going down when Rome was at its height and up when it was in crisis and even more up when it crumbled. J-L283 did only go down in the initial conquest, but so drastically still most lineages likely being affected, but then it had a stable growth up to the collapse, exactly when E-V13 goes up with Slavic R-Z280 and I2a-din.

But that would mean you can really perfectly correlate the graphs in time (initial conquest versus Rome at it's height versus decline). With a shift of 200 years that picture changes completely from being caused by Roman invasion to being caused by Roman decline. So for sure there was a bottleneck that impacted V13, but I'm not sure we understand it yet. It's a promising approach though.

[rafc]

The Maslomecz Goths don't have any "obvious" Dacian influence. You're claiming that there is "Dacian influence" in Maslomecz because there is one reported E-V13 sample among them. What you leave out of your narrative is that there are two J-L283 samples in Maslomecz (different research teams). There are more J-L283 than E-V13 samples in Maslomecz. Is this the result of "Dacian influence"?

The J-L283 individuals may have moved there from the northern Balkans or even from areas of Roman Transylvania, but they have nothing to do with "Dacians". The J-L283 sample examined by the Węgleński-Kokowski team belongs to J-Z38241+, an Iron Age Illyrian lineage of Croatia and his autosomal profile points to a southern European origin which is different from that of the other 17 samples (14 I1a, 4 R1a) of the study. A similar route from the northern Balkans is more likely than not for the E-V13 individual who moved to the north with J-L283 to live among the Goths.

Regardless of genetic results or whether V13 is involved, there was clear Dacian and Sarmatian influence in the Maslomecz group.

[rafc]

ArchetypeOne, I don't doubt your mathematical and statistical skills.

But the picture is really simple when we have a tone of aDna J-L283 found especially in Italy and the West Balkans and then little to none E-V13 in those places.
Then again, E-V13 popped up in the aDna record in an IA context further east in Moldova(no J-L283 there) and Slovakia(not sure if there was J-L283 among the Vekerzug samples).

Why do you need complex formulas when you have aDna which clearly speaks against shared history of J-L283 and E-V13?

I think this also shows the limits of using such graphs. In the end the broad picture of V13 will reflect trends over a much wider area. It's initial expansion happened in the EBA when there was population expansion and increasing societal complexity over a wide area. The MBA saw a general downturn with depopulation almost everywhere, while the LBA was a period of renewed growth and expansion. So it's quite likely the peaks and lows of V13 would repeat in a haplogroup that was geographically completely separate from V13.
I agree that the aDNA record is a much stronger indicator, and it clearly shows that V13 did not have a shared history with L283 until the late Iron age or even Roman age. Take that L283 group from Croatia that was also found in Italy, Tunesia, and now with the Goths. It's only in that last instance that it was found together with V13, which makes it unlikely they shared a long history.

[Riverman]

But that would mean you can really perfectly correlate the graphs in time (initial conquest versus Rome at it's height versus decline). With a shift of 200 years that picture changes completely from being caused by Roman invasion to being caused by Roman decline. So for sure there was a bottleneck that impacted V13, but I'm not sure we understand it yet. It's a promising approach though.

But J-L283 and other lineages did well or by far not that bad, so the question is why was E-V13 so massively, negatively affected, when Rome was at its height. The J-L283 downturn is way before the Roman height being reached, its exactly when the Illyrians and Thracians being conquered and suppressed (Illyrian wars, Macedonian wars, Illyrian and Thracian uprisings) - I highlighted the J-L283 pattern for the Roman era:


They go down with the Illyrian wars, and they go up once being colonised.

E-V13 goes down too in that phase, I would bet on some Illyrians being hit harder plus the Thracians which were hit hard too. But they keep half of their growth, presumably even different lineages, which corresponds perfectly with the strong Dacian proportion among E-V13 carriers.
They just go down, with the Dacian wars.

The Illyrians were defeated and largely integrated by 200 BC, finally about 160 BC:
https://en.wikipedia.org/wiki/Illyrian_Wars

The Dacian Wars (101–102, 105–106) were two military campaigns fought between the Roman Empire and Dacia during Emperor Trajan's rule.

https://en.wikipedia.org/wiki/Trajan%27s_Dacian_Wars

So the Illyrians grew, once pacified, in the Roman Empire, but E-V13 only grew with tribals.

This doesn't speak, in my opinion, for the E-V13 carriers being not present in Roman territories, not at all. But the J-L283 might have had a more favourable position, becoming better integrated or keeping up more tribal ways of life, better for their demography. The E-V13 carriers on the other hand just had losses within Rome, zero growth. All their growth in that time period and up to the Early Medieval period seems to have come primarily from the tribal and free from Rome areas, like from the Free Dacians and those lineages which joined the Slavs.
Clearly, some Illyrian tribes quickly joined Roman ranks or became valuable allies, soon after pacification, some even before. The Thracians and Dacians, as well as many Celtic and Sarmatian tribes, on the other hand seem to have been much less successful in the Roman sphere.

The timing differences come from different conditions and positions within the Roman sphere and beyond. Note especially the contradictory movement of J-L283 vs. E-V13 around 600 AD, its exactly the date of R-Z280 and I2a-din going straight up in one of the biggest Slavic expansions. And E-V13 lineages being very obviously part of it. If counting them, they probably are just about 1:20, which is, funnily, about the number E-V13 has today in many Slavic regions (4-5 %).
At the same time, J-L283 goes straight down, has a significant dent. They were not growing with early Slavs like E-V13.

This also means that the E-V13 in the Balkans had very little growth, throughout that whole period. Most of the growth came from tribals at the fringe and from outside. About 900 AD, J-L283 and E-V13 going in synchrony on a large scale, that's the Vlach and Albanian expansions starting to get significant. This also when R-CTS9219 joined ranks, with these Balkan groups, which is quite telling as well. Contrary to that, R-PF7563 looks like a low level independent player throughout all those times. Having no big wins or losses, just existing on a fairly low level.