No more modern sheets in G25

[Lucas]

https://eurogenes.blogspot.com/2020/08/housekeeping-stuff.html

Better download modern sheets. Soon they won't be available:(:picard2:



I'm about to phase out the use of the Global25 datasheets with modern-day samples.
.....
Thus, I'll no longer be updating these datasheets and, from next week, I'll also stop linking to them at this blog (like here). The links will remain live for the next few months, so that users can adjust to the change.

But luckily:


Ymir said...

If G25 coordinates in the future will be still compatible with modern (not-updated) sheets?

Davidski said...

Yes, the old and new datasheets will all be compatible with each other, but I won't be linking to the current modern datasheets.

[Lucas]

I will still keep it here for archive and future use (updated today with most recent sheets)


G25 Modern Averages scaled (http://g25vahaduo.genetics.ovh/G25modern-scaled-averages.htm)
G25 Modern Individual Samples scaled (http://g25vahaduo.genetics.ovh/G25modern-scaled.htm)


G25 Modern Averages (http://g25vahaduo.genetics.ovh/G25modern-averages.htm)
G25 Modern Individual Samples (http://g25vahaduo.genetics.ovh/G25modern.htm)

Or download and store on your computer. For few years it will be unique resource.

[Ion Basescul]

Thank god for Vahaduo and your collections on G25. K13 and Dodecad K12b have way more regional samples than Davidski ever had.
One thing that I might suggest for you is to have a K13 list with all the averages (both national and regional) and then perhaps a list with individual samples.
But that will get easily very massive, because only I am ready to give about 700 coordinates from Romania and Moldova, so I am not sure if that will affect the performance of the website.

[Token]

That is a stupid move in a commercial sense, but whatever.

[Lucas]

That is a stupid move in a commercial sense, but whatever.

I hope for few months he will change his mind because of that. Maybe I'm wrong but most people bought his coords for modern estimation. Ancient was a nice addition. Most laymen don't understand those Yamnaya, WHG and stuff.

[Token]

I hope when he will see his Paypal account for few months he will change his mind. Maybe I'm wrong but most people bought his coords for modern estimation. Ancient was a nice addition.
Yes, he lost most of his potential customers out there. Most people outside TA and Anthrogenica don't care about this ancient stuff at all, you know it better than me since you also sell these kind of services.

[Lucas]

One thing that I might suggest for you is to have a K13 list with all the averages (both national and regional) and then perhaps a list with individual samples.


I can but what would be the difference between actual SOURCE list? Give me some details what to do (in PM).

But that will get easily very massive, because only I am ready to give about 700 coordinates from Romania and Moldova, so I am not sure if that will affect the performance of the website.
Better make averages.

[Jana]

Not a huge loss, his modern sheets were trash anyway except maybe for NW Euros.
However his ancient sheets are excellent.

[Lucas]

Yes, he lost most of his potential customers out there. Most people outside TA and Anthrogenica don't care about this ancient stuff at all, you know it better than me since you also sell these kind of services.

Maybe he is in blog and AG bubble and really think most people buy it because of ancients. But almost all posters there already bought it...
Probably he doesn't even know about numerous Facebook groups dedicated to his stuff.

[JamesBond007]

That is a stupid move in a commercial sense, but whatever.

Depends, if the modern samples are untrustworthy, in the scientific sense, which they are than being innaccurate is a commercial liability. The modern average datasheet says I am closest to the Dutch and then Norwegians while I suppose that could be true I am 75% British isles AFAIK so I doubt it. There has always been Scots close to the Norwegians and English close to the Dutch, more so compared to the average, but that does not make them not British.

Davidski is right if you want to know your modern ancestry then you have to create Iron Age or Medieval models. I did that for myself here :

https://www.theapricity.com/forum/showthread.php?329550-England-Germanic-invader-vs-native-calculator&highlight=England+Germanic+invader

[Lucas]

Depends, if the modern samples are untrustworthy, in the scientific sense, which they are than being innaccurate is a commercial liability. The modern average datasheet says I am closest to the Dutch and then Norwegians while I suppose that could be true I am 75% British isles AFAIK so I doubt it. There has always been Scots close to the Norwegians and English close to the Dutch, more so compared to the average, but that does not make them not British.

Davidski is right if you want to know your modern ancestry then you have to create Iron Age or Medieval models. I did that for myself here :

https://www.theapricity.com/forum/showthread.php?329550-England-Germanic-invader-vs-native-calculator&highlight=England+Germanic+invader

We don't have Medieval not to mention IA samples for every country and probably for many years not.
And many nations were created or got self-identification just few centuries ago.

[Chris596]

Depends, if the modern samples are untrustworthy, in the scientific sense, which they are than being innaccurate is a commercial liability. The modern average datasheet says I am closest to the Dutch and then Norwegians while I suppose that could be true I am 75% British isles AFAIK so I doubt it. There has always been Scots close to the Norwegians and English close to the Dutch, more so compared to the average, but that does not make them not British.

Davidski is right if you want to know your modern ancestry then you have to create Iron Age or Medieval models. I did that for myself here :

https://www.theapricity.com/forum/showthread.php?329550-England-Germanic-invader-vs-native-calculator&highlight=England+Germanic+invader

Yes, this is indeed a remarkable issue. Sometimes I'm closest to Bosnians, sometimes to Serbs, sometimes Bulgarians, etc. The list of my top modern populations says I'm closer to Slovenians than to Moldavians too which is questionable. Other than that it's pretty okay I guess. Even though I'm ethnically mostly Hungarian, genetically I'm basically a Serb.

[J. Ketch]

I will still keep it here for archive and future use (updated today with most recent sheets)


G25 Modern Averages scaled (http://g25vahaduo.genetics.ovh/G25modern-scaled-averages.htm)
G25 Modern Individual Samples scaled (http://g25vahaduo.genetics.ovh/G25modern-scaled.htm)


G25 Modern Averages (http://g25vahaduo.genetics.ovh/G25modern-averages.htm)
G25 Modern Individual Samples (http://g25vahaduo.genetics.ovh/G25modern.htm)

Or download and store on your computer. For few years it will be unique resource.
The modern averages can at least be updated with mono-ethnic G25 users.

[Token]

We don't have Medieval not to mention IA samples for every country and probably for many years not.
And many nations were created or got self-identification just few centuries ago.

At the moment only Iberians and Lazians got very thorough studies of the IA period onwards. The modelling of most of the other ethnicities relies on several extrapolations and unverified hypotheses. The situation of Slavs (most of the Europeans) is particularly worrisome: a Czech singleton, two Avars of debatable origins and maybe the Sunghir guy.

Geneticists are too busy sampling every single alley of Bronze Age Europe.

[Lucas]

The modern averages can at least be updated with mono-ethnic G25 users.

I already added two Berbers with strict regional ancestry. They asked me to do that. Looks legit. I saw also their Gedmatch.

unscaled
Berber_ALG_Chaoui,-0.0022,0.0143,-0.003,-0.0225,0.011,-0.0116,-0.0121,0.001,0.0314,0.0152,0.0026,-0.0023,0.01,-0.0123,0.0105,-0.0077,0.0031,-0.0157,-0.0333,0.0039,-0.0077,-0.0334,0.0205,0.0027,0.0058
Berber_Zenatan,-0.0017,0.0136,-0.0003,-0.0215,0.0085,-0.0144,-0.0115,0.0016,0.0319,0.0118,0.0012,-0.0071,0.0145,-0.0102,0.0124,-0.0088,0.0036,-0.016,-0.033,0.0063,-0.0086,-0.0306,0.0173,-0.0098,0.0042

scaled
Berber_ALG_Chaoui,-0.025041,0.145221,-0.011314,-0.072675,0.033852,-0.032351,-0.028436,0.002308,0.06422,0.0277,0.004222,-0.003447,0.014866,-0.016928,0.014251,-0.010209,0.004042,-0.01989,-0.041857,0.004877,-0.009608,-0.0413,0.025266,0.003253,0.006945
Berber_Zenatan,-0.01935,0.138112,-0.001131,-0.069445,0.026159,-0.04016,-0.027026,0.003692,0.065243,0.021504,0.001949,-0.010641,0.021556,-0.014038,0.016829,-0.011668,0.004694,-0.02027,-0.04148,0.007879,-0.010731,-0.037838,0.021322,-0.011809,0.005029

[J. Ketch]

I already added two Berbers with strict regional ancestry. They asked me to do that. Looks legit. I saw also their Gedmatch.

unscaled
Berber_ALG_Chaoui,-0.0022,0.0143,-0.003,-0.0225,0.011,-0.0116,-0.0121,0.001,0.0314,0.0152,0.0026,-0.0023,0.01,-0.0123,0.0105,-0.0077,0.0031,-0.0157,-0.0333,0.0039,-0.0077,-0.0334,0.0205,0.0027,0.0058
Berber_Zenatan,-0.0017,0.0136,-0.0003,-0.0215,0.0085,-0.0144,-0.0115,0.0016,0.0319,0.0118,0.0012,-0.0071,0.0145,-0.0102,0.0124,-0.0088,0.0036,-0.016,-0.033,0.0063,-0.0086,-0.0306,0.0173,-0.0098,0.0042

scaled
Berber_ALG_Chaoui,-0.025041,0.145221,-0.011314,-0.072675,0.033852,-0.032351,-0.028436,0.002308,0.06422,0.0277,0.004222,-0.003447,0.014866,-0.016928,0.014251,-0.010209,0.004042,-0.01989,-0.041857,0.004877,-0.009608,-0.0413,0.025266,0.003253,0.006945
Berber_Zenatan,-0.01935,0.138112,-0.001131,-0.069445,0.026159,-0.04016,-0.027026,0.003692,0.065243,0.021504,0.001949,-0.010641,0.021556,-0.014038,0.016829,-0.011668,0.004694,-0.02027,-0.04148,0.007879,-0.010731,-0.037838,0.021322,-0.011809,0.005029
In an updated version many of the small regional averages could be collated into larger regions or nations. There's a lot of (basically) mono-ethnic people here and on Anthrogenica to add to the averages or create new ones. At the end of the day modern averages don't matter that much and can be a little off; the more people you have, the more margin for error in individuals.

[J. Ketch]

This is not such a bad thing tbh, to stop the constant whingeing from some quarters about unrepresentative samples, and curtail the number of pointless modern G25 calculators being posted.

[Jana]

At the moment only Iberians and Lazians got very thorough studies of the IA period onwards. The modelling of most of the other ethnicities relies on several extrapolations and unverified hypotheses. The situation of Slavs (most of the Europeans) is particularly worrisome: a Czech singleton, two Avars of debatable origins and maybe the Sunghir guy.

Geneticists are too busy sampling every single alley of Bronze Age Europe.

Also obsessed to sample every corner of Siberia/steppe

I don't understand shocking lack of interest in crucial period of late Iron Age to early medieval/dark age populations? That period is key to understand ethnogenesis of most European nations.

Why?

I saw last paper from Croatia coming up, got excited only to find out it's another almost useless neolithic study.

Can't understand thair reasoning.

[Lucas]

I saw last paper from Croatia coming up, got excited only to find out it's another almost useless neolithic study.

Can't understand thair reasoning.

Yep, almost every neo farmer is fucking Sardinian-like. What is interesting here...

[MarlboroFS]

Why Davidski has delete only the modern spreadsheet? He is developing new version for modern samples? He should delete Eurogenes Gedmatch first IMO

[Token]

Also obsessed to sample every corner of Siberia/steppe

I don't understand shocking lack of interest in crucial period of late Iron Age to early medieval/dark age populations? That period is key to understand ethnogenesis of most European nations.

Why?

I saw last paper from Croatia coming up, got excited only to find out it's another almost useless neolithic study.

Can't understand thair reasoning.

I think everyone already realized that there were massive migrations from Anatolia during the Neolithic, and massive migrations from the steppe during the Eneolithic, but they will not stop digging communal farmer graves and kurgans until they confirm that the same occured for every part of Europe. I understand the obsession with the IE urheimat question, it was and still is a major problem in archeology and linguistics, but seriously, what are the chances that it wasn't in the Pontic-Caspian steppe based on what we've seen so far? In the end we will never get a full confirmation, unless someone happens to stumble into an PIE inscription in cuneiform inside of a kurgan.

[Aren]

That is a stupid move in a commercial sense, but whatever.

Since there has been a small but very vocal community that complain about the average of their ethnicity being off, he probably thinks it's unprofessional of him to have a datasheet on his blog with many averages inlcuding outliers.
I agree it's wrong of him. Those that complain about these averages are biased people that want their ethnicity to look more Slavic, less Med, les North African etc and it's never good to cater to these kind of people even if they are right about the average being wrong. We should be just happy that there is a datasheet at all, produced by an amateur available to everyone.

[Mingle]

Since there has been a small but very vocal community that complain about the average of their ethnicity being off, he probably thinks it's unprofessional of him to have a datasheet on his blog with many averages inlcuding outliers.
I agree it's wrong of him. Those that complain about these averages are biased people that want their ethnicity to look more Slavic, less Med, les North African etc and it's never good to cater to these kind of people even if they are right about the average being wrong. We should be just happy that there is a datasheet at all, produced by an amateur available to everyone.

His averages were really bad for some of them. There were also people that used to complain about bad averages on GEDmatch. And those complaints were also valid. But at least with those, the averages were still mostly sensible, just off by a bit. His averages on G25 are riddled with outliers (with outlier range samples outnumbering normal samples in some cases). Another thing is his averages had a small sample size and people kind of muffed by why so many academic samples were excluded.

Anyways, people have the right to voice their disapproval of bad averages. It's not our fault if he can't handle criticism. If we see something wrong, there's nothing wrong with pointing it out. If he doesn't like that we point out such things, then it's his fault for publishing such bad averages in the first place. You can't create a tool and make money off of it, but not be open to criticism about the references you use. That said, I don't think he had any sort of agenda like some people may think. They just seemed like honest mistakes, which could have happened out of carelessness or some other reason.

[Aren]

His averages were really bad for some of them. There were also people that used to complain about bad averages on GEDmatch. And those complaints were also valid. But at least with those, the averages were still mostly sensible, just off by a bit. His averages on G25 are riddled with outliers (with outlier range samples outnumbering normal samples in some cases). Another thing is his averages had a small sample size and people kind of muffed by why so many academic samples were excluded.
I disagree, what averages are really bad in your opinion? They surely exist but I can't think of any off the top of my head.


Anyways, people have the right to voice their disapproval of bad averages. It's not our fault if he can't handle criticism. If we see something wrong, there's nothing wrong with pointing it out. If he doesn't like that we point out such things, then it's his fault for publishing such bad averages in the first place. You can't create a tool and make money off of it, but not be open to criticism about the references you use. That said, I don't think he had any sort of agenda like some people may think. They just seemed like honest mistakes, which could have happened out of carelessness or some other reason.
We have a different mindset then I suppose. To me David didn't need to create such a thorough datasheet with moderns from all over the world, I hadn't heard of maybe 50% of all these ethnicites prior to experimenting with his spreadsheets. So I'm just grateful that this spreadsheet exist. I'm also confident he does his best with the time he has to add the most optimal and available genomes out there. Are there averages that are wrong and include many outliers? For sure, but anyone can see that so there's no reason to get all worked up about it and create threads like the Italian guy here who was so displeased with Campanians being so southern in his spreadsheet(the average isn't even way off it's just slightly more southern shifted than one would expect) that he creates a thread here suggesting David is one of the chosen ones.

It's obvious David isn't deleting the spreadsheet cause he can't handle criticism but because he has never put much importance to it, since his tool is best used using aDNA and cause it's difficult to find available modern genomes who are representative for every ethnicity out there. But when the complaints start getting more vocal he probably just thinks it's better to remove it ones and for all.

[Jana]

I disagree, what averages are really bad in your opinion? They surely exist but I can't think of any off the top of my head.


We have a different mindset then I suppose. To me David didn't need to create such a thorough datasheet with moderns from all over the world, I hadn't heard of maybe 50% of all these ethnicites prior to experimenting with his spreadsheets. So I'm just grateful that this spreadsheet exist. I'm also confident he does his best with the time he has to add the most optimal and available genomes out there. Are there averages that are wrong and include many outliers? For sure, but anyone can see that so there's no reason to get all worked up about it and create threads like the Italian guy here who was so displeased with Campanians being so southern in his spreadsheet(the average isn't even way off it's just slightly more southern shifted than one would expect) that he creates a thread here suggesting David is one of the chosen ones.

It's obvious David isn't deleting the spreadsheet cause he can't handle criticism but because he has never put much importance to it, since his tool is best used using aDNA and cause it's difficult to find available modern genomes who are representative for every ethnicity out there. But when the complaints start getting more vocal he probably just thinks it's better to remove it ones and for all.

Moldovan average sucked. It was more southern than Romanian one.
However in his defence he did upload some newer academic Moldovan averages after Ion Basecul mailed him.

Serb average was pretty off, more southern than reality and more southern than Montenegrin one.
I heard complaints south Italian average sucked too and was based on Calabrian outliers IIRC.

Than you got Polish average, very NE shifted and bit too Slavic for average of Poland (probably because David is NE Polish himself). That way pure western poles like Peterski who constitute important part of Polish population end up scoring Czech before Polish.

Austrian average seemed too much eastern biased (I may be wrong however). Slovenian average is bit too southern IMO, and Croatian is partly Germanic while average Croats don't really have that.

Those are slight mistakes. However he could have used monoethnic customers of his to improve those that needed improvement.

Moldovan average are example of those that were not slightly off, but way off from genetic average of that country.

[vbnetkhio]

Also obsessed to sample every corner of Siberia/steppe

I don't understand shocking lack of interest in crucial period of late Iron Age to early medieval/dark age populations? That period is key to understand ethnogenesis of most European nations.

Why?

I saw last paper from Croatia coming up, got excited only to find out it's another almost useless neolithic study.

Can't understand thair reasoning.

they probably don't care much about modern European nations and find ancient population movements much more interesting.

[Aren]

Moldovan average sucked. It was more southern than Romanian one.
However in his defence he did upload some newer academic Moldovan averages after Ion Basecul mailed him.

Serb average was pretty off, more southern than reality and more southern than Montenegrin one.
I heard complaints south Italian average sucked too and was based on Calabrian outliers IIRC.

Than you got Polish average, very NE shifted and bit too Slavic for average of Poland (probably because David is NE Polish himself). That way pure western poles like Peterski who constitute important part of Polish population end up scoring Czech before Polish.

Austrian average seemed too much eastern biased (I may be wrong however). Slovenian average is bit too southern IMO, and Croatian is partly Germanic while average Croats don't really have that.

Those are slight mistakes. However he could have used monoethnic customers of his to improve those that needed improvement.

Moldovan average are example of those that were not slightly off, but way off from genetic average of that country.

I am or was aware of all these "faults" and none of them are really bad IMO, some that you mention we are not even sure if they are wrong. Such as the Polish average being too NE shifted. Based on what? Peterski's family is from Western Poland, with deep roots there, he should be shifting West/Southwest of the Polish average.

[Ion Basescul]

Moldovan average sucked. It was more southern than Romanian one.
However in his defence he did upload some newer academic Moldovan averages after Ion Basecul mailed him.

Serb average was pretty off, more southern than reality and more southern than Montenegrin one.
I heard complaints south Italian average sucked too and was based on Calabrian outliers IIRC.

Than you got Polish average, very NE shifted and bit too Slavic for average of Poland (probably because David is NE Polish himself). That way pure western poles like Peterski who constitute important part of Polish population end up scoring Czech before Polish.

Austrian average seemed too much eastern biased (I may be wrong however). Slovenian average is bit too southern IMO, and Croatian is partly Germanic while average Croats don't really have that.

Those are slight mistakes. However he could have used monoethnic customers of his to improve those that needed improvement.

Moldovan average are example of those that were not slightly off, but way off from genetic average of that country.

And the GEDmatch Moldovan average was the opposite, a mix of East Slavs and regular Moldovans to produce a Northern Moldovan proxy, which kind of works for me and my family. But the average wasn't produced ethically. I'm pretty comfortable with the average and individual samples on Global 25 right now, as they theoretically cover all the genetic variety, from North to South. But then again, all of them and significantly more are available with the updated Eurogenes K13.

[Jana]

I am or was aware of all these "faults" and none of them are really bad IMO, some that you mention we are not even sure if they are wrong. Such as the Polish average being too NE shifted. Based on what? Peterski's family is from Western Poland, with deep roots there, he should be shifting West/Southwest of the Polish average.

NE Poland has pretty low population except for Warsaw (where live people from all regions) AFAIK, bulk of population lives in south/west. Correct me if wrong.
From what I know Polish average is based on academic Mazovian samples, that region has low population historically if compared to something like Lesser Poland/greater Poland and national average should be average of all regions weighted according to population percentage in overall numbers.

But, there is user Michal who is from south/lesser Poland and he is pretty eastern/northeastern so maybe this average is not bad.

[Lucas]

NE Poland has pretty low population except for Warsaw (where live people from all regions) AFAIK, bulk of population lives in south/west. Correct me if wrong.
From what I know Polish average is based on academic Mazovian samples, that region has low population historically if compared to something like Lesser Poland/greater Poland and national average should be average of all regions weighted according to population percentage in overall numbers.

.

It is only rumour. No direct regional evidence for that in papers. They are just Poles in study. But yes those samples are rather NE shifted but not extremaly.

Of course we don't talk here about so called Estonian Poles average which is disaster.

[ph2ter]

From my own analysis it appears that Polish samples are pretty representative. I managed to group the Polish samples into all main Polish clusters. The same I accomplished with German samples.
I don't think that there was any bias in choosing the samples.
You people act like you don't know or don't want to know how hard is to collect and compute the values for several thousand of samples from the whole world.
And we all got that for just 12$.
Nobody prevented anybody from making their own averages from the samples according to which this anybody have come to the conclusion that Davidski averages are not right.

[Thracian]

I hope he will change his mind.

[J. Ketch]

Re: the focus on Bronze Age and lack of Medieval samples, David wrote this on his blog:


We have to shift the focus to ancient samples, and if users are interested in more recent ancestry, then they should focus on the Iron Age and Medieval periods.

There will be a lot of samples available from the Middle Ages soon, so this won't be a problem.

[Lucas]

. The same I accomplished with German samples.
.

I did it earlier:)

[Luso]

subscribed.