Share Your IllustrativeDNA results Ancient and Modern

[qh777]

I personally only like using Eurogenes and only on Gedmatch with modern populations. I don't like most of the calculators on DNAGenics. I mainly used the site to merge together my DNA files into one.

 

[ssha_cha7]

I personally only like using Eurogenes and only on Gedmatch with modern populations. I don't like most of the calculators on DNAGenics. I mainly used the site to merge together my DNA files into one.

Agreed. Most of them leave me somewhat wondering... However there seems to be this very critical advantage:

"Adjusting the distance parameter in the analysis will change the number of samples in the output. A smaller distance will result in more samples and more distant ancestry being included, while a larger distance will result in fewer samples and closer ancestry being included."

This seems invaluable.

It seems to be based on the same technique that the "timeline" tool uses. And frankly it is quite accurate (e.g. Siberian was 18 generations ago).

It's a feature I've leveraged but because my understanding was it adjusted for "trace" which is of great interest to me. It seems that's only one aspect. A closer look will be necessary

 

[ssha_cha7]

On second look I think I'm misinterpreting the word "distant" as time, instead of genetically distant. I still value the tool and even the Chebyshev over-fits, partly because there seems something honest about them, in the sense that we are pieced together from all over. And how these pieces interact.

Anyway it's only the beginning of this tech. So probably about time for me to call it until the next level of things

 

[Who Cares?]

First off, I don't trust any of the calculators like G25 or K13/15/etc.

Second, and about DNAGenics, the only two things worthy of mentioning there is Shared Root Matches (because they're using IBS methodology) and they tell you on which parts of your genome you don't have any historic matches.

 

[qh777]

First off, I don't trust any of the calculators like G25 or K13/15/etc.

Second, and about DNAGenics, the only two things worthy of mentioning there is Shared Root Matches (because they're using IBS methodology) and they tell you on which parts of your genome you don't have any historic matches.

I liked looking at some of the matches from the shared roots but there are some where I match like some Iron Age Italians but on 23&me it says I don’t. One thing that bothers me is that on DNAGenics they include ancient. Egyptian jk2888 as a potential match but it’s been labeled as contaminated so it cannot really be considered. As for G25 I get results that more or less make sense. And I have checked it versus actual DNA matches and the fit numbers are reasonable. For example I match Bal002 which is 500ad Pictish individual. Not sure on the archaeological ID not on pc to check. There are merged g25 coordinates for it and my distance was 0.029 which is reasonable given my ancestry and other dna test data.

 

[Who Cares?]

I liked looking at some of the matches from the shared roots but there are some where I match like some Iron Age Italians but on 23&me it says I don’t. One thing that bothers me is that on DNAGenics they include ancient. Egyptian jk2888 as a potential match but it’s been labeled as contaminated so it cannot really be considered. As for G25 I get results that more or less make sense. And I have checked it versus actual DNA matches and the fit numbers are reasonable. For example I match Bal002 which is 500ad Pictish individual. Not sure on the archaeological ID not on pc to check. There are merged g25 coordinates for it and my distance was 0.029 which is reasonable given my ancestry and other dna test data.

Yes, regarding this Egyptian JK2888, this dumb calculators predicted I have something to do with him, but when I got the actual IBS results they of course did not had me having any African sample. I did get some exotic samples from Asia and North America but that's it

 

[ssha_cha7]

First off, I don't trust any of the calculators like G25 or K13/15/etc.

Second, and about DNAGenics, the only two things worthy of mentioning there is Shared Root Matches (because they're using IBS methodology) and they tell you on which parts of your genome you don't have any historic matches.

It's all just for fun for me. At any rate DNA has lots of unfun applications. Do I grow toenails like a Celt or a Germanic? It's all merely to do with piss and blood.

I imagine, whatever your side or vision, that the uses of DNA once it becomes powerful will sadly become commonplace, political, but leveraged for health.

These are the carefree innocent days.

For today, health is the aspect I ~100% ignore.

Everything is in good fun. For now

Not to be overly serious, but the entire foundation of DNA testing for ethnicity purposes is sand. This is 10000% subjectively my opinion.

 

[Who Cares?]

It's all just for fun for me. At any rate DNA has lots of unfun applications. Do I grow toenails like a Celt or a Germanic? It's all merely to do with piss and blood.

I imagine, whatever your side or vision, that the uses of DNA once it becomes powerful will sadly become commonplace, political, but leveraged for health.

These are the carefree innocent days.

For today, health is the aspect I ~100% ignore.

Everything is in good fun. For now

Not to be overly serious, but the entire foundation of DNA testing for ethnicity purposes is sand. This is 10000% subjectively my opinion.

Yes, but fun has to be based on some solid evidence. These calculators are good estimators, but the estimations can also go wrong and become misleading. That's why I said finding related historic samples by IBS methodology is only relevant thing in all of this.

 

[qh777]

Yes, regarding this Egyptian JK2888, this dumb calculators predicted I have something to do with him, but when I got the actual IBS results they of course did not had me having any African sample. I did get some exotic samples from Asia and North America but that's it

I just feel like their algorithm is too loose DNAGenics. I get matches from the Americas too which leads me to believe that. On IllustrativeDNA some of the three way models assign <1% of various Native American groups which I attribute to genetic overlap between Europeans and Native Americans having Ancestral North Eurasian admixture.

Not to be overly serious, but the entire foundation of DNA testing for ethnicity purposes is sand. This is 10000% subjectively my opinion.

I can understand while you feel that way. There is quite a lot of genetic overlap between particular groups such as Northwest Europeans and it's not necessarily because of post A.D. 476 and the Viking age. For Instance when modeling steppe/ANF/WHG admixture, Celts, Slavs, and Germanic peoples broadly have similar Steppe/ANF/WHG admixture. And some of that has to do with interactions that took place long before the Romans and Vikings. There is some correlation if the right samples are used for your own personal ancestry modeling. I've made fun lists in the past and I wouldn't completely discredit them. If anything it illustrates the interconnectedness of Europe, the Near East and West Eurasia as a whole.

The one slight difference is that the Scandinavians have a bit more Steppe than the other four.

 

[ssha_cha7]

I guess it's interesting that everyone is researching for a different reason. For me my questions are very small or match-based.

Just for your own comparison with these new coordinates and reckless Chebyshev experiments it similarly likes to use especially South American to balance. First time I've encountered this, usually it's an odd Siberian offshoot like Evenk

All in all this is a dangerous past time

 

[qh777]

I guess it's interesting that everyone is researching for a different reason. For me my questions are very small or match-based.

Just for your own comparison with these new coordinates and reckless Chebyshev experiments it similarly likes to use especially South American to balance. First time I've encountered this, usually it's an odd Siberian offshoot like Evenk

All in all this is a dangerous past time

View attachment 17320

Tiwanaku are a pre Columbian native American civilization that existed around Bolivia in South America. I put them in the list because they have left behind some interesting ruins. As I mentioned in another post when I see native American results like that, I think it's from the shared ancestry (overlap) between Europeans and Native Americans from the upper paleolithic population known as Ancestral North Eurasians. ANEs abbreviated. Northern Gaul would not be isolated from Southern Gaul.

This is a bit of info if you click on Iron Age Briton in IllustrativeDNA. It mention's what I have screenshotted.

This is the study that discusses this large-scale migration from France into Britain starting about 1300 BC : https://www.nature.com/articles/s41586-021-04287-4

So If you have British Ancestry I wouldn't be too shocked to see that.

 

[ssha_cha7]

Tiwanaku are a pre Columbian native American civilization that existed around Bolivia in South America. I put them in the list because they have left behind some interesting ruins. As I mentioned in another post when I see native American results like that, I think it's from the shared ancestry (overlap) between Europeans and Native Americans from the upper paleolithic population known as Ancestral North Eurasians. ANEs abbreviated. Northern Gaul would not be isolated from Southern Gaul.

View attachment 17327

This is a bit of info if you click on Iron Age Briton in IllustrativeDNA. It mention's what I have screenshotted.

This is the study that discusses this large-scale migration from France into Britain starting about 1300 BC : https://www.nature.com/articles/s41586-021-04287-4

So If you have British Ancestry I wouldn't be too shocked to see that.

I'm on board. South Gaul is just my big fingers

The ANE however is new. The good old days in Siberia, perhaps. Extremely remotely maybe Denisovan overlap, maybe a Paupan sample if it's not already there, would appear

 

[qh777]

I guess it's interesting that everyone is researching for a different reason. For me my questions are very small or match-based.

Just for your own comparison with these new coordinates and reckless Chebyshev experiments it similarly likes to use especially South American to balance. First time I've encountered this, usually it's an odd Siberian offshoot like Evenk

All in all this is a dangerous past time

View attachment 17320

I'd definitely reduce to remove that

 

[Who Cares?]

I just feel like their algorithm is too loose DNAGenics. I get matches from the Americas too which leads me to believe that. On IllustrativeDNA some of the three way models assign <1% of various Native American groups which I attribute to genetic overlap between Europeans and Native Americans having Ancestral North Eurasian admixture.

It means you just share some SNP with that sample. It could be at any given instance of time. It does not mean that you descend from that sample.

Some of the samples I got were about 20 000 years old, so its normal they have a lot of descendants and all over the world right now.

 

[ssha_cha7]

I'd definitely reduce to remove that

Did so with Scandinavian, Langobard, and Goth removed. Enough overlap with Sca/Jute, and probably not too much Langobard or Goth in Britain.

Basically Jute and Frank are perhaps proxies for Scandinavian and more continental Germanic. Seems like a pretty balanced mix.

You might also see how long this model froze my phone

 

[qh777]

It means you just share some SNP with that sample. It could be at any given instance of time. It does not mean that you descend from that sample.

Some of the samples I got were about 20 000 years old, so its normal they have a lot of descendants and all over the world right now.

You are correct even though sample y isn't your great,great,great,great,great etc. mother or father it still is meaningful and incredible. At least to me it's quite fascinating but the SNPs with the native Americans in my case I think would actually be from ANEs and not actually a pre-columbian Cuban.

This one makes more sense as because this may be derived from later admixture events that happened during the Hellenistic period or Roman.

23&Me's Historical Matches section as of now does not have me matching any of the Native American samples or Ancient Chinese Millet Farmers which is why I mentioned that DNAGenics' algorithm is too "loose" so to speak.
it also does not show me matching SFI-35:

But maybe 23&Me could be wrong who knows

 

[qh777]

I do think global is a bit more accurate in my case. I am 75 Brit / 25 Swede, US. There's some Swiss that may be providing Italic in various calculators, but perhaps I read that this is really allocation of some neolithic component. Then there's the trace SSA which appears once and seems remotely possibly a component of Berber, which itself would have been a component of Iberian, etc. Of course it's possible, the Roman Empire had a lot of movement. But doesn't really show elsewhere.

And a whole span of models here, too

Was wondering did you upload you dna data from a dna ancestry test you took or did you upload simulated g25 coordinates?

 

[ssha_cha7]

Was wondering did you upload you dna data from a dna ancestry test you took or did you upload simulated g25 coordinates?

So far MyHeritage and I used that file for IllustrativeDNA.

(Thankfully my autosomal results were very accurate. For the file itself, and matches, I'd probably take Ancestry next)

I just visited the site, and saw the IllustrativeDNA upgrade as well. Not sure of the news, but I'll post any new results

Looks like they developed their own coordinate system which is great. Hopefully their modern admixture will be more useful. And interested to see if any Kievan-Rus or Saami appear, both categories I'm surprised not to see presently.

 

[ssha_cha7]

This post by on Reddit seems to cover it. So small changes to the models, but more samples in a few months. All in all, still not at the snp level. A long journey to go

So I reached out to admins and appearently this "new" system is not exactly a new thing. It is still based on PCA with 25 vectors. They mainly did this to decrease waiting times for the results, meaning results will be ready in 24 hours instead of days. They also claim they did some improvements with the system to provide better models. On top of all, there will be a major update with lots new samples in a few months as well.

 

[qh777]

This post by on Reddit seems to cover it. So small changes to the models, but more samples in a few months. All in all, still not at the snp level. A long journey to go

Nice that’s news to me. They had IDB in development but then it just disappeared on the site