PDA

View Full Version : Do you trust 23andme's Ancestry Composition or GEDmatch calculators more?



Pages : [1] 2

Sikeliot
05-27-2015, 02:00 AM
Some people say that 23andme predicts ancestry from the last 500 years, and that it is therefore more reliable than GEDmatch.

I don't really agree because I've noticed very peculiar things happen with the 23andme calculator. I've seen families where one parent is Sicilian and the other is Irish. The Sicilian parent scores 15% Middle Eastern, while the child scores 0%. I've yet to see a half North European, half SE European (Cretan, Sicilian, Greek islander) score significant Middle Eastern even though full members of said groups score a significant amount.

Likewise, I suspect 23andme underestimates admixture when it is small. I have seen that in people with under 10% of African, the African is greatly underestimated on 23andme.

What do others think? I think due to its algorithms, 23andme ends up giving misleading results, but it is good for telling if a given ancestral component exists... just not necessarily the amounts.

Tįltos
05-27-2015, 03:13 AM
Well here is the thing, there is NO standard to how these admixture tests should be analyzed between the various companies and third party tools.

I personally prefer the analysis of 23andme, because it is conservative. I do enjoy thinking about deep ancestry, but I'm very interested in connecting with my roots in a paper trail time frame. If that's even possible! :)

Trojet
05-27-2015, 03:15 AM
In my case, I think 23andme was pretty accurate. 99.7% European including 91.1% Balkan. That fits pretty well with what I know about my ancestry.

Sikeliot
05-27-2015, 03:28 AM
23andme works better for people of one full ancestry, IMO. Not for people who are mixed with many things, in small amounts. I notice the most accurate results in relatively unmixed people.

Tįltos
05-27-2015, 03:38 AM
23andme works better for people of one full ancestry, IMO. Not for people who are mixed with many things, in small amounts. I notice the most accurate results in relatively unmixed people.

Well I'm an American mutt, and 23andme IMHO was able to reasonably estimate my unknown Jewish ancestry. It is very far back from what I can gather. I finally have paper records into the 1700s for the line that it is coming from, and no obvious sign of it there (from paper record). OTOH, I have an Arbėreshė grandparent who I have inherited some West Asian from. FTDNA gives an inflated Asia Minor for this, which 23andme does not show.

Trojet
05-27-2015, 03:52 AM
Well I'm an American mutt, and 23andme IMHO was able to reasonably estimate my unknown Jewish ancestry. It is very far back from what I can gather. I finally have paper records into the 1700s for the line that it is coming from, and no obvious sign of it there (from paper record). OTOH, I have an Arbėreshė grandparent who I have inherited some West Asian from. FTDNA gives an inflated Asia Minor for this, which 23andme does not show.

Wow that is really cool that you have an Arbėresh grandparent.

I'm not an expert in autosomal calculations, but as far as I know, Albanians/Arbėresh don't score much on West Asian anyway. We are mainly Balkanic and Southern/SE European.

ArmandoR1b
05-27-2015, 04:31 AM
For me, 23andme is the best because it correctly distinguishes Iberian autosomal DNA from other southern European DNA. None of the other companies or Gedmatch calculators can do that. 23andme also gives a more accurate amount of Native American whereas the Gedmatch calculators have up to 2% noise in western Europeans. FTDNA myOrigins under reports Native American ancestry. Ancestry.com also provides a more accurate amount of Native American but it doesn't do as well for separating Iberian from other southern Europeans.

Hands down 23andme is the best for Latin Americans and Iberians even if it might give us slightly lower amount of North African, Middle Eastern, and West African components than the actual genealogical ancestry.

MonkeyDLuffy
05-27-2015, 04:55 AM
The 23&me gave a big chunk of SA ancestry, and knowing that SA includes bangladesh, nepal, pakistan, srilanka and India (23&me aslo include afghanistan), it didn't help much. So gedmatch calculators were better for understanding my ancestry.

Tįltos
05-27-2015, 04:58 AM
Wow that is really cool that you have an Arbėresh grandparent.

I'm not an expert in autosomal calculations, but as far as I know, Albanians/Arbėresh don't score much on West Asian anyway. We are mainly Balkanic and Southern/SE European. Thanks, from San Nicola dell'Alto, Italy.
At FTDNA my mom who would have been half scores I think off the top of my head 23% Asia Minor. I'm 12% there. Over at 23andme she shows much more Balkan and just a noisy 0.1% MENA. My small amount of MENA actually shows as mostly coming from my father in split view at 23andme, with only a 0.1% showing from my mom there.

I do think FTDNA made their admixture test slightly better from their original, but that's just my opinion. I really don't know if any standard in these autosomal tests can ever be imposed as autosomal DNA is randomly inherited from each line. I still prefer 23andme for this type of reading though. :)

Sikeliot
05-27-2015, 05:52 AM
For me, 23andme is the best because it correctly distinguishes Iberian autosomal DNA from other southern European DNA. None of the other companies or Gedmatch calculators can do that.

When Iberian is not a reference sample, Iberians find themselves scoring various combinations of Italian/Greek and British/Irish, as Iberians are genetically somewhere between the two.

Cinnamon orange
05-27-2015, 06:08 AM
Wow that is really cool that you have an Arbėresh grandparent.

I'm not an expert in autosomal calculations, but as far as I know, Albanians/Arbėresh don't score much on West Asian anyway. We are mainly Balkanic and Southern/SE European.

That is interesting. I am a cousin of Taltos on the Arberesh side. On Ancestry's test I score eight percent West Asian, four from the Middle East and four from the Caucuses.
My grandfather was Arberesh. I think it likely the Arberesh mixed some with locals both in Greece and southern Italy. That may be why we are higher west Asian.

Sikeliot
05-27-2015, 06:15 AM
Sicilian Arbereshe have mixed with other Sicilians. I have a woman on 23andme who is a Sicilian Arbereshe and she scores some Middle Eastern, though not as much as most Sicilians.

vettor
05-27-2015, 06:20 AM
I once thought 23andme was poor and gedmatch better , but I have found 4 matches to 23andme people in the past year ( the latest 9 hours ago, ..............I am awaiting the woman's reply after I gave her the data she requested - fingers crossed )

I have never found a match in gedmatch even with 2 different sets of uploads - using only my first 20 matches ( 2 different sets due to 23andme upload and ftdna upload )

Sikeliot
05-27-2015, 06:52 AM
23andme definitely has its advantages. Finding recent relatives, like cousins, is one of those advantages. I always get people on GEDmatch messaging me as matches though. I am one of the few I guess!

ArmandoR1b
05-27-2015, 07:20 AM
When Iberian is not a reference sample, Iberians find themselves scoring various combinations of Italian/Greek and British/Irish, as Iberians are genetically somewhere between the two.
Yes, I know that Iberians are genetically somewhere between the two. I have run the results of numerous people through the Gedmatch calculators and seen the results from 23andme, FTDNA, and Ancestry and I have also read all of the published studies that include Iberians.

That doesn't change anything in my initial post in this thread so it doesn't change the fact that 23andme is better for Iberians and Latin Americans than any of the Gedmatch calculators.

Trojet
05-27-2015, 11:26 AM
That is interesting. I am a cousin of Taltos on the Arberesh side. On Ancestry's test I score eight percent West Asian, four from the Middle East and four from the Caucuses.
My grandfather was Arberesh. I think it likely the Arberesh mixed some with locals both in Greece and southern Italy. That may be why we are higher west Asian.

I am not aware on how we score on other tests, but on 23andme all of my Albanian relatives including me, score pretty much 100% European with at least 90% Balkan, and no West Asian, or anything else outside of Europe.

ArmandoR1b
05-27-2015, 12:28 PM
I am not aware on how we score on other tests, but on 23andme all of my Albanian relatives including me, score pretty much 100% European with at least 90% Balkan, and no West Asian, or anything else outside of Europe.

Greeks also score a lot of Balkan. There is a pro and a con for this. The pro is that it distinguishes Greeks from Italians and Iberians which Gedmatch, FTDNA, and Ancestry can't do. The con is that it doesn't distinguish Albanians and other Balkan states or ethnicities from Greeks.

What are your Countries of Ancestry? https://www.23andme.com/you/labs/ancestry_finder/ That is another somewhat useful tool that Gedmatch, FTDNA, and Ancestry don't have. I understand that it is the ancestry of the matches and not necessarily where a common ancestor is from but the default is 4 grandparents from the same country so the higher percentages at the top of he list should be countries your ancestry is from but not always. I am curious as to what your case is.

Reith
05-27-2015, 12:44 PM
Gedmatch Eurogenes is right on for my ancestor's locations..

Trojet
05-27-2015, 12:58 PM
Greeks also score a lot of Balkan. There is a pro and a con for this. The pro is that it distinguishes Greeks from Italians and Iberians which Gedmatch, FTDNA, and Ancestry can't do. The con is that it doesn't distinguish Albanians and other Balkan states or ethnicities from Greeks.

What are your Countries of Ancestry? https://www.23andme.com/you/labs/ancestry_finder/ That is another somewhat useful tool that Gedmatch, FTDNA, and Ancestry don't have. I understand that it is the ancestry of the matches and not necessarily where a common ancestor is from but the default is 4 grandparents from the same country so the higher percentages at the top of he list should be countries your ancestry is from but not always. I am curious as to what your case is.

Yes, you are right that Greeks as well score pretty high on "Balkan" on 23andme, it would also make sense historically.

I am actually a Gheg Albanian from Western Macedonia with no known migrations to Macedonia. I think we (Gheg Albanians) score the highest Balkan. Historically, we have been the most isolated, therefore less mixed group in the Balkans, so the high Balkan percentage would make sense.

My countries of ancestry on 23andme are in this order: Greece, Albania (close second), Romania, Bulgaria, Serbia. Surprisingly no Macedonia. Important to note that my matches in Greece all have Arvanite/Albanian ancestry and are from NW Greece close to the border with Albania.

Cinnamon orange
05-27-2015, 06:19 PM
I think the Arbereshe, (provided it is the norm) show higher Middle East from either mixing in Italy or Greece, as Greeks show decent levels of Middle Eastern. But yes, it seems we veer off from modern day Albanians. Some Arbereshe in Italy came by way of Greece (the Peloponnesus and Attica to my understanding). We have some DNA matches in Greece.
We also show some North African, and that (i think) must be from southern Italy not Greece.
Odd because the group of towns my grandfathers family came from kept their language and identity and I just assumed they were not mixed. As usually minority's melt into the majority and lose their identity like the Huguenots in England.
My family was from what is now Crotone province.

Tįltos
05-27-2015, 07:23 PM
That is interesting. I am a cousin of Taltos on the Arberesh side. On Ancestry's test I score eight percent West Asian, four from the Middle East and four from the Caucuses.
My grandfather was Arberesh. I think it likely the Arberesh mixed some with locals both in Greece and southern Italy. That may be why we are higher west Asian.

Hi! I'm sure if you tested at FTDNA they would give you an even higher amount of Asia Minor. :biggrin1:

Cinnamon orange
05-27-2015, 07:26 PM
Hi! I'm sure if you tested at FTDNA they would give you an even higher amount of Asia Minor. :biggrin1:

Hehe, I will pass for now:)

BalkanKiwi
05-27-2015, 09:57 PM
I score 8% Asia Minor on FTDNA but get no type of Asian or Middle Eastern on 23andMe. I assume the Asia Minor could be representing my Croatian as I get no Southern Europe on FTDNA.

Sikeliot
05-28-2015, 04:31 AM
23andme seems more accurate than AncestryDNA from what I can see. AncestryDNA is not good at distinguishing Iberian, and does not attempt to differentiate Italians and Greeks.

DMXX
05-28-2015, 08:30 AM
23andMe's allocations in the Middle-Eastern and North African components are wholly inconsistent with the actual geographical distinctions in the region.

Middle-Eastern blurb:



The Middle East, represented here by people of Turkey, Iran and Syria, is an important crossroads in human history. Genetically the people of this region reflect that role, with connections to neighbors in every cultural, linguistic, and geographic direction.


North African breakdown:




Population Source Sample Size
Palestinian HGDP 51
Bedouin HGDP 48
Mozabite HGDP 30
Egypt 23andMe 28
Palestine 23andMe 28
Morocco 23andMe 19
Algeria 23andMe 14
Saudi Arabia 23andMe 8
Tunisia 23andMe 7
Jordan 23andMe 5
Yemen 23andMe 5
Kuwait 23andMe 3
United Arab Emirates 23andMe 2
Bahrain 23andMe 1


That alone renders AC component breakdown quite uninformative/misleading for those with West or Southwest Asian ancestry.

Dauv
05-28-2015, 08:37 AM
Some Arbereshe in Italy came by way of Greece (the Peloponnesus and Attica to my understanding). We have some DNA matches in Greece.

Not some but most Arbereshe in Italy came from Greece and were Arvanites. Yes, from Peloponnese and Central Greece but not only. Some were already mixed with Greeks. Admixture with southern Italians is more recent.

Source: I have an Arberesh ancestor from Piana.

Tįltos
05-28-2015, 11:38 AM
Not some but most Arbereshe in Italy came from Greece and were Arvanites. Yes, from Peloponnese and Central Greece but not only. Some were already mixed with Greeks. Admixture with southern Italians is more recent.

Source: I have an Arberesh ancestor from Piana.

I would love to learn more about where exactly my ancestors came from before they arrived in Italy. I'm not so sure that will be possible. I do have quite a few matches to Greeks. Some will say we match because the Albanians had such large families. :)

DMXX,
Great point!

Trojet
05-28-2015, 02:07 PM
I would love to learn more about where exactly my ancestors came from before they arrived in Italy. I'm not so sure that will be possible. I do have quite a few matches to Greeks. Some will say we match because the Albanians had such large families. :)

DMXX,
Great point!

You guys have probably read this, but it is is a very informative article from Wikipedia: http://en.m.wikipedia.org/wiki/Arbėreshė_people

It seems like there have been different waves of migrations mostly after the death of Albania's hero Skanderbeg who fought the Ottomans, but mainly coming from today's southern Albania, NW Greece, and later Central Greece. I'm sure there were migrations from northern Albanian lands as well just not as much.

Very interesting to point that all Albanians around that time were called amongst themselves "Arbėreshė", and then later I guess sometime after Ottoman occupation, we started calling ourselves as "Shqiptar" meaning people who speak/pronounce alike (same language).
And to this day, Arbėresh communities in Italy continue to use the old name Arbėreshė.

Dauv
05-28-2015, 02:37 PM
You guys have probably read this, but it is is a very informative article from Wikipedia: http://en.m.wikipedia.org/wiki/Arbėreshė_people

It seems like there have been different waves of migrations mostly after the death of Albania's hero Skanderbeg who fought the Ottomans, but mainly coming from today's southern Albania, NW Greece, and later Central Greece. I'm sure there were migrations from northern Albanian lands as well just not as much.

Very interesting to point that all Albanians around that time were called amongst themselves "Arbėreshė", and then later I guess sometime after Ottoman occupation, we started calling ourselves as "Shqiptar" meaning people who speak/pronounce alike (same language).
And to this day, Arbėresh communities in Italy continue to use the old name Arbėreshė.

That wikipedia article is based on poor sources, mostly internet websites. Arbėreshė is not so true that is an old name in south Italy. Arbėreshė in Sicily and Calabria called themselves "Greci" ("Greeks") until a few generations ago. I will be back on this (to be clear I'm not saying that Arbėreshė=Greeks).


I would love to learn more about where exactly my ancestors came from before they arrived in Italy. I'm not so sure that will be possible. I do have quite a few matches to Greeks. Some will say we match because the Albanians had such large families. :)

My uncle has rebuilt our family tree (paternal line) until the end of 1400 AD. We have most of the documents: birth certificates, marriage, census. It took over 15 years. Now we know that our paternal line descends from one of the Gallo-Italic towns of east Sicily (Oppida Lombardorum). When my ancestor at the end of 1700 left our hometown to settle in Palermo he married an Arberesh woman from Piana. From this marriage descends our branch. At the beginning of 1900 AD his descendant married a woman from Genoa, north Italy and their son moved to north Italy where my family has lived in the last century. In the last year I started doing some research about my Arberesh ancestors.

Trojet
05-28-2015, 02:53 PM
That wikipedia article is based on poor sources, mostly internet websites. Arbėreshė is not so true that is an old name. Arbėreshė in Italy called themselves "Greci" ("Greeks") until a few generations ago. I will be back on this (to be clear I'm not saying that Arbėreshė=Greeks).

Yes, I know how wikipedia works, but this article looks pretty accurate, considering what I know about Albanian history.

As far as the name goes, I'm sure some Arbėreshė who came from Greek areas (Arvanites) called themselves Greci, but the vast majority called themselves as Arbėreshė, just like their people back home :)
And yes that is their old/original name, one of the reasons why we all know and call the as such (Arbėreshė), otherwise we would call / know them as "Greci".

Cinnamon orange
05-28-2015, 03:34 PM
Yes, I know how wikipedia works, but this article looks pretty accurate, considering what I know about Albanian history.

As far as the name goes, I'm sure some Arbėreshė who came from Greek areas (Arvanites) called themselves Greci, but the vast majority called themselves as Arbėreshė, just like their people back home :)
There is a reason why we all know and call the as such, otherwise we would call / know them as "Greci".

Yes, I have read that Arvanite is what the Greeks called Albanians and that they used among themselves a word very similar to Arbereshe. It may be the same actually, just how I saw it spelled was different. The Albanian lands used to be called Arberia, so it makes sense.

Dauv, do you have any sources in English for the Albanian migrations? I think it seems that most Albanians moved to the Greek lands as mercenaries, for the Venetians or whoever was in control. Many areas became depopulated during the Ottoman period such as Attica, or so I have read. That may have made the Albanians the majority and locals merged with them.

My family's group of villages in what is now Crotone province southern Italy was settled by the first soldiers of Skanderbeg to come to Italy. I have no indication they came via what is now Greece, more likely the area where Scanderbeg was from but likely later refugees joined them. Later migrants may be in part how they were able to maintain their identity so long.

I have heard of an uprising in the Peloponnesus that brought many people to Italy and as far as I can tell, they integrated into existing Arbereshe communities.

My aunts explanation when I asked around age fourteen of how our family ended up in Italy was 'The Turks forced us out'.

As for Greci, it may be what the Italians called the Albanians, as I have read about old 'Greek' communities in southern Italy and think they may be Arbereshe. Maybe the ones who kept the Greek rite were more recent than those who changed to the Roman rite.

We have been in Italy a long time! I would love to know about later waves of migration as the initial settlement of our community is clear just not the exact origin of the soldiers. In fact they could have been from various areas. I wonder to how many took local wives. That could explain our MENA on 23andme. Most is 23andme's Middle East designation but Some is North African, which fits the pattern in southern Italy for 23andme results.

Trojet
05-28-2015, 03:50 PM
Yes, I have read that Arvanite is what the Greeks called Albanians and that they used among themselves a word very similar to Arbereshe. It may be the same actually, just how I saw it spelled was different. The Albanian lands used to be called Arberia, so it makes sense.

Yes, absolutely. Arvanites is what Greeks (greek pronounciation of Arbėreshė) called Arbėreshė (today's Albanians).

As I pointed out Albanians were called Arbėreshė until Ottoman occupation. Skanderbeg called Albania "The land of Arbėria". There was also some principalities in modern day Albania in the middle ages called the Principality of Arbėr.
Also, to this day a lot of Albanians name their boys "Arbėr", I have a cousin by that name BTW...

Mamluk
05-28-2015, 09:30 PM
Just a side comment--

My family called our Albanian ancestors, Arna'ut (Turkish: Arnavut). When I mentioned this to some friends from Kosovo, they snickered, because they said it is actually a derogatory label that the Turks gave to the Albanians. They said Arna'ut means "stubborn," supposedly because the Albanians were difficult to conquer, and even afterwards were challenging to control. The Serbs still use this name for Albanians, I believe.

(I had a great-great grandmother who was Albanian and supposedly never really learned Arabic after her father [Ömer Ağa] brought her to Palestine. She married my great-great grandfather (Y haplogroup is: J2b1), who was a descendant of an Albanian/Balkan janissary yeniēeri.)

basmaci
05-28-2015, 10:07 PM
They said Arna'ut means "stubborn,"
no in Turkish the word Arnavut has not a meaning outside Alban :) most likely derived from Arvanite. "Arnavut inadı" is just a idiom/phrase, means "stubbornness of Alban" or "stubborn like Alban". and is also not a offensive word in Turkish, either has to do with wars etc. there is also similar others : stubbornness of Circassian, stubbornness of Tatar...

Trojet
05-28-2015, 10:19 PM
Just a side comment--

My family called our Albanian ancestors, Arna'ut (Turkish: Arnavut). When I mentioned this to some friends from Kosovo, they snickered, because they said it is actually a derogatory label that the Turks gave to the Albanians. They said Arna'ut means "stubborn," supposedly because the Albanians were difficult to conquer, and even afterwards were challenging to control. The Serbs still use this name for Albanians, I believe.

(I had a great-great grandmother who was Albanian and supposedly never really learned Arabic after her father [Ömer Ağa] brought her to Palestine. She married my great-great grandfather (Y haplogroup is: J2b1), who was a descendant of an Albanian/Balkan janissary yeniēeri.)

It is great to see that there is a lot of people around the world with Albanian ancestry :)
Yes it is true that Turks call Albanians "Arnavut", although I think the original meaning is a pronunciation of the Greek "Arvanite".

As a side note, it is unfortunate that a lot of Arvanites/Albanians were deported against their will to Turkey by Greece during the population exchange between Greece and Turkey in the early 20th century. The Greek government over the ages has implemented a lot of policies (one of tthem repressing the use of their language Arvanitika) which resulted in assimilation of the remaining Arvanite population and unfortunately a lot of them are now assimilated, even though they played an important role during the Greek War of Independence.

http://en.m.wikipedia.org/wiki/Arvanites

Sikeliot
05-29-2015, 02:27 AM
We have been in Italy a long time! I would love to know about later waves of migration as the initial settlement of our community is clear just not the exact origin of the soldiers. In fact they could have been from various areas. I wonder to how many took local wives. That could explain our MENA on 23andme. Most is 23andme's Middle East designation but Some is North African, which fits the pattern in southern Italy for 23andme results.


Southern Italians also have genuine West Asian influences, not just North African.

Cinnamon orange
05-29-2015, 07:12 AM
Southern Italians also have genuine West Asian influences, not just North African.

I know but that is common I think in Greeks as well. I am curious about what admixture my Arbereshe family might have, such as Italian and/or Greek. (I think North African or what 23andme terms North African, is less common in Greeks but both southern Italians and Greeks get 23andm's Middle East designation in varying amounts.)

Mamluk
05-29-2015, 02:33 PM
The metropolitan centers of the Middle East, Anatolia and the Mediterranean coast (especially Palestine, Syria) have been melting pots for centuries. If samples from these regions are just being erroneously lumped into a general MENA classification, it will skew autosomal results for everyone else.

Unless these samples all come from Negev bedouins, or endogamous villages (even some of the Christians might have Greek and Italian input), then I cannot fully trust these calculators.

I believe the East Asian and north Eurasian component to be underestimated in Iran and Anatolia also.

Tolan
05-30-2015, 08:30 AM
23andme works better for people of one full ancestry, IMO. Not for people who are mixed with many things, in small amounts. I notice the most accurate results in relatively unmixed people.

This is not true for me!
All my ancestors of these last 500 years come all of a part of the Northwest of France.
23andme gives me:
34% British & Irish: False
Only 16.4% French & German
6.7% Iberian: false
1.4% Italian: false
0.6 Eastern European: false

I think, in fact, that there are no genes specifically French, British, German and therefore their methods is not good (as others companies).

Calculators in Gedmatchs, compare admixtures with the average found for each country: the method is more accurate in my case (in a single population, with two population, obviously the result is not true)

Cinnamon orange
05-30-2015, 08:44 AM
The metropolitan centers of the Middle East, Anatolia and the Mediterranean coast (especially Palestine, Syria) have been melting pots for centuries. If samples from these regions are just being erroneously lumped into a general MENA classification, it will skew autosomal results for everyone else.

Unless these samples all come from Negev bedouins, or endogamous villages (even some of the Christians might have Greek and Italian input), then I cannot fully trust these calculators.

I believe the East Asian and north Eurasian component to be underestimated in Iran and Anatolia also.

Maybe but I saw a study not long ago that mentioned the crusader impact on Lebanon through a Y DNA study. Also with the slave trade from Africa through the centuries there has been continued admixture. Not to mention Europeans sold both in North Africa and via Constantinople.... I doubt we can recreate the past. Even bedouins have more recent African admixture. And well the past was pretty admixed as well!

23andme at least is implying 500 years. I doubt there are any areas of the Middle East untouched by outside admixture. Even in more closed groups like Druze or Samaritans.

I do not fully understand the logic of 23andme's groupings. Palestinians in North Africa and Maltese in the Balkans. I do think an Anatolian category would be useful.

Cinnamon orange
05-30-2015, 08:47 AM
This is not true for me!
All my ancestors of these last 500 years come all of a part of the Northwest of France.
23andme gives me:
34% British & Irish: False
Only 16.4% French & German
6.7% Iberian: false
1.4% Italian: false
0.6 Eastern European: false

I think, in fact, that there are no genes specifically French, British, German and therefore their methods is not good (as others companies).

Calculators in Gedmatchs, compare admixtures with the average found for each country: the method is more accurate in my case (in a single population, with two population, obviously the result is not true)

Can you trace all of your GG etc grandparents? There were Cossacks in France at one time, which could give you the east euro. Especially, if the descendants then merged into a smaller community where the DNA passed back and forth. It may not have been via a marriage, so the ancestor may not be listed.

vettor
05-30-2015, 05:57 PM
I find 23nadme very helpful if you use - Countries of Ancestry with these selections
- In advanced controls, select 2+ grandparents and then move the segment slider to about 7cm ( use the USA selector option if you like also)

With this method I have found many links in family trees

Ron from PA
06-07-2015, 11:40 PM
For my known paper trail 23 is very accurate. Where ancestry is way off.

Mellow
06-12-2015, 03:13 AM
It seems accurate for me.I didn't know I had distant "oceanian" ancestry(although minor 0.1% and could be noise).
"Balkan" category doesn't mean much to me but I've noticed each Balkan ethnicity has their own ratio of categories.Greeks seem to differ from Albanians by scoring high "Italian" and Slavic Balkanites score some East European...

I don't know whether to trust it since a friends results changed (I think it was in 2013); some Ashkenazi Jewish became "Italian" .Now I thought 23andme had a reliable reference for AJ for something like that to happen.

AJL
06-12-2015, 02:59 PM
I don't know whether to trust it since a friends results changed (I think it was in 2013); some Ashkenazi Jewish became "Italian" .Now I thought 23andme had a reliable reference for AJ for something like that to happen.

23andme has a large number of AJ customers but this doesn't mean they all have a great paper trail. So, less well represented populations in the database that have contributed to, or which are similar to, the AJ gene pool could be confused for AJ.

Darko
06-12-2015, 11:29 PM
The problem with 23andme is that it is available only for the residents of some countries.

Joe B
06-15-2015, 05:20 PM
GEDmatch is down while we deal with a major data loss caused by a recently introduced software bug. We are working to restore data. We apologize for the inconvenience, and ask for your understanding.
That's what the GEDmatch front page says today.

Joe B
06-15-2015, 10:23 PM
GEDmatch is back up.

2015 June 15 - GEDmatch has experienced a significant loss of data due to a bug in a recent programming change. The loss affects the kit profiles that contain the person's name, and connect it to your login email account. Unfortunately, some information was not recovered. Affected are kits uploaded after approximately Feb 20, 2015....

Kaido
06-15-2015, 10:27 PM
Not sure what they did, but I have a bunch of new top matches.

Tįltos
06-16-2015, 04:46 AM
Not sure what they did, but I have a bunch of new top matches.

On ISOGG Facebook it was mentioned that some private kits were suddenly public since they went back online. This might be why you have a bunch of new matches? It was also mentioned that someone had lost all of their aliases and had to reenter them. I only noticed that for one of my kits I had to reenter the haplogroups.

Cinnamon orange
06-28-2015, 05:25 PM
23andme has a large number of AJ customers but this doesn't mean they all have a great paper trail. So, less well represented populations in the database that have contributed to, or which are similar to, the AJ gene pool could be confused for AJ.

I am wondering if some eastern euro is confused for AJ, as there is a small amount of eastern euro in AJ's. Most of my east euro matches show AJ. Seems unlikely so many have it and more likely in an inbred population like the AJ, that the eastern euro was passed about and became part of the component.
I have a small amount of AJ from my moms western euro side. I think that is actual ancestry. I have a match with two people, one in Spain and one in Mexico in the same segment. So I am wondering maybe a Sephardic ancestor with some in common AJ ancestry far back. Of course it could just be an Iberian ancestor, non Sephardic, who left DNA in Germany or the Spanish Netherlands way back when....

AJL
07-10-2015, 06:41 PM
I am wondering if some eastern euro is confused for AJ, as there is a small amount of eastern euro in AJ's. Most of my east euro matches show AJ. Seems unlikely so many have it and more likely in an inbred population like the AJ, that the eastern euro was passed about and became part of the component.


That would probably also best explain how a few minor yDNA lineages found their way into the Ashkenazi gene pool.

Reith
07-10-2015, 07:12 PM
Gematch seems much more accurate for me..

paulgill
07-10-2015, 08:28 PM
No, 23andMe's component composition calculator is nothing compared to the calculators available at gedmatch.com. 23andMe is an obsolete test, so are many more like the most of them at FTDNA.

Whole Genome Sequencing WGS test at Full Gemones is the solution, one test covers all and is complete, not just a few bits and pieces that you find at 23andMe and FTDNA. https://www.fullgenomes.com/

ArmandoR1b
07-10-2015, 10:46 PM
No, 23andMe's component composition calculator is nothing compared to the calculators available at gedmatch.com. 23andMe is an obsolete test, so are many more like the most of them at FTDNA.

Yet if a person doesn't know if they have large amounts of Iberian or Italian ancestry 23andme is the only test that correctly and consistently shows it. Gedmatch can't do that and it has false amounts of Native American of up to 2% in people that are from Europe. Gedmatch has a long ways to go for those.

paulgill
07-10-2015, 11:27 PM
Yet if a person doesn't know if they have large amounts of Iberian or Italian ancestry 23andme is the only test that correctly and consistently shows it. Gedmatch can't do that and it has false amounts of Native American of up to 2% in people that are from Europe. Gedmatch has a long ways to go for those.

23andMe fails miserably in my case while Geno 2.00 and the FF at FTDNA Myorigins does a much better job, and gedmatch.com a superb one, tons of free calculators there.

AJL
07-11-2015, 01:03 AM
Paul:

I know you have some sort of vendetta against 23andme lately, but this isn't an entirely fair comparison.

Gedmatch is wonderfully well run and useful, and if you spend the time on it you'll be well ahead of the game, but it can't run without some genetic data, which 23andme provides. Myorigins has gotten better (for some people) than the original FTDNA ancestry tool, but I can still find some kits I manage where 23andme has better analysis -- though these are platform 2 and 3 tests, not ones done on the new chip with fewer SNPs. And Geno 2.0 gives very "flat" results for most Europeans and West Asians, and seems to use a bizarre calculator for genetic distances to populations for them.

The bottom line is no single test will work perfectly for everyone, and the sheer size of the 23andme customer database in comparison with all others is something that it does and always will have going for it. I mean, it's so widespread a product that one of my known third cousins who lives in the same province as me even happened to buy it without knowing I had bought it years before. I wouldn't call 23andme the only tool in the chest but you can't say that any other genealogically useful DNA test has that kind of market saturation (over 1 million kits), and that in itself is a kind of usefulness.

ArmandoR1b
07-11-2015, 02:19 AM
23andMe fails miserably in my case while Geno 2.00 and the FF at FTDNA Myorigins does a much better job, and gedmatch.com a superb one, tons of free calculators there.

Failing miserably for some populations doesn't make 23andme obsolete, especially when it is the ONLY calculator to correctly distinguish Iberians from other European groups including Italians and Greeks and Gedmatch fails miserably at it.

paulgill
07-11-2015, 03:42 AM
Failing miserably for some populations doesn't make 23andme obsolete, especially when it is the ONLY calculator to correctly distinguish Iberians from other European groups including Italians and Greeks and Gedmatch fails miserably at it.


Good for a few and bad for most, agreed.

Boudicca
07-11-2015, 08:08 PM
My Dad gets 1.1% Eastern European on AC on Chr 3 and Chr 6. In what appears to be the same spots on the chromosome paintings at Gedmatch it appears as either West Asian or East Med. He is mostly British with brick walls in his tree. Frustratingly he gets no matches to any regions in these locations on Countries of Ancestry to compare but he does get 0.9% AJ 4GPs. I'm not sure which programmes are more likely to be real for these populations? 23andme or Gedmatch? I have been told 23andme is very accurate at picking up Eastern European?

paulgill
07-11-2015, 08:37 PM
My Dad gets 1.1% Eastern European on AC on Chr 3 and Chr 6. In what appears to be the same spots on the chromosome paintings at Gedmatch it appears as either West Asian or East Med. He is mostly British with brick walls in his tree. Frustratingly he gets no matches to any regions in these locations on Countries of Ancestry to compare but he does get 0.9% AJ 4GPs. I'm not sure which programmes are more likely to be real for these populations? 23andme or Gedmatch? I have been told 23andme is very accurate at picking up Eastern European?

No, it is not. All my people have a certain amount of East European component that differentiates them from other groups around them, and it is lousy at picking that up compared to many more free tools out there.

Boudicca
07-11-2015, 09:24 PM
No, it is not. All my people have a certain amount of East European component that differentiates them from other groups around them, and it is lousy at picking that up compared to many more free tools out there.

That's interesting. Which populations are you comparing against? What do you mean "certain amount of EE that differentiates from other groups?"
I share with 28 Brits and not one gets above 0.1% Eastern European. Not the biggest sample admittedly but even so Dad's 1.1% seems big compared to them! In terms how how well it differentiates between neighbouring groups eg Balkans I'm not so sure as I have seen a 7/8 British Isles 1/8 Croatian get 2% Eastern European and no Balkan or Southern European and it does seem a little conservative...I have seen a 7/8 Brit 1/8 Lithuanian get 5% Eastern European.

ADW_1981
07-11-2015, 09:47 PM
No, it is not. All my people have a certain amount of East European component that differentiates them from other groups around them, and it is lousy at picking that up compared to many more free tools out there.

No, I would say she is correct. It is very unusual for a British person to get East European at 23andME AC. It's more common with someone with recent German/Central European ancestry.

paulgill
07-12-2015, 04:41 AM
That's interesting. Which populations are you comparing against? What do you mean "certain amount of EE that differentiates from other groups?"
I share with 28 Brits and not one gets above 0.1% Eastern European. Not the biggest sample admittedly but even so Dad's 1.1% seems big compared to them! In terms how how well it differentiates between neighbouring groups eg Balkans I'm not so sure as I have seen a 7/8 British Isles 1/8 Croatian get 2% Eastern European and no Balkan or Southern European and it does seem a little conservative...I have seen a 7/8 Brit 1/8 Lithuanian get 5% Eastern European.

Do what pleases you. I have now left them all for WGS for good.

paulgill
07-12-2015, 04:44 AM
No, I would say she is correct. It is very unusual for a British person to get East European at 23andME AC. It's more common with someone with recent German/Central European ancestry. That is because it fails to differentiate one component from the other that makes up the British aDNA soup.

MonkeyDLuffy
07-12-2015, 05:33 AM
No, it is not. All my people have a certain amount of East European component that differentiates them from other groups around them, and it is lousy at picking that up compared to many more free tools out there.

What kind of east euro is exclusive to jatts?

Helgenes50
07-12-2015, 05:51 AM
@ ADW_1981

No, I would say she is correct. It is very unusual for a British person to get East European at 23andME AC. It's more common with someone with recent German/Central European ancestry.


The East European at 23andme is probably more ancient than 500 years.

I notice that some of their french customers are in this case, without sharing this recent ancestry.

My 4 grandparents are French ( Normand), I get in the last AC 0.9% of East European, but one of my ggg grandfather is supposed to be Russian.
I share with others of their custumers with a recent Eastern European Ancestry, without or just with a bit of this component.
So difficult to say whether it's recent or not
The problem, we are all a mix of Western, Eastern, Southern and Northern, in different proportions of course, often of ancient ancestry, and recent for some of us

Boudicca
07-12-2015, 07:59 AM
@Helgenes50: Thank you for your response. I suppose the difficult thing is how can anyone guarantee 100% of any ancestry? I'm a prime example of that as I haven't been able to trace my full family tree back to even 1800!! I could be completely wrong but I can imagine migration being easier on the continent for Eastern European migrants to move around and settle compared to Britain. You say that you get 0.9% and others with known Eastern European ancestry either get some or none, that suggests to me it is very conservative?

Boudicca
07-12-2015, 08:04 AM
That is because it fails to differentiate one component from the other that makes up the British aDNA soup.

I guess I need to share with more Brits just to be sure...I could be wrong but I didn't think many Slavs made up the genetic genepool of Britain either that or school taught me wrong!!! Yes the Angles, Jutes and Saxons came over but surely their mix together with the Celtic and Scandinavian would ultimately create the British/Irish category or fall into French/German and Scandinavian (as I have observed with my 28 British sharers!)?

paulgill
07-12-2015, 08:23 AM
What kind of east euro is exclusive to jatts?

We are talking about proportions here, which set Jatts, Panya, Tarkhan and other apart from each other .

anglesqueville
07-12-2015, 10:28 AM
Back to the original question, I seem to recall that a user of 23&me went to a professional geneticist, in a colloquium, and questionned him about the reliability of the Ancestral Composition of 23. Roughly the answer was: okay in standard mode, and only for the regional resolution. Of course this is very frustrating, because we all hope to get very more than an only result like

99.5% European
69.6% Northern European
4.5% Southern European
25.4% Broadly European
0.1% East Asian & Native American
0.1% Broadly East Asian & Native American
0.4% Unassigned

My personal philosophy is now to run all programs I find, compare the results, without focusing on anyone. Of course I have my special preferences, because likely of more or less conscious fantasies. For example I like the calculators which give me siberian or saami or amerindian admixture, probably because I love the northern countries, and dream of toundras and long winters. But nothing in all this is science. It's not wrong to dream, but it's dangerous to confuse dreams and science. The fact is that in these topics, the borderline between them is thin and moving.

Boudicca
07-12-2015, 11:30 AM
In my case it highlights something I wasn't expecting and has caused me a headache!!! Some people say anything over 1% on AC is real, others say not. Some say 0.1% is noise, others not. Some say it's recent, others ancient. No one agrees with what any results mean so how can you use tests like this to help back up traditional genealogy which is what I'm more interested in compared to what happened thousands of years ago? In this respect given the original question I don't think I trust any!

BalkanKiwi
07-12-2015, 11:35 AM
At the end of the day the only hard evidence we have are paper trails and bringing different information sources together, and not simply relying on one. I'd like to believe, for example, the 2% North African I get on the K23b is real, however I can't prove it and likely never will. Frustrating? Yes, however delving into genetics is only going to give you questions you wont find answers too. Stick to the stuff you know is accurate and take the smaller minorities from different sources with a grain of salt.

ArmandoR1b
07-12-2015, 12:07 PM
Good for a few and bad for most, agreed.

Talk about an over simplification. Seeing as how there are more than 754 million people of Iberian, Italian, or Native American ancestry and 23andme helps easily distinguish these people from the rest of the European and North American population of almost 1 billion people it is good for a significant number of people. Gedmatch admixture calculators can't do the above as simply or accurately as 23andme.

Boudicca
07-12-2015, 12:48 PM
At the end of the day the only hard evidence we have are paper trails and bringing different information sources together, and not simply relying on one. I'd like to believe, for example, the 2% North African I get on the K23b is real, however I can't prove it and likely never will.

But that's the thing...what scores ARE significant?! Is it 1% as standard across ALL calculators? Probably not but perhaps some kind of guide from them all would help us decide?

Helgenes50
07-12-2015, 01:43 PM
But that's the thing...what scores ARE significant?! Is it 1% as standard across ALL calculators? Probably not but perhaps some kind of guide from them all would help us decide?

All scores are probably significant, even very low, when they are repeated in each calculator

evon
07-12-2015, 02:52 PM
I dont think any one calculator is 100% correct at this point (that includes 23andme and FTDNA), but those components that are found in most of them should be seen as real, while that which is visible in only one or less should be taken with a grain of salt...

Tįltos
07-12-2015, 02:54 PM
My Dad gets 1.1% Eastern European on AC on Chr 3 and Chr 6. In what appears to be the same spots on the chromosome paintings at Gedmatch it appears as either West Asian or East Med. He is mostly British with brick walls in his tree. Frustratingly he gets no matches to any regions in these locations on Countries of Ancestry to compare but he does get 0.9% AJ 4GPs. I'm not sure which programmes are more likely to be real for these populations? 23andme or Gedmatch? I have been told 23andme is very accurate at picking up Eastern European?

Hi Boudicca,
In the Countries of Ancestry have you adjusted your settings down to just say 3,2,1 GP level instead of 4 and down to 5 cm? This might offer more information, though I know I like to have more certainty with the 4 GP. Keep in mind though 4GP might not be accurate either. I have 4GP for me from USA, does not mean I'm Native American? No of course not. :) My mom's kit has 2GP from Italy. Though they are really Albanians who have been in Italy since the late 1400s.

I know lgmayka gave you this link in your other thread. http://untoldlondon.org.uk/articles/read/polish_migration_to_london_a_drama_in_three_acts
Consider this as well, as I think I read your paper trail only goes back to 18th century? Your father may have connections to Eastern Europe this way. http://polishscottishheritage.co.uk/?heritage_item=scots-in-poland http://www.krakowpost.com/2014/03/scots-in-poland-poles-in-scotland/

I think 23andme is very good at finding Eastern European ancestry, though they can't say Lithuanian vs Russian. I'll take it! The original population finder at FTDNA could not tell me I was Eastern European though I'm half Eastern European.
My 23andme AC result-41.9% Eastern Europe
My current FTDNA-46% Eastern Europe (big improvement!)

You had also mentioned some 4GP AJ you found for your father. From what I can see most of the fully AJ that I match at 23andme show very small percentages of Eastern European around 0.4%. Though I realize others that I don't share with *might* show a little more. What I'm trying to say is that Eastern European and AJ should not be getting confused at 23andme. They are very good a distinguishing between the two.

EDIT-Also to let you know my AJ is only 0.9 in all three modes over at 23andme. Very small percent. Most will say that is noise. Given that I do have matches to other AJ on the chromosome that is painted in AJ, and what my brother's Y DNA turned out to be suffice it to say I have AJ ancestry. Though it is very far back. I certainly have not been able to find any records for this conversion, and not sure that I ever will! So some of your AJ 4GP I'm sure do have connections to this Eastern European component. Just that they won't mistake the AC between the two.

Boudicca
07-12-2015, 03:06 PM
Hi Taltos,
Thanks for the information. My Dad's paper trail is iffy in the 19th century!!! That's a pretty good score for you on AC!

Tįltos
07-12-2015, 03:12 PM
Hi Taltos,
Thanks for the information. My Dad's paper trail is iffy in the 19th century!!! That's a pretty good score for you on AC!

You're welcome! I hope you are able to start piecing together the puzzle of your father's results. :)

evon
07-12-2015, 05:29 PM
My results in both 23andme and FTDNA to show just how different they can be:

23andme:
http://i412.photobucket.com/albums/pp207/vulcanphoto/Screenshot%20from%202015-07-12%20192445.png

FTDNA:
http://i412.photobucket.com/albums/pp207/vulcanphoto/mew.png

Both calculations fail to pick up the Balkan and South Asian derived from my Romani ancestry, as well as German ancestry etc, which is visible in Segment sharing with these populations and documented papertrails..so they are far from perfect, but you clearly see a similar trend among them, which is likely correct...

AnnieD
07-12-2015, 06:49 PM
Since both of your results show high Scandinavian percentages, it makes me curious about the FTDNA results that my father and I received. Unfortunately, his results are loaded only at FTDNA which does not have transfer compatibility with 23andme at this time. However, I am tested at both companies plus several others and consistently get 0% to just a few percentages for Scandinavian AC. We are of primarily Colonial American - British descent, so I am certainly primed for finding more AC discrepancies than Europeans of less migratory heritage would. However, my father's 47% Scandinavian vs. 0% for me results seems to stretch the boundaries of ethnicity accuracy. Given the recent Nature Study in Britain which showed that Scandinavian genomes are more confined to just a few regions, it also seems likely that his results are very conflated. To those more astute in genomic analysis, is it considered within the current realms of scientific possibly for genetic inheritance to be so unequal and mish-mashed that the parent could have nearly half of genome classified as one ethnicity vs. the child as zero of that ethnicity? Or, does this seem to be calculator interpretation issue for populations with some shared ancestral components?

paulgill
07-12-2015, 08:08 PM
Talk about an over simplification. Seeing as how there are more than 754 million people of Iberian, Italian, or Native American ancestry and 23andme helps easily distinguish these people from the rest of the European and North American population of almost 1 billion people it is good for a significant number of people. Gedmatch admixture calculators can't do the above as simply or accurately as 23andme.

Our population is twice as big as that, and 23andMe is of no use for us and for many billions of the other ethnicities. Fine, if it works for you, but stop selling it to us as it is useless for us.

evon
07-12-2015, 08:09 PM
Since both of your results show high Scandinavian percentages, it makes me curious about the FTDNA results that my father and I received. Unfortunately, his results are loaded only at FTDNA which does not have transfer compatibility with 23andme at this time. However, I am tested at both companies plus several others and consistently get 0% to just a few percentages for Scandinavian AC. We are of primarily Colonial American - British descent, so I am certainly primed for finding more AC discrepancies than Europeans of less migratory heritage would. However, my father's 47% Scandinavian vs. 0% for me results seems to stretch the boundaries of ethnicity accuracy. Given the recent Nature Study in Britain which showed that Scandinavian genomes are more confined to just a few regions, it also seems likely that his results are very conflated. To those more astute in genomic analysis, is it considered within the current realms of scientific possibly for genetic inheritance to be so unequal and mish-mashed that the parent could have nearly half of genome classified as one ethnicity vs. the child as zero of that ethnicity? Or, does this seem to be calculator interpretation issue for populations with some shared ancestral components?

It is certainly possible that you inherited no Scandinavian DNA, but segment sharing should be able to tell you if this is the case. Do you have Scandinavian matches at 23andme or FTDNA?

evon
07-12-2015, 08:11 PM
Our population is twice as big as that, and 23andMe is of no use for us and for many billions of the other ethnicities. Fine, if it works for you, but stop selling it to us as it is useless for us.

23andme and FTDNA are both very Eurocentric, but I would not say they are useless for South Asians, that is going a bit too far, but they do not at this time give a correct picture for people from South Asia for sure..

paulgill
07-12-2015, 08:29 PM
Talk about an over simplification. Seeing as how there are more than 754 million people of Iberian, Italian, or Native American ancestry and 23andme helps easily distinguish these people from the rest of the European and North American population of almost 1 billion people it is good for a significant number of people. Gedmatch admixture calculators can't do the above as simply or accurately as 23andme.


23andme and FTDNA are both very Eurocentric, but I would not say they are useless for South Asians, that is going a bit too far, but they do not at this time give a correct picture for people from South Asia for sure..

We have another member here who I share my 23andMe results with. Genetically he is very close to me on gedmatch calculators, but on 23andMe he have three times as much Euro as I do, yet on the Global Similarity Map, I am shown being much closer to Euorpeans than him. Global Similarity Map paints a much truer picture than the Ancestry Composition at 23andMe, I have the proof for that.

nuadha
07-12-2015, 10:33 PM
It is certainly possible that you inherited no Scandinavian DNA, but segment sharing should be able to tell you if this is the case. Do you have Scandinavian matches at 23andme or FTDNA?

How would segment sharing confirm what is claimed in ancestry composition. What feature shows segment sharing?

Tįltos
07-12-2015, 11:36 PM
How would segment sharing confirm what is claimed in ancestry composition. What feature shows segment sharing?

By using either Countries of Ancestry or the Family Inheritance: Advanced. https://www.23andme.com/you/labs/multi_ibd/

evon
07-13-2015, 10:20 AM
How would segment sharing confirm what is claimed in ancestry composition. What feature shows segment sharing?

You have to triangulate the segment in order to get a proper picture of the segment in question (as Taltos explained), which can be done by sharing with multiple people and looking for actual matching between them, I have done it for many segments, but here is one i made a mindmap of so as to show how it is clearly linked to Jewish populations:
http://i412.photobucket.com/albums/pp207/vulcanphoto/New-Mind-Map_1w38vwl7.jpg

ArmandoR1b
07-13-2015, 04:22 PM
Our population is twice as big as that, and 23andMe is of no use for us and for many billions of the other ethnicities. Fine, if it works for you, but stop selling it to us as it is useless for us.

You must be forgetting that you were the person that stated that 23andme is now obsolete. It isn't obsolete and it is the best calculator for significant components. Accept that fact that it isn't obsolete and stop trying to claim it is useless for everyone. There are almost a billion that it isn't useless for.

paulgill
07-13-2015, 08:02 PM
You must be forgetting that you were the person that stated that 23andme is now obsolete. It isn't obsolete and it is the best calculator for significant components. Accept that fact that it isn't obsolete and stop trying to claim it is useless for everyone. There are almost a billion that it isn't useless for.

23andMe is an obsolete test as WGS from Full Genomes is available, and WGS renders most of the tests at FTDNA useless too, FTDNA SNP Panels, I believe, are still good for most people. As a matter of fact FTDNA MyOrigin is much better than 23andMe Ancestry composition for me. 23andMe may be OK for people like you, but why should I praise something that is totally useless for me?

Boudicca
07-13-2015, 08:43 PM
By using either Countries of Ancestry or the Family Inheritance: Advanced. https://www.23andme.com/you/labs/multi_ibd/

Which is more reliable, AC or Countries of Ancestry? Dad's getting some nice Eastern European overlapping segments and a good few matches to Poland but nothing out of the ordinary compared to 2 other Brits who also get 17 matches to Poland at 4GPs 5cM and he has no matches where the EE appears on AC!!

evon
07-13-2015, 10:10 PM
Which is more reliable, AC or Countries of Ancestry? Dad's getting some nice Eastern European overlapping segments and a good few matches to Poland but nothing out of the ordinary compared to 2 other Brits who also get 17 matches to Poland at 4GPs 5cM and he has no matches where the EE appears on AC!!

You cant look at only visible overlapping's for a true picture, you need to check it using the "Advanced Family inheritance tool" between the two or more parties, as contrary to what Countries of Ancestry show you, each chromosome is divided into two "bars", which is visible in Ancestry Composition. On another note, you cant use Ancestry Composition to check segments in Countries of Ancestry, as these tools are not compatible it seems, which is due to Ancestry Composition not being correct.

ArmandoR1b
07-16-2015, 06:01 PM
23andMe is an obsolete test as WGS from Full Genomes is available, and WGS renders most of the tests at FTDNA useless too, FTDNA SNP Panels, I believe, are still good for most people.
WGS is too expensive for the general public. Since most people can't afford it there is no way to find enough reports on the quality of it as an ethnicity calculator and I truly doubt that a large enough dataset exists for it to be accurate as far as an ethnicity calculator. Since this thread is about 23andme Ancestry Composition vs Gedmatch then it is obviously about ethnicity calculators and therefore the FTDNA SNP panels have nothing to do with the original topic.


As a matter of fact FTDNA MyOrigin is much better than 23andMe Ancestry composition for me.
That's fine. You still haven't understood that I am not interested in anecdotal experiences. I am interested in large groups of people and whether or not their results fit your statement that 23andme is obsolete. Since the 23andme results are very good for a large population then the blanket statement that "23andme is obsolete" is wrong.



23andMe may be OK for people like you, but why should I praise something that is totally useless for me?
That post shows what you have been failing to understand. I am not asking you to praise anything. I am asking you to not make blanket statements that are factually incorrect.

paulgill
07-16-2015, 08:04 PM
WGS is too expensive for the general public. Since most people can't afford it there is no way to find enough reports on the quality of it as an ethnicity calculator and I truly doubt that a large enough dataset exists for it to be accurate as far as an ethnicity calculator. Since this thread is about 23andme Ancestry Composition vs Gedmatch then it is obviously about ethnicity calculators and therefore the FTDNA SNP panels have nothing to do with the original topic.


That's fine. You still haven't understood that I am not interested in anecdotal experiences. I am interested in large groups of people and whether or not their results fit your statement that 23andme is obsolete. Since the 23andme results are very good for a large population then the blanket statement that "23andme is obsolete" is wrong.



That post shows what you have been failing to understand. I am not asking you to praise anything. I am asking you to not make blanket statements that are factually incorrect.

I understand you very well, but what I don't understand is why you be promoting one useless calculator over the tens of other better ones, because that equals misleading the masses. Also new better calculators are showing up everyday, looks like either you are ignoring them or are not aware of them?

AJL
07-17-2015, 05:41 PM
^ Just because calculators are new does not mean they are better.

As far as I can see the algorithms have been pretty much the same for a few years, and only the number of clusters and what they are called differ.

Sometimes different names for these clusters have been bandied about, and while the centroids of the clusters are anthropologically interesting, the names of the clusters are not especially useful for determining the origins of individuals. It is no less helpful to say "you have more of Component 3 than you should," if you know where Component 3 is centred, than to say "you have more of Mediterranean than you should."

Beyond a certain point, accuracy does not improve with increased clusters or runs, so in all likelihood, pretty much the only things that will improve calculators are:

(1) More data points (i.e. more SNPs tested by companies);
(2) Better/more/more accurate sample sets for populations; and
(3) Better understanding of linkage of SNPs to remove redundant data points.

Points 1 and 3 are potentially contradictory, however, because more data points can actually add more noise if the SNPs are not ancestrally informative. So until we have mass full genomic testing with well screened and highly regionalized population samples, I suspect we are a few years off from any real improvements to the calculators.

ArmandoR1b
07-19-2015, 02:17 PM
I understand you very well, but what I don't understand is why you be promoting one useless calculator over the tens of other better ones, because that equals misleading the masses. Also new better calculators are showing up everyday, looks like either you are ignoring them or are not aware of them?

I have been very clear about this. The other calculators are not able to accurately distinguish Iberians from Italians, other southern Europeans, and other Europeans. It also has a very low false positive of Native American. The strengths and weaknesses of each calculator need to be recognized. Just because a calculator is weak in one area does not mean that it is obsolete when it is very strong in other areas.

Show me a calculator that can do what 23andme does for western and southern Europeans, Latin Americans, and Native Americans and I will stop promoting 23andme for these people or for people without knowledge of their ancestry. Until then 23andme is not obsolete.

paulgill
07-19-2015, 06:37 PM
I have been very clear about this. The other calculators are not able to accurately distinguish Iberians from Italians, other southern Europeans, and other Europeans. It also has a very low false positive of Native American. The strengths and weaknesses of each calculator need to be recognized. Just because a calculator is weak in one area does not mean that it is obsolete when it is very strong in other areas.

Show me a calculator that can do what 23andme does for western and southern Europeans, Latin Americans, and Native Americans and I will stop promoting 23andme for these people or for people without knowledge of their ancestry. Until then 23andme is not obsolete.

Anyways, I have now said a big good bye to 23andMe, FTDNA and the like. The only ones that I find useful are Full Genomes and YSEQ for testing, also the Gedmatch and other free tools for ancestory composition. On this very site, right here, many people are creating and testing all kinds of calculators, one just have to participate in those tests and see which one works best for them.

Christina
07-19-2015, 07:33 PM
All calculators will suffer from what computer programmers somewhat harshly call GIGO: "Garbage in, garbage out." This expresses the principle that results can only be as good as the data on which they rest.

In other words, the baselines are largely calculated by self-reporting. If enough people define their ancestry wrong, the baselines will be calculated wrong.

I am active in the genealogy community, and I've helped countless people trace their ancestries. I find that just about every American family of northern European descent who I encounter has some kind of unsubstantiated family mythology. To wit, someone with the last name Cole, who thinks that they are British, have no idea that their original surname was Kohl, and that they were German. Others are the other way around: they infer "exciting" myths upon their ancestors, for flavor and color. I can't tell you how many people claim to have some Native American ancestry who don't.

Even in the old world, the data suffer from being able to distinguish between recent and ancient exotic ancestry. Take someone from a corner of South Italy. Let's assume that village was populated during ancient times (before any modern concept of "ethnos" emerged). The original population was substantially farmers from the east. But these people formed the Italian ethnic group, and have been there for 3000 years. The presence of some alleles similar to eastern populations may or likely may not indicate recent ancestry. Obviously if you go back far enough, almost every European passed through the Middle East. And way far back enough, everyone is African.

So all calculators could be clearer on timing. In other words, do they test in common alleles that indicate very little, except shared ancestry before our ethnic and regional concepts even emerged? Or do they test true divergences, whose presence genuinely indicates a modern introgression?

"Insignificant" rural areas suffer from under-representation in the basal calibration data as well. There is a paucity of data from many many regions in Europe; one imagines how poor the coverage must be in other regions. And thus the calculators are far from perfect.

ArmandoR1b
07-19-2015, 07:35 PM
Anyways, I have now said a big good bye to 23andMe, FTDNA and the like. The only ones that I find useful are Full Genomes and YSEQ for testing, also the Gedmatch and other free tools for ancestory composition. That's fine. Like I said many times already, I am just asking you to not label 23andme as obsolete when it works very well for so many people.

I also think that BigY, FGC, and Yseq should be used when needed and they when they are affordable but those don't have anything to do with ethnicity calculators so that is going off-topic.



On this very site, right here, many people are creating and testing all kinds of calculators, one just have to participate in those tests and see which one works best for them.
I don't have a need to spend time and energy participating in those tests and finding one that works for Iberians, Latin Americans, southern Europeans, and Native Americans when that test already exists and it provides other valuable tools including a database of distant relatives not found elsewhere.

kjjohnston
07-19-2015, 11:16 PM
5246

None of the admixture comparisons do a perfect job. Part of the reason is that segments are not identified through phasing AND mapped back to obvious ancestors. We need to follow phased segments back in time. This is where genealogists can help the population biologists.

I am showing the X chromosome because brothers who have the same segments generally should have the same admixture. They are phased naturally because they only have one X chromosome to compare. In addition, where you see distinct crossover borders between the grandparents, then you might expect more of a transition to a different population.

At 23andMe, it appears that there is a lot of fitting and smoothing going on by mixing the segments from different ancestors. I am showing how each sibling matches the two maternal grandparents.

At least 23andMe does attempt to show populations on the X chromosome which is more than we can say for the population researchers in general.

Zionas
08-04-2015, 12:13 PM
I take GEDmatch with a grain of salt, because many of my admixture results on there don't make sense to me. They give me lots of weird components like Atlantic-Baltic, West Asian etc. 23andMe shows I'm pretty much fully East Asian, although my 16% Unassigned in conservative mode is still a mystery. I'll get my parents tested. There is a Vietnamese guy who has shared genomes with me and he's 61.2% Unassigned in conservative mode! A Thai Chinese woman who's of Hakka descent (Hakkas are believed to be the most pure Han Chinese) shows 100% EA and 98.5% Chinese in speculative mode, while even I lag slightly behind at 99.7% EA and 97.1% Chinese.


If you would like to know my AC results, please view my 'Welcome' post or PM me.

Zionas
08-04-2015, 12:16 PM
My FTDNA Family Finder Results:
55% NEA
45% SEA

Thorbjorn
08-21-2015, 03:31 PM
Hi all...

This is my first visit here, stumbling on it from a Google search '23andMe GEDmatch discrepancy'. This is a bit long, but bear with me. I want to be as thorough as possible.

My Y-DNA haplogroup is T, Middle/Near Eastern. My family always believed we are fully of Italian and Sicilian descent (born American). However, our surname is of Norman French origin as far as I can tell, not Italian, and is cognate with two forms of a common British surname. Our surname is concentrated in the Palermo Province region, the seat of the Norman kingdom in southern Italy and Sicily. It is sparse on the Italian mainland, even in the south of the boot. My Countries of Ancestry file from 23andMe shows a good number of UK, Sweden, Norway, Ireland with the DNA matches' 4 grandparents from those countries, predominately UK. I have suspected from that circumstantial evidence that we have Norman ancestry.

My AC from 23andMe shows 70% Italian, but I think that's misleading and not very finely broken down.

70.3 Italian
10.8 Broadly Southern European
1.2 Broadly Northern European
3.2 Broadly European
12.9 Middle Eastern & North African

Eurogenes K15 shows (K13 is not much different):

North_Sea - 9.33
Atlantic - 16.77
Baltic - 2.81
Eastern_Euro - 3.71
West_Med - 18.09
West_Asian - 12.72
East_Med - 29.52
Red_Sea - 4.64
Northeast_African - 0.24
Sub-Saharan - 2.17

The West Asian, East and West Mediterranean, Red Sea, African are no doubt the T haplogroup. The European numbers are probably my mtDNA HV4a haplogroup.

What I want to find, and have not yet been able to do so, is a finer breakdown of the regions of North Sea, Atlantic, and Baltic. Those are the areas where the Germanic tribes held sway: Norse, Anglo-Saxon, Norman, Franks, Vandals, etc. Does anyone know how to get a finer breakdown, and why there's such a discrepancy between 23andMe and GEDmatch?

Humanist
08-21-2015, 04:26 PM
The West Asian, East and West Mediterranean, Red Sea, African are no doubt the T haplogroup. The European numbers are probably my mtDNA HV4a haplogroup.

First, welcome to the forum. Your haplogroups represent an infinitesimally small percentage of your overall ancestry. Unless those haplogroups were a recent introduction into your ancestral homeland(s) (e.g. Sicily), I would not associate percentages from your ancestry composition with your haplogroups.

Thorbjorn
08-21-2015, 04:52 PM
Thank Humanist. I should mention I understand that the haplogroups represent only a direct line of genetic descent, father's father's father... etc. on back. I probably misspoke in associating them with the admixtures. I'm thinking more along the lines of the autosomal matches in the discrepencies of percentages. That's what I'm presuming is being reported as the percentages (autosomal). In mentioning the Mediterranean, Red Sea, etc. I am thinking of the locations where T is found, and some of my direct paternal ancestors may have settled. I hope I explained that right, though I think I just confused it further... :P this is all really new to me. So I guess I should have kept it to a simple "why are 23andMe and GEDmatch so different?" (wow, that was a lot easier to say :)).

vettor
08-21-2015, 11:47 PM
Thank Humanist. I should mention I understand that the haplogroups represent only a direct line of genetic descent, father's father's father... etc. on back. I probably misspoke in associating them with the admixtures. I'm thinking more along the lines of the autosomal matches in the discrepencies of percentages. That's what I'm presuming is being reported as the percentages (autosomal). In mentioning the Mediterranean, Red Sea, etc. I am thinking of the locations where T is found, and some of my direct paternal ancestors may have settled. I hope I explained that right, though I think I just confused it further... :P this is all really new to me. So I guess I should have kept it to a simple "why are 23andMe and GEDmatch so different?" (wow, that was a lot easier to say :)).

Hi I am also T from Italy and have ancestors from Italy since the 17th century .
there has been T in central europe from at least 5500BC ( as per Haak findings , I suspect when they finally do the dna on the other 26 skeletons they will find more T )

You have far more italian than I in 23andme 70% to 30 % .

What "non-italian" surnames do you have so I can check the information I have

Thorbjorn
08-22-2015, 12:30 PM
Hi I am also T from Italy and have ancestors from Italy since the 17th century .
there has been T in central europe from at least 5500BC ( as per Haak findings , I suspect when they finally do the dna on the other 26 skeletons they will find more T )

You have far more italian than I in 23andme 70% to 30 % .

What "non-italian" surnames do you have so I can check the information I have

Hi vettor...

I haven't found any non-Italian surnames. I used Ancestry.com to try to trace my family tree. I don't know how accurate it is though I tried to be as thorough as possible. All of my ancestors I found were born in Italy and Sicily from the 1800s on... Pantano, Fazio, Scauzza, Tedesco, Dispenza, Scalera, Pascarella, Petrillo and several more that represent back to my 3x great-grandparents.

vettor
08-22-2015, 06:08 PM
Hi vettor...

I haven't found any non-Italian surnames. I used Ancestry.com to try to trace my family tree. I don't know how accurate it is though I tried to be as thorough as possible. All of my ancestors I found were born in Italy and Sicily from the 1800s on... Pantano, Fazio, Scauzza, Tedesco, Dispenza, Scalera, Pascarella, Petrillo and several more that represent back to my 3x great-grandparents.

well, it looks like they are all sicilian names except for

Petrillo which is in Campania

Tedesco which means German , is in all of Italy ...........usually non-noble germans heading for the levant for the crusades

and
Scauzza which is a sopranome ( nick name ) very recent, like last 150 to 200 years.......

Thorbjorn
08-22-2015, 06:24 PM
Well, I think I know what's going on here now. This is a slide from GEDmatch's tutorial. I take it this is the "typical" genetic make-up of the "typical" southern Italian (and Sicilian).

http://slides.com/kittycooper/gedmatch#/14

This is mine (not sure if it's visible to anyone but me, I could make it an attachment).

http://www4.gedmatch.com/gifs/M436538_A73375.gif

It looks like 23andMe aggregates some of the other components. If 1,000 people from Italy show the same admixture (+/-) I can understand the results. It looks like GEDmatch just breaks down the admixture to a finer degree. It looks like the "typical" southern Italian is indeed descended from peoples who migrated from northern Europe, hence the high number of fair-skinned, light-haired Italians and Sicilians.

Erlembaldo
09-05-2015, 10:46 PM
What I want to find, and have not yet been able to do so, is a finer breakdown of the regions of North Sea, Atlantic, and Baltic. Those are the areas where the Germanic tribes held sway: Norse, Anglo-Saxon, Norman, Franks, Vandals, etc. Does anyone know how to get a finer breakdown, and why there's such a discrepancy between 23andMe and GEDmatch?

I'm no expert, but it could probably be broken down as such:

Atlantic=NW Europe
North Sea=Scandinavia
Baltic=NE Europe

Thorbjorn
09-16-2015, 05:55 PM
I'm no expert, but it could probably be broken down as such:

Atlantic=NW Europe
North Sea=Scandinavia
Baltic=NE Europe

Thanks, That makes sense. I haven't been on the site for a while and didn't see this until now.

Dorkymon
09-16-2015, 09:26 PM
It looks like the "typical" southern Italian is indeed descended from peoples who migrated from northern Europe, hence the high number of fair-skinned, light-haired Italians and Sicilians.

This should be the most deluded deduction I had read this month.

jfttrh44
03-12-2016, 05:28 PM
Some people say that 23andme predicts ancestry from the last 500 years, and that it is therefore more reliable than GEDmatch.

I don't really agree because I've noticed very peculiar things happen with the 23andme calculator. I've seen families where one parent is Sicilian and the other is Irish. The Sicilian parent scores 15% Middle Eastern, while the child scores 0%. I've yet to see a half North European, half SE European (Cretan, Sicilian, Greek islander) score significant Middle Eastern even though full members of said groups score a significant amount.

Likewise, I suspect 23andme underestimates admixture when it is small. I have seen that in people with under 10% of African, the African is greatly underestimated on 23andme.

What do others think? I think due to its algorithms, 23andme ends up giving misleading results, but it is good for telling if a given ancestral component exists... just not necessarily the amounts.

The thing about the "500 years" is absolutely rubbish. I am from Spain, all my ancestors (I did my family tree until 1500) are fully Spaniards, and on 23andme and on gedmatch, 99% of my matches have non-spanish surnames. I practically have matches from many places in Europe except on Spain, and I am fully Spaniard for the last 500 years.

This means that the ancestry is older, the matches are older, and that very few Spaniards have tested on gedmatch. But my matches are much older than those 500 years, probably 1000-2000 years older.

ArmandoR1b
03-13-2016, 11:39 AM
The thing about the "500 years" is absolutely rubbish. I am from Spain, all my ancestors (I did my family tree until 1500) are fully Spaniards, and on 23andme and on gedmatch, 99% of my matches have non-spanish surnames. I practically have matches from many places in Europe except on Spain, and I am fully Spaniard for the last 500 years.

This means that the ancestry is older, the matches are older, and that very few Spaniards have tested on gedmatch. But my matches are much older than those 500 years, probably 1000-2000 years older.

I agree that the autosomal DNA in Ancestry Composition at 23andme, which is what the 500 years old comment is about, is older than 500 years because it takes thousands of years, or extreme endogamy, for the autosomal DNA to become identifiable as a separate component and for it to also be very widespread. Iberians, from all over Iberia, and Latin Americans, from all over Latin America, get a very high amount of Iberian DNA but it is very low outside those populations. That in itself shows that it is more than 500 years old.

What 23andme is really saying about the 500 years old DNA is that is what your ancestors from 500 years ago had in their Ancestry Composition. At least the ones that you inherited the autosomal DNA from.

jfttrh44
03-13-2016, 12:08 PM
What 23andme is really saying about the 500 years old DNA is that is what your ancestors from 500 years ago had in their Ancestry Composition. At least the ones that you inherited the autosomal DNA from.

I still don“t get what you say about this. So, if you have a more recent ancestor (from a very different ethnicity: for example someone from Finland), it would not appear because it will not affect the total?

So, in fact, all of these tools are always for older ancestry and never for newer (less than 500 years). And what happens with the people from America? All of them are mixed in the last 500 years, inside their race, and outside also.

E_M81_I3A
03-13-2016, 12:24 PM
23andme compares your DNA to modern populations and not to the ancient components that made up these modern populations while most of GedMatch calculators compare your DNA to ancient components instead of modern populations.

For example, some North Africans get 100 % North African ancestry at 23andme which is correct, but it does not say anything about ancient SSA ancestry because it is included in the North African modern populations. Same for some Iberians who get 100% Iberian at 23andme which is correct, but in the same manner it does not say anything about ancient Maghrebi ancestry because it is included in the Iberians modern populations.

When they use GedMatch calculators, North Africans will generally get 15-25% SSA and Iberians will get 5-15% Maghrebi ancestry as ancient components that make up modern populations used by 23andme are here isolated.

So both 23andme and Gedmatch calculators are useful, it just depends on what we want to know.

ArmandoR1b
03-13-2016, 12:58 PM
Same for some Iberians who get 100% Iberian at 23andme which is correct, but in the same manner it does not say anything about ancient Maghrebi ancestry because it is included in the Iberians modern populations.It is extremely uncommon for an Iberian to get 100% Iberian at 23andme and most have a small amount of SSA or MENA along with small amounts of Italian and northern European at 23andme.


When they use GedMatch calculators, North Africans will generally get 15-25% SSA and Iberians will get 5-15% Maghrebi ancestry as ancient components that make up modern populations used by 23andme are here isolated. That of course depends on the person and the Gedmatch calculator. The Spaniards that I have seen results for don't get Maghrebi at Gedmatch although there is almost no doubt that they have some Maghrebi in their ancestry. Which calculator are you using that shows Maghrebi in Iberians?

Kurd
03-13-2016, 01:00 PM
23andMe's speculative mode greatly overestimates major components, and underestimates minor components. This is due to their methodology of snipping the genome into 100 SNP segments to compare against the limited references they have. So for example, if 60% of the the segment indicates Middle Eastern, and 40% indicates S Asian, that segment is assigned 100% Middle Eastern. In effect 40% of the segment, which is S Asian is ignored, and the whole segment is assigned Middle-Eastern.

Also, their methodology includes segment smoothing, which means if there are chunks of minor components in a segment, they are ignored.

That is how Iranians and W Asians turn out 98-100% Middle Eastern, and folks in neighboring Pakistan turn out 98-100% S Asian in speculative mode.

This naturally is unrealistic and uninformative, because you don't need a test to tell you that. Conservative mode is better with regards to inflation of major components and underestimation of minor components, but the trouble here is that people get 5-70% unassigned. This is where your minor components are hidden.

The above translates to 23andMe being useless for figuring out your minor components to any degree of accuracy

E_M81_I3A
03-13-2016, 01:45 PM
The Spaniards that I have seen results for don't get Maghrebi at Gedmatch although there is almost no doubt that they have some Maghrebi in their ancestry. Which calculator are you using that shows Maghrebi in Iberians?

Some calculators have a North african component like Dodecad 12b, MDLP 23... Whenever there is a North african component then almost all Iberians get 5-15% of it. If there is no North african component then they get 1-4% SSA which is generally a signal of more important North African ancestry.

For example, a Galician's Dodecad 12b results (7% NAF and 3% SSA):

http://img11.hostingpics.net/thumbs/mini_519248GalicianDodecad12b.jpg (http://www.hostingpics.net/viewer.php?id=519248GalicianDodecad12b.jpg)

and a Portuguese (7% NAF and 1.5% SSA)

http://img11.hostingpics.net/thumbs/mini_659850PortugueseGedmatch.jpg (http://www.hostingpics.net/viewer.php?id=659850PortugueseGedmatch.jpg)


So far, I have not seen any Iberians, at least Western Iberians, at GedMatch without North african and/or SSA ancestry, but I am sure there are some in Eastern part of Iberia. Which region of Iberia is your ancestry from? have you tried Dodecad 12b for example ?

ArmandoR1b
03-13-2016, 02:16 PM
Some calculators have a North african component like Dodecad 12b, MDLP 23...

Like I said "That of course depends on....the Gedmatch calculator". Not all of the Gedmatch calculators have a North African component and there isn't a huge consensus on which Gedmatch calculator is the best. I know for sure there is a significant amount of Maghrebi ancestry in Iberians but I haven't found a calculator or a DNA company that more easily identifies Iberians vs other southern Europeans.

jfttrh44
03-13-2016, 03:13 PM
23andme compares your DNA to modern populations and not to the ancient components that made up these modern populations while most of GedMatch calculators compare your DNA to ancient components instead of modern populations.

For example, some North Africans get 100 % North African ancestry at 23andme which is correct, but it does not say anything about ancient SSA ancestry because it is included in the North African modern populations. Same for some Iberians who get 100% Iberian at 23andme which is correct, but in the same manner it does not say anything about ancient Maghrebi ancestry because it is included in the Iberians modern populations.

When they use GedMatch calculators, North Africans will generally get 15-25% SSA and Iberians will get 5-15% Maghrebi ancestry as ancient components that make up modern populations used by 23andme are here isolated.

So both 23andme and Gedmatch calculators are useful, it just depends on what we want to know.

You know very little about Iberian genetics.

1ŗ. This "Maghrebi" ancestry is Neolithic and in fact does not mean that Iberians have old north-african ancestry, but that they share a small common ancestor (it can be the reverse: the "Maghrebi" have old Iberian ancestry.

2ŗ. If we suppose that the North-African component is Neolithic (as many studies confirm), it is not "Maghrebi", it is not "muslim", it is not necesarily "non white", because North-African have also white/caucasoid genes. In fact, there is a theory that says that during the Neolithic some proto-berebers came to Western Spain, and also went up North, to other countries. I have seen French, Belgium, and UK people with 0.1-0.3% North-African, that could also be from the same origin than the Iberian one (old Neolithic).

3ŗ: That 5-15% is ridiculous. Most are 0.5-1.5%. In the case they get something. It is more a West Iberian than an East Iberian. Most Spaniards are >99% European (the other 1% is ancient), and that is more than many other countries, including nordic ones, that have a 96-98% as a media (due to Uralic ancient ancestry).

jfttrh44
03-13-2016, 03:20 PM
Some calculators have a North african component like Dodecad 12b, MDLP 23... Whenever there is a North african component then almost all Iberians get 5-15% of it. If there is no North african component then they get 1-4% SSA which is generally a signal of more important North African ancestry.

For example, a Galician's Dodecad 12b results (7% NAF and 3% SSA):

http://img11.hostingpics.net/thumbs/mini_519248GalicianDodecad12b.jpg (http://www.hostingpics.net/viewer.php?id=519248GalicianDodecad12b.jpg)

and a Portuguese (7% NAF and 1.5% SSA)

http://img11.hostingpics.net/thumbs/mini_659850PortugueseGedmatch.jpg (http://www.hostingpics.net/viewer.php?id=659850PortugueseGedmatch.jpg)


So far, I have not seen any Iberians, at least Western Iberians, at GedMatch without North african and/or SSA ancestry, but I am sure there are some in Eastern part of Iberia. Which region of Iberia is your ancestry from? have you tried Dodecad 12b for example ?

2 examples in a population of 45 million of people.

If I search I could probably find French examples with ancient North-Africa or SSA ancestry. From 100% ethnic french people.

Again, having North-African ancestry means very little if you pretend to extract some conclussions as berebers and other unmixed (SSA) north-africans obviously have a lot of white/caucasoid genes. You are french, so you know who Zidane is. 5000 years ago, the neolithic North-African ancestry you see in Western Iberians, is more likely similar to a "Zidane" than to a normal North-African of today (which are heavily mixed with SSA).

As you can see in the distance of the 1ŗ example, his first population is Galicia at 8, that means it is not an average Galician, but a more "extreme-rare" one.



And finally, look at this thread (a northern french test):
http://www.anthrogenica.com/showthread.php?6399-my-family-s-results-%28Northern-France-at-the-border-with-Belgium


0.1% Middle Eastern & North African
0.1% North African

0.1% East Asian & Native American
East Asian
0.1% Broadly East Asian

South_Asian 0.88%
Amerindian 0.46%

His father:

0.1% Sub-Saharan African
0.1% West African

0.1% East Asian & Native American
East Asian
0.1% Broadly East Asian

Red_Sea 0.76%
South_Asian 0.25%
Amerindian 1.64%

His girlfriend:
0.2% Middle Eastern & North African
0.2% North African
< 0.1% Broadly Middle Eastern & North African

Red_Sea 0.05%
South_Asian 0.70%
Siberian 0.47%
Amerindian 0.23%
Oceanian 0.55%
Northeast_African 0.18%

His gf's maternal grandmother:
0.1% Middle Eastern & North African
0.1% Broadly Middle Eastern & North African
0.2% Unassigned

Amerindian 0.36%
Oceanian 0.56%
Sub-Saharan 0.26%




An example of a Northern-French with also ancient Middle Eastern & North African and SSA ancestry.

I am sure that many French have similar or greater results in these aspects.

ffoucart
03-13-2016, 03:39 PM
I'm approving about some ancient admixture for North African and SSA. Too many Northern European have those kinds of minor admixture. There is no way that all of them have some recent North African or SSA ancestor. The same can be said to East Asian (very common too).
As an example, I share with Scandinavians and some of them have some SSA.

Lugus
03-13-2016, 03:41 PM
2 examples in a population of 45 million of people.

If I search I could probably find French examples with ancient North-Africa or SSA ancestry. From 100% ethnic french people.

Again, having North-African ancestry means very little if you pretend to extract some conclussions as berebers and other unmixed (SSA) north-africans obviously have a lot of white/caucasoid genes. You are french, so you know who Zidane is. 5000 years ago, the neolithic North-African ancestry you see in Western Iberians, is more likely similar to a "Zidane" than to a normal North-African of today (which are heavily mixed with SSA).

As you can see in the distance of the 1ŗ example, his first population is Galicia at 8, that means it is not an average Galician, but a more "extreme-rare" one.

What's the problem with North African ancestry or any other? I get 5% on Dodecad 12b but as far as I'm concerned I could also get 50% plus another 50% SSA... I'm proud of ALL my SNPs. Each one is there for some historical and biological reason and that's what makes genetic genealogy fascinating.

By the way, I think Dodecad is correct with 5% and not 23andme, which gives me only a ridiculous 0.2%.

ffoucart
03-13-2016, 04:08 PM
What's the problem with North African ancestry or any other? I get 5% on Dodecad 12b but as far as I'm concerned I could also get 50% plus another 50% SSA... I'm proud of ALL my SNPs. Each one is there for some historical and biological reason and that's what makes genetic genealogy fascinating.

By the way, I think Dodecad is correct with 5% and not 23andme, which gives me only a ridiculous 0.2%.
I'm only speaking for myself, and I have no problem with MENA or SSA (jfttr44 used my own results).

But the problem is not only about MENA and SSA, it's for the ethnicities used by 23andme. Some are completely left out (due to lack of samples), and French & German is also quite problematic (look at my level of British).

Originally, I was 100% French and German (I answered correctly that all my gp were French). Absurd. They changed their algorithm and I got admixed results, and others after being phased against my father.

The problem is that I could have some ancestry in England 500 years ago, but not at this level.

The same can be said for SSA and MENA. There were some Africans who went to Low Countries. But it's unlikely that I have more than 1 or 2 of them on thousands of ancestors. Moreover it's unlikely that many people can have the same.

About Southern European, there were flux with North Africa in both ways. So having some common ancestry can be from either side. And we know by History that: Reconquista was also a colonization from Northern Spain/South of France, and many Spanish Muslims didn't stay in Spain but went to North Africa.

So, between Prehistoric flux and Historical flux, what is surprising is the low level of admixture.

Probably because, North Africa did have its own flu. From Sub Sahara Africa and Middle East.

Scarlet Ibis
03-13-2016, 04:26 PM
Thread closed for review.

Scarlet Ibis
03-14-2016, 06:46 AM
Thread has been cleaned up, and reopened.

A reminder to keep the conversation civil. Thank you.

E_M81_I3A
03-14-2016, 07:22 AM
3ŗ: That 5-15% is ridiculous. Most are 0.5-1.5%. In the case they get something. It is more a West Iberian than an East Iberian. Most Spaniards are >99% European (the other 1% is ancient), and that is more than many other countries, including nordic ones, that have a 96-98% as a media (due to Uralic ancient ancestry).

Sorry, in fact I forgot to mention the sources. See below.

According to several of the biggest autosomal studies about European genetics in the last 3 years, Iberians get, on average, about 12% North African ancestry :

1) Lazaridis et al 2014 : "Ancient human genomes suggest three ancestral populations for present-day Europeans"

http://biorxiv.org/content/early/2013/12/23/001552

Authors found an average African admixture of 14.8% (12.6% Mozabite and 2.2% Mbuti/Yoruba) in the Spanish population, confirming that gene flow from Sub-Saharan and North African populations has occurred.

See Supplementary Information page 65, Table S11.2 below

Samples analysed from all regions of spain:

Spanish_Pais_Vasco_IBS Pais_Vasco 5)
Spanish_ Leon_IBS Spanish 5
Spanish_Andalucia_IBS Spanish 4
Spanish_Aragon_IBS Spanish 6
Spanish_Baleares_IBS Spanish 4
Spanish_Cantabria_IBS Spanish 5
Spanish_Cataluna_IBS Spanish 5
Spanish_Extremadura_IBS Spanish 5
Spanish_Galicia_IBS Spanish 5
Spanish_Mancha_IBS Spanish 5
Spanish_Murcia_IBS Spanish 5
Spanish_Valencia_IBS Spanish 5

http://img15.hostingpics.net/pics/620068Lazaridis2014.jpg (http://www.hostingpics.net/viewer.php?id=620068Lazaridis2014.jpg)



2) Botigué et al 2013, "Gene flow from North Africa contributes to differential human genetic diversity in southern Europe"

http://www.pnas.org/content/110/29/11791.full

"Southwestern European populations average between 4% and 20% of their genomes assigned to a North African ancestral cluster (SI Appendix, Fig. S3), whereas this value does not exceed 2% in southeastern European populations."

According to David Comas, one of the authors of the study : "We have analyzed a large battery of SNPs in 2,099 individuals from 43 populations across North Africa and Europe, and have shown that there has been extensive gene-flow (in substantially higher amounts than previously thought) between North Africa and the Iberian Peninsula (ranging from ~5 to 15%, with low levels in Basques), but not in other regions in Europe" (http://biologiaevolutiva.org/dcomas/estimating-gene-flow-from-north-africa-to-southern-europe/).

See also from David Comas about the same study : "La cifra del 20% sólo se da en Canarias, para el resto del paķs oscila entre el 10% y 12%" (http://www.huffingtonpost.es/2013/06/03/los-espanoles-somos-los-e_n_3379814.html)

ffoucart
03-14-2016, 07:57 AM
I find extremely interesting that, overall, the map of North African admixture in Europe given by Botigué is very similar to the Roman Empire. Probably meaning that some of the admixture is from the Roman period, and not recent.

The areas where there is a high level of North African admixture seem to be well defined in the Iberian Peninsula (Tenerife?). Extreme South West. Could also be linked to the Spanish colonial empire.

Odyss
03-14-2016, 08:41 AM
The areas where there is a high level of North African admixture seem to be well defined in the Iberian Peninsula (Tenerife?). Extreme South West. Could also be linked to the Spanish colonial empire.

Tenerife is a Canary Island , so it would fit with the indigenous contribution here.

Lugus
03-14-2016, 11:02 AM
"Southwestern European populations average between 4% and 20% of their genomes assigned to a North African ancestral cluster (SI Appendix, Fig. S3), whereas this value does not exceed 2% in southeastern European populations."


That's what every genetic study of Iberian populations shows. It's also clear from both maternal and paternal haplogroups. See also Moorjani's paper The History of African Gene Flow into Southern Europeans, Levantines, and Jews:

http://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1001373

We Iberians are an amazing mix of different peoples and that is exactly what enabled us to get our place in history. Just compare how many people speak Portuguese and German. The Portuguese language is still expanding in the number of native speakers (especially in Africa) whereas almost every other European language is shrinking, except Spanish.

ffoucart
03-14-2016, 11:20 AM
Tenerife is a Canary Island , so it would fit with the indigenous contribution here.

Tenerife is not cited in the article. I simply don't remember the name of a specific town in South Western Spain (my mistake).

There is a more recent study about African Admixture by Hernandez et alia (October 28, 2015), at PLOS one, using Mt haplogroups, and concluding to sizeable ancient African contribution, depending on the region.

The more recent one (account between 35% and 65% of the total) can be attributed to late Roman Empire and Islamic domination (with slave trade).

So it means that a lot of North African admixture in European, even Southern Europeans, and even in the Iberian Peninsula predates the Islamic conquest.

It's also interesting to note a low level African admixture even in Northern Europeans.

ArmandoR1b
03-14-2016, 12:38 PM
I still don“t get what you say about this. So, if you have a more recent ancestor (from a very different ethnicity: for example someone from Finland), it would not appear because it will not affect the total?That's not at all what I was saying. What I was saying is that if you a more recent ancestor (from a very different ethnicity: for example someone from Finland), it would show your Finnish DNA as it your ancestors had it 500 years. I'm also saying that if you had a more distant ancestor (from a very different ethnicity: for example someone from Finland), from 5-7 generations ago, that you might not see that DNA in your Ancestry Composition at 23andme because of how recombination works. However, more ancient ancestry from a very different ethnicity will show up, to some degree, if enough of your ancestors were admixed from the distant ancestry such as the North African in Iberians. Of course, if the populations are closely related, or have migrations in both directions, the admixture can see be hard to define to a high degree of accuracy.


So, in fact, all of these tools are always for older ancestry and never for newer (less than 500 years). And what happens with the people from America? All of them are mixed in the last 500 years, inside their race, and outside also.This is all explained in my response above.

jfttrh44
03-14-2016, 01:28 PM
What's the problem with North African ancestry or any other? I get 5% on Dodecad 12b but as far as I'm concerned I could also get 50% plus another 50% SSA... I'm proud of ALL my SNPs. Each one is there for some historical and biological reason and that's what makes genetic genealogy fascinating.

By the way, I think Dodecad is correct with 5% and not 23andme, which gives me only a ridiculous 0.2%.

Who said there was a problem? I said that it was false.


Sorry, in fact I forgot to mention the sources. See below.

According to several of the biggest autosomal studies about European genetics in the last 3 years, Iberians get, on average, about 12% North African ancestry :

1) Lazaridis et al 2014 : "Ancient human genomes suggest three ancestral populations for present-day Europeans"

http://biorxiv.org/content/early/2013/12/23/001552

Authors found an average African admixture of 14.8% (12.6% Mozabite and 2.2% Mbuti/Yoruba) in the Spanish population, confirming that gene flow from Sub-Saharan and North African populations has occurred.

See Supplementary Information page 65, Table S11.2 below

Samples analysed from all regions of spain:

Spanish_Pais_Vasco_IBS Pais_Vasco 5)
Spanish_ Leon_IBS Spanish 5
Spanish_Andalucia_IBS Spanish 4
Spanish_Aragon_IBS Spanish 6
Spanish_Baleares_IBS Spanish 4
Spanish_Cantabria_IBS Spanish 5
Spanish_Cataluna_IBS Spanish 5
Spanish_Extremadura_IBS Spanish 5
Spanish_Galicia_IBS Spanish 5
Spanish_Mancha_IBS Spanish 5
Spanish_Murcia_IBS Spanish 5
Spanish_Valencia_IBS Spanish 5

http://img15.hostingpics.net/pics/620068Lazaridis2014.jpg (http://www.hostingpics.net/viewer.php?id=620068Lazaridis2014.jpg)



2) Botigué et al 2013, "Gene flow from North Africa contributes to differential human genetic diversity in southern Europe"

http://www.pnas.org/content/110/29/11791.full

"Southwestern European populations average between 4% and 20% of their genomes assigned to a North African ancestral cluster (SI Appendix, Fig. S3), whereas this value does not exceed 2% in southeastern European populations."

According to David Comas, one of the authors of the study : "We have analyzed a large battery of SNPs in 2,099 individuals from 43 populations across North Africa and Europe, and have shown that there has been extensive gene-flow (in substantially higher amounts than previously thought) between North Africa and the Iberian Peninsula (ranging from ~5 to 15%, with low levels in Basques), but not in other regions in Europe" (http://biologiaevolutiva.org/dcomas/estimating-gene-flow-from-north-africa-to-southern-europe/).

See also from David Comas about the same study : "La cifra del 20% sólo se da en Canarias, para el resto del paķs oscila entre el 10% y 12%" (http://www.huffingtonpost.es/2013/06/03/los-espanoles-somos-los-e_n_3379814.html)

Those studies are biased, and done with bad methodology. There are tons of other different studies that says that the "African" blood in Iberia is between 0.5-2%, and in other parts of Europe between 0.2 and 1%. So, basically, very little difference. I would like to see I very complete study with ALL the regions of Spain and with a bigger number of samples (and not 100 or 200), because they always take Andalucia, Canarias, and maybe Galicia, which are probably the most special regions in terms of genetics.

The last one is just a newspaper article (and that study was even denied later by the same authors)

For the Mozebites, etc: they took Huelva (the open door to slave trade in Spain), Galicia and Canarias (the place in Spain with a special distinction in genetics) and they did a promedium.

It is like you have a place with a 20% a place with a 0.5% and you do the promedium and it says that in "all the country" there is a 10%.

It has also been denied because it was done with bad methodology:
http://www.eupedia.com/forum/threads/31580-Role-of-recent-admixture-in-last-1500-years/page3

I would also like to see the same studies with people from other parts of Europe. There are more Asian genes in Northern Europe than African genes in Southern Europe.

BTW, if the studies are done now with today“s populations, your country: France, will be the one with more SSA and NorthAfrican by far, probably x4, or x5 than Iberia.

Lugus
03-14-2016, 04:42 PM
There are tons of other different studies that says that the "African" blood in Iberia is between 0.5-2%, and in other parts of Europe between 0.2 and 1%

I would like to see one of those studies about the frequency of North African alleles in Europe.

For what it's worth, I'll post my Dodecad 12b results (not showing 0.8% East African and 0.43% SSA). What do you get in yours?

Admix Results (sorted):

# Population Percent
1 Atlantic_Med 42.39
2 North_European 25.03
3 Caucasus 14.22
4 Gedrosia 7.20
5 Northwest_African 5.00
6 Southwest_Asian 4.46


Finished reading population data. 223 populations found.
12 components mode.

--------------------------------

Least-squares method.

Using 1 population approximation:
1 Galicia @ 7.263640
2 Extremadura @ 7.506670
3 Baleares @ 7.803439
4 Portuguese @ 8.049597
5 N_Italian @ 9.889225
6 Castilla_Y_Leon @ 10.388924
7 Murcia @ 10.600319
8 North_Italian @ 10.766053
9 Cataluna @ 11.318843
10 Spaniards @ 11.592388
11 Canarias @ 11.790305
12 Spanish @ 11.822064
13 Andalucia @ 12.439163
14 Castilla_La_Mancha @ 13.758139
15 Cantabria @ 13.892420
16 French @ 14.096252
17 French @ 14.336542
18 Valencia @ 14.359070
19 Aragon @ 15.347192
20 TSI30 @ 16.890079

Using 2 populations approximation:
1 50% N_Italian +50% Portuguese @ 4.001249


Using 3 populations approximation:
1 50% Extremadura +25% Irish +25% Sicilian @ 1.343012


Using 4 populations approximation:
1 Canarias + Cataluna + Orkney + Sicilian @ 0.743195
2 Argyll + Canarias + Cataluna + Sicilian @ 0.748027
3 C_Italian + Canarias + Extremadura + Kent @ 0.773323
4 Canarias + Cataluna + Orcadian + Sicilian @ 0.808625
5 C_Italian + Canarias + CEU30 + Extremadura @ 0.822483
6 C_Italian + Canarias + English + Extremadura @ 0.824081
7 Canarias + Cataluna + Irish + Sicilian @ 0.855611
8 Canarias + Orkney + Sicilian + Spanish @ 0.861647
9 Canarias + Mixed_Germanic + Murcia + O_Italian @ 0.871774
10 Canarias + Orkney + S_Italian_Sicilian + Spanish @ 0.894902
11 Argyll + Canarias + Sicilian + Spanish @ 0.899696
12 Canarias + Orkney + Sicilian + Spaniards @ 0.936799
13 C_Italian + Canarias + Kent + Portuguese @ 0.940603
14 C_Italian + Canarias + CEU30 + Portuguese @ 0.948301
15 Argyll + Canarias + Cantabria + Sicilian @ 0.972160
16 C_Italian + Canarias + English + Portuguese @ 0.972325
17 Canarias + Castilla_Y_Leon + Orkney + S_Italian_Sicilian @ 0.989692
18 Argyll + C_Italian + Canarias + Extremadura @ 1.007597
19 Canarias + Mixed_Germanic + Portuguese + Tuscan @ 1.008576
20 Canarias + Orcadian + Sicilian + Spanish @ 1.018060

Lugus
03-14-2016, 04:44 PM
Using 1 population approximation:
1 Galicia @ 7.263640
2 Extremadura @ 7.506670
3 Baleares @ 7.803439
4 Portuguese @ 8.049597
5 N_Italian @ 9.889225
6 Castilla_Y_Leon @ 10.388924
7 Murcia @ 10.600319

As far as I know (4 generations back) I don't have any Galician ancestry. All my 4 grandparents come from Central Portugal.

ffoucart
03-14-2016, 06:46 PM
As far as I know (4 generations back) I don't have any Galician ancestry. All my 4 grandparents come from Central Portugal.

You have Galicia at 7. It's not near. Anyway, it's logical given that Portugal was created in the Reconquista mainly by people from Galicia.

Moreover, Dodecad 12b is a bit outdated.

And frankly, 4 generations is very recent, genealogically speaking. It's difficult to say "I'm not admixed" with only 4 generations (it's only up to your 16 great great grand parents).

Lugus
03-14-2016, 07:18 PM
Moreover, Dodecad 12b is a bit outdated.

Still waiting for the next generation of more sophisticated autosomal tests and admixture tools.

Yes, 4 generations is nothing and in 23andme I do have DNA relatives in the north of Portugal.

Sikeliot
03-15-2016, 12:18 AM
Lugus, how much SSA do you score on 23andme or any other test you've done (FTDNA, Ancestry, etc.)?

Lugus
03-15-2016, 06:23 AM
Lugus, how much SSA do you score on 23andme or any other test you've done (FTDNA, Ancestry, etc.)?

So far I've only done 23andme. I got 0.1% east Africa. I recently ordered Family Finder from FTDNA and I'll post the results as soon as I get them.

Just to show how fluid things are regarding autosomal tests: my initial result was 0.3% west Africa and it went down after phasing against both my parents, but then I also got 0.2% unassigned (total: 99.5% European).

ffoucart
03-15-2016, 06:33 AM
Link tio Hernandez et al 2015
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4624789/

Sikeliot
03-15-2016, 06:51 AM
So far I've only done 23andme. I got 0.1% east Africa. I recently ordered Family Finder from FTDNA and I'll post the results as soon as I get them.

Just to show how fluid things are regarding autosomal tests: my initial result was 0.3% west Africa and it went down after phasing against both my parents, but then I also got 0.2% unassigned (total: 99.5% European).

Interesting. I inherited little of my SSA and I still score more than you. :lol:

ADW_1981
03-15-2016, 01:28 PM
In my case, I'm about 1% SSA at 23andMe as well as Gedmatch. It doesn't affect my oracle results at all since it appears to be from just 1 ancestor who lived 400 years ago or more and is represented by a single segment of DNA. It traces back somewhere in Europe I am certain.

Lugus
03-15-2016, 05:07 PM
In my case, I'm about 1% SSA at 23andMe as well as Gedmatch. It doesn't affect my oracle results at all since it appears to be from just 1 ancestor who lived 400 years ago or more and is represented by a single segment of DNA. It traces back somewhere in Europe I am certain.

Are you sure? Even an admixture of 50% will disappear without trace after 10 generations (250-300 years) unless it gets "replenished" by the population's gene pool. In Portugal practically everyone has a little bit of SSA, so if by chance you lose it in one generation, for sure you'll get it back the next. But that's not the case in the countries of your flags, I think.

My wife isn't Portuguese so my kids immediately lost their SSA. In exchange they got 51% Ashkenazy, which was accurately detected by 23andme (impressive).

Sikeliot
03-15-2016, 05:28 PM
Are you sure? Even an admixture of 50% will disappear without trace after 10 generations (250-300 years) unless it gets "replenished" by the population's gene pool. In Portugal practically everyone has a little bit of SSA, so if by chance you lose it in one generation, for sure you'll get it back the next. But that's not the case in the countries of your flags, I think.

My wife isn't Portuguese so my kids immediately lost their SSA. In exchange they got 51% Ashkenazy, which was accurately detected by 23andme (impressive).

Do you think Portuguese people scoring SSA on 23andme means an ancestor from the slave trade? I find it difficult to believe it's ancient/native to the area or that would be over 500 years old and would likely not show up.

Some people from Madeira score as high as 5% SSA.

ADW_1981
03-15-2016, 05:53 PM
Are you sure? Even an admixture of 50% will disappear without trace after 10 generations (250-300 years) unless it gets "replenished" by the population's gene pool. In Portugal practically everyone has a little bit of SSA, so if by chance you lose it in one generation, for sure you'll get it back the next. But that's not the case in the countries of your flags, I think.

My wife isn't Portuguese so my kids immediately lost their SSA. In exchange they got 51% Ashkenazy, which was accurately detected by 23andme (impressive).

This 1 segment has carried 3 generations. My mother also has Iberian and Non-Specific South European but in small quantities as well. Perhaps they arrived together? Maybe the 400 years was the upper bound but it's difficult to know for sure. It's not recent, but unlikely ancient either.

You're right the 23andMe AC is great. My non-European ex's ancestry is measured exactly correct at 50% in my daughter.

Lugus
03-15-2016, 06:09 PM
Do you think Portuguese people scoring SSA on 23andme means an ancestor from the slave trade? I find it difficult to believe it's ancient/native to the area or that would be over 500 years old and would likely not show up.

Some people from Madeira score as high as 5% SSA.

It can be both ancient and more recent depending on the case. Even when a component is ancient if everyone has it it won't go away.

Madeira is a different story. We know there was a recent African contribution to the population. The same is true about some particular areas of the mainland.

Anyway, African ancestry shouldn't be a big deal in Portugal considering the hundreds of years of close and often problematic connection we had and still have with Africa. We might be geographically in Europe but perhaps the most significant parts of our history unfolded in other continents, at least since the 15th century and until recently.

AJL
03-16-2016, 12:54 AM
Are you sure? Even an admixture of 50% will disappear without trace after 10 generations (250-300 years) unless it gets "replenished" by the population's gene pool. In Portugal practically everyone has a little bit of SSA, so if by chance you lose it in one generation, for sure you'll get it back the next. But that's not the case in the countries of your flags, I think.

My wife isn't Portuguese so my kids immediately lost their SSA. In exchange they got 51% Ashkenazy, which was accurately detected by 23andme (impressive).

23andme has a huge Ashkenazi base, in part because of funding into Parkinson's research (there is a more or less Ashkenazi-specific Parkinson's mutation (http://www.nejm.org/doi/full/10.1056/NEJMc055509)). And Ashkenazim are also historically endogamous. This combination makes 23andme a great "Ashkenazi detector."

One factor that's immensely useful in thinking about how likely a segment is to persist is historical population density (https://armedea.files.wordpress.com/2014/10/mosaic_density_map_homogen_light.jpg).

MoTheaglachNaCruinne
03-30-2016, 02:10 PM
I have to say, I am very impressed by 23andme. I didn't find FtDNA's MyOrigins very helpful which I did for my paternal grandfather. It pinned him at aroudn 10% Eastern European, but he showed nothing of the sort with his 23andme's Ancestry Composition. The more macro-scale admixture percentages reported by FtDNA were very useful, since they put him at 99% European and 1% Unstated, whereas 23andme put him at 99.7% European and 0.2% East Asian; I just chalked that down to being that FtDNA tests more regions in the genome than 23andme, and that there was a little more non-European than what 23andme found. However, it was still less than 1% which is why it was left unreported.

Morges
04-30-2016, 10:34 AM
I think GEDmatch is much more deeper than 23andme.

sciencediver
05-01-2016, 04:24 PM
Gedmatch>ftDNA>AncestryDNA>Geno>DNA.land>23andme

shazou
05-02-2016, 12:41 AM
I might end up trusting AncestryDNA more once their new ethnicity estimate is rolled out this year, especially once I see their new breakdown for the Asia region.

Giuseppina
05-31-2016, 01:02 AM
What about the national geographic project, you think is accurate?

BalkanKiwi
05-31-2016, 01:23 AM
What about the national geographic project, you think is accurate?

National Geographic isn't great for minor ancestry, however it gives a more accurate and broader picture of your overall ancestry. From what I understand Next Gen is better due to the new clusters. I have done Geno 2.0 and didn't get much from it, however it may be beneficial to a new person starting genetic testing. The only downside is the data can't be uploaded to GEDMatch, however it can be modified to be used on DIY calculators (and then possibly onto GEDMatch, I haven't tried).

A Norfolk L-M20
05-31-2016, 01:36 AM
Neither. They're quite different things anyway. As an Englishman with a pretty well recorded, homogeneous English ancestry, I received a mixture of uncertainty, and pan-European composition (on speculative). It's all well to say that this most likely is down to a failure of speculative mode to distinguish the populations of NW Europe, with all of their admixture and shared origins, reaching back to prehistory - but 23andMe claims that this is your ancestry 500 years ago. It is not! If I hadn't belonged to a homogeneous population, with a recorded paper ancestry - if I had been for example, an American customer, with only a slight knowledge of my ancestry, I could have well have swallowed it as accurate. It isn't. Indeed, phasing with one parent, did some bizarre things to my composition. My Scandinavian percentage fell from 7% to 1%! My British & Irish increased from 32% to 37%. More percentage moved into the uncertainty of "broadly" NW European. That was following phasing. So why should I take any unphased Ancestry Composition in speculative mode as remotely accurate?

As for GEDMATCH - the calculators measure genetic distance, not ancestral composition. We tend to pick which ever works best for us. But isn't that bias? For our English family, Eurogenes K13 wins. However, South Dutch sometimes pops up above SE English. It keeps chucking Iberian into the mixtures. It's fun, and certainly interesting, but not really the same sort of tool as 23andMe ancestry composition.

Giuseppina
05-31-2016, 02:29 AM
Then ... Which one you suggest, I'm just intrigued, did n/geo... My mtndna j1c1.. Which find bizarre, am Sicilian...

A Norfolk L-M20
05-31-2016, 02:53 AM
Whatever you think that you can afford. For all of my criticism, I think that genetics genealogy is a lot of fun. I just despair for those that think that autosome testing has any sort of accuracy when it comes to sorting out the populations of NW Europe.

Mike_G
05-31-2016, 02:53 AM
I could see where people from tightly defined geographic ancestries with known bloodlines could take issue with any results, but for me GEDmatch is an extension of 23andme. I might be way off base with this, but for a guy of mixed Western and Eastern European ancestry, GEDmatch is like a paint mixing application. It's taken my basic 23andme results (58/32 NWE/EE) and made a few educated guesses beyond the basic results plus 15% English/Irish and 8.8% Scandinavian results that 23andme provided.

I scratched my head over all the German and Scandinavian results that Eurogenes and MDLP spit out since my reported nationalities are English/Dutch (mother) and Polish (father), but it made sense to me after I thought about it a bit. How do I end up with almost exclusively German or Swedish one population results, Oracle mixes in the North and Baltic seas, and a bunch of Norwegian, Ukrainian and Belorussian results in 3 or 4 population admixes? Well, like 23andme said, I'm a product of those two little areas of water in Northern Europe.

Depending on the formula (going back to the paint mix theory) let's say we assign navy blue to the English, royal blue to the Dutch and a greenish yellow to Polish ancestries (remembering that Polish genes are pretty mixed themselves)....and we end up with forest green. Depending on the GEDmatch calculators forest green could represent German or Swedish roots. Or in the case of Eurogenes K15, Southwest Finland. So we give a lady from Newcastle a few drinks and a guy from Gdansk a couple beers and they have a nice night together. They end up with a kid that wouldn't draw even a passing glance as an outsider in Rostock or Malmo (except for the dark brown hair). I might be wrong with all this, but it's the only way I can square up the GEDmatch results I've gotten.

I'll probably do more testing to get more haplogroup clarity, but I accept 23andme and GEDmatch for what they are...confirmation that the paperwork was as correct as it could be and trying to refine the guesswork a little more. The only disappointing thing is that I don't have any "exotic" stuff like Jewish or Romani ancestry. Or even Italian. Damn...

BalkanKiwi
05-31-2016, 03:12 AM
I could see where people from tightly defined geographic ancestries with known bloodlines could take issue with any results, but for me GEDmatch is an extension of 23andme. I might be way off base with this, but for a guy of mixed Western and Eastern European ancestry, GEDmatch is like a paint mixing application. It's taken my basic 23andme results (58/32 NWE/EE) and made a few educated guesses beyond the basic results plus 15% English/Irish and 8.8% Scandinavian results that 23andme provided.

I scratched my head over all the German and Scandinavian results that Eurogenes and MDLP spit out since my reported nationalities are English/Dutch (mother) and Polish (father), but it made sense to me after I thought about it a bit. How do I end up with almost exclusively German or Swedish one population results, Oracle mixes in the North and Baltic seas, and a bunch of Norwegian, Ukrainian and Belorussian results in 3 or 4 population admixes? Well, like 23andme said, I'm a product of those two little areas of water in Northern Europe.

Depending on the formula (going back to the paint mix theory) let's say we assign navy blue to the English, royal blue to the Dutch and a greenish yellow to Polish ancestries (remembering that Polish genes are pretty mixed themselves)....and we end up with forest green. Depending on the GEDmatch calculators forest green could represent German or Swedish roots. Or in the case of Eurogenes K15, Southwest Finland. So we give a lady from Newcastle a few drinks and a guy from Gdansk a couple beers and they have a nice night together. They end up with a kid that wouldn't draw even a passing glance as an outsider in Rostock or Malmo (except for the dark brown hair). I might be wrong with all this, but it's the only way I can square up the GEDmatch results I've gotten.

I'll probably do more testing to get more haplogroup clarity, but I accept 23andme and GEDmatch for what they are...confirmation that the paperwork was as correct as it could be and trying to refine the guesswork a little more. The only disappointing thing is that I don't have any "exotic" stuff like Jewish or Romani ancestry. Or even Italian. Damn...

As someone with similar ancestry, with Southern Europe thrown in as well, I know how you feel.

A Norfolk L-M20
05-31-2016, 03:30 AM
I agree with you both. Ancestry Composition can throw up some unexpected result that is real. It does identify some exotic populations well. I just despair when I see people quoting their unphased 23andMe AC results in speculative mode, as though they are written in stone.

Mike_G
05-31-2016, 04:17 AM
As someone with similar ancestry, with Southern Europe thrown in as well, I know how you feel.

Those of us from the "new world" are quite the mixed salads, aren't we? Given New Zealand's history I imagine you have a lot of interesting mixes there. When I was stationed in Japan I played golf with a guy from New Zealand who could have been a cousin of mine. Very similar physically and he had the same quiet but friendly demeanor as me. His Australian shipmates were another story. Those guys were rough. A lot of fun though...even if your Kiwi countryman had to do a little translating for me.


I agree with you both. Ancestry Composition can throw up some unexpected result that is real. It does identify some exotic populations well. I just despair when I see people quoting their unphased 23andMe AC results in speculative mode, as though they are written in stone.

I agree. It's fun to see what it can tell you....and not tell you. I wouldn't make any definitive ancestry claims based on any of this stuff without some kind of additional corroborating information.

BalkanKiwi
05-31-2016, 04:23 AM
Those of us from the "new world" are quite the mixed salads, aren't we? Given New Zealand's history I imagine you have a lot of interesting mixes there. When I was stationed in Japan I played golf with a guy from New Zealand who could have been a cousin of mine. Very similar physically and he had the same quiet but friendly demeanor as me. His Australian shipmates were another story. Those guys were rough. A lot of fun though...even if your Kiwi countryman had to do a little translating for me.

I live in Australia, and still have trouble understanding some of them :biggrin1: I don't know if you caught my signature, but my known ancestry on paper is English, Scottish, Irish, Welsh, Croatian, Bosnian, Ashkenazi, Polish and Maori.

Like you said, it's pretty interesting to see how these people from all over the world converge on the same area over time. I think DNA testing is more useful for those with some paper trail, because for those who don't, it becomes a lot of guess work and false assumptions.

Mike_G
05-31-2016, 05:02 AM
I did notice. That is a crazy background on your side. Not crazy in a bad way, but wow. Without cheating and looking into your post history, I'm having a tough time adding it up. English/Scottish/Irish/Welsh + Croatian/Bosnian + Ashkenazi/Polish + Maori? No matter the composition, congrats on getting that sorted out. That's pretty diverse no matter where in the world you are.

BalkanKiwi
05-31-2016, 11:33 PM
I did notice. That is a crazy background on your side. Not crazy in a bad way, but wow. Without cheating and looking into your post history, I'm having a tough time adding it up. English/Scottish/Irish/Welsh + Croatian/Bosnian + Ashkenazi/Polish + Maori? No matter the composition, congrats on getting that sorted out. That's pretty diverse no matter where in the world you are.


Thank you sir! :beerchug: Its probably a good example of those out there who do genetic testing with such diverse backgrounds but don't have a paper trail. It would a confusing genetic puzzle. You mentioned Gdansk, do you have ancestors from Pomerania as well?

Mike_G
06-01-2016, 12:15 AM
I have no idea about my specific roots. I just threw Gdansk out there because it was the first Baltic city that came to mind. I was adopted at birth, but the paperwork said Polish birth father and English/Dutch mother. 23andme indicated the same thing more or less (32% EE). 23andme also says my paternal haplogroup is R1a1a, which is pretty broad as I've discovered. I'd like to see if I could get more specific information than that (FTDNA maybe???).

The cool thing about my adopted parents is that my mom has German and British Isles ancestry while my dad is 100% Polish. So although we don't share the same genes, we do share common European geographic origins. I've pestered my dad for years about where in Poland they came from. Figures it would be the adopted kid who was more curious than the rest. My aunt finally ran it down a few years ago, and it turns out they're from just southwest of Warsaw on one side and southeastern Poland on the other.

Tomenable
04-27-2017, 12:02 AM
23andMe's speculative mode greatly overestimates major components, and underestimates minor components. This is due to their methodology of snipping the genome into 100 SNP segments to compare against the limited references they have. So for example, if 60% of the the segment indicates Middle Eastern, and 40% indicates S Asian, that segment is assigned 100% Middle Eastern. In effect 40% of the segment, which is S Asian is ignored, and the whole segment is assigned Middle-Eastern.

Also, their methodology includes segment smoothing, which means if there are chunks of minor components in a segment, they are ignored.

That is how Iranians and W Asians turn out 98-100% Middle Eastern, and folks in neighboring Pakistan turn out 98-100% S Asian in speculative mode.

This naturally is unrealistic and uninformative, because you don't need a test to tell you that. Conservative mode is better with regards to inflation of major components and underestimation of minor components, but the trouble here is that people get 5-70% unassigned. This is where your minor components are hidden.

The above translates to 23andMe being useless for figuring out your minor components to any degree of accuracy

But isn't this precisely why they claim that they tell us only about the last 500-1000 years of our ancestry? Large unbroken segments indicate recent admixtures, small segments - remote, ancient admixtures.

JerryS.
04-27-2017, 01:51 AM
I used Ancestry for my DNA test. it was in the right countries, but the percentages were way off on all but one ethnicity. loaded raw data to FamilyTree and got the right regions and a little closer to correct percentages, until they redid their new version of my origins and its way wrong. GEDmatch however has things as close as I could have guessed.

geebee
04-27-2017, 02:57 AM
I agree with you both. Ancestry Composition can throw up some unexpected result that is real. It does identify some exotic populations well. I just despair when I see people quoting their unphased 23andMe AC results in speculative mode, as though they are written in stone.

I don't think any company's results should be taken as "written in stone". However, one thing is true of Ancestry Composition that isn't true of the competitors: 23andMe shows you where on the chromosomes segments representing your ancestries are actually located. This makes it possible to compare relatives in ways that mere percentages don't allow.

For example, both FTDNA and Ancestry only tell me I have a "trace" amount of Native American ancestry. For Ancestry, this is estimated as 1%; while FTDNA doesn't put a percentage on it at all -- just "less than 2%", as if the difference between 0.1% and 1.9% doesn't matter.

23andMe not only gives me a percentage (2.0%), it also shows me a total of seven segments it believes to be of Native American origin. I can then compare these to the locations of Native American segments identified for my five siblings, who all show similar percentages. I can also compare my daughter's segments, and see that all of her segments are located in the same places as mine; she merely has fewer of them.

Not only that, but I've been able to compare my results against those of more distant relatives. For instance, I show a Native American segment on chromosome 15 in the same location where a 3rd cousin shares a Native American segment. As it happens, we both descend from the same documented Native American ancestor.

Obviously, this doesn't guarantee that the segment is from her, but it certainly is suggestive. At Ancestry, I'm in a 50-member DNA Circle centered on a descendant of the Native American ancestor. Many of the other members of this circle have Native American at least at the trace level, and often more. However, because there is not a chromosome browser at Ancestry, or any sort of attempt to say where any ancestral segments might be located, there's no way to see what sort of overlap there might be.

So while 23andMe's Ancestry Composition isn't perfect, I think it's the tool that has the greatest potential for verification.

EDIT: Incidentally, it was 23andMe that first allowed me to see that I inherited Native American segments from both of my maternal grandparents. I'd already seen documentation of a Native American ancestor on my grandmother's side, but only recently have I learned anything about my grandfather's Native American ancestry.

What was particularly interesting was that while all of my siblings have at least one NA segment on chromosome 6, we weren't all lining up quite right. I don't simply mean that our segments weren't in exactly the same locations, but that some of us had segments in locations where we didn't share grandparental DNA with others.

Eventually, I realized that we matched two different sets of people with different Native American ancestry. The one group includes people with French colonial ancestry from the Louisiana/Mississippi Gulf coast. They have in common the Native American ancestor of my grandmother.

The other group leads by to folks that Ancestry includes in its "Early Settlers of Eastern Kentucky & Northeast Tennessee" Genetic Community.

2nd EDIT: I should probably add that my own results are actually not unphased, but phased against my father's results. My daughter's results, which are consistent with mine, are phased against both the results of parents.

Dibran
04-27-2017, 11:13 PM
Yes, you are right that Greeks as well score pretty high on "Balkan" on 23andme, it would also make sense historically.

I am actually a Gheg Albanian from Western Macedonia with no known migrations to Macedonia. I think we (Gheg Albanians) score the highest Balkan. Historically, we have been the most isolated, therefore less mixed group in the Balkans, so the high Balkan percentage would make sense.

My countries of ancestry on 23andme are in this order: Greece, Albania (close second), Romania, Bulgaria, Serbia. Surprisingly no Macedonia. Important to note that my matches in Greece all have Arvanite/Albanian ancestry and are from NW Greece close to the border with Albania.

I will have to post mine later but my countries of ancestry were pretty much the same as you. In order I had Greece, Albania, Romania, Montenegro, Bulgaria.

I came up 94 percent Balkan. My father came up 88 percent. Gheg from Okshtun with origin from Mirdita, surprisingly(assuming Ghegs 90+). So I assume my mother contributed some extra.


Me:
European
98.5%
Southern European
97.1%
Balkan
94.0%
Italian
0.9%
Broadly Southern European
2.2%
Northwestern European
0.3%
British & Irish
0.3%
Broadly European
1.1%
Middle Eastern & North African
0.8%
Middle Eastern
0.8%
East Asian & Native American
0.5%
East Asian
0.4%
Japanese
0.3%
Mongolian
< 0.1%
Broadly East Asian & Native American
< 0.1%
Unassigned
0.3%


Father:


European
98.7%
Southern European
97.4%
Balkan
88.2%
Broadly Southern European
9.2%
Northwestern European
0.7%
British & Irish
0.3%
Broadly Northwestern European
0.4%
Broadly European
0.7%
Middle Eastern & North African
1.1%
Middle Eastern
1.1%
South Asian
< 0.1%
Broadly South Asian
< 0.1%
Unassigned
0.1%


How did you tell they were arvanites btw? Most mine were also in northwest Greece. One had the name arvanitakis which is a give away. The rest I couldn't be sure if arvanites or just Greeks.

JerryS.
04-28-2017, 12:37 AM
I will have to post mine later but my countries of ancestry were pretty much the same as you. In order I had Greece, Albania, Romania, Montenegro, Bulgaria.

I came up 94 percent Balkan. My father came up 88 percent. Gheg from Okshtun with origin from Mirdita, surprisingly(assuming Ghegs 90+). So I assume my mother contributed some extra.


Me:
European
98.5%
Southern European
97.1%
Balkan
94.0%
Italian
0.9%
Broadly Southern European
2.2%
Northwestern European
0.3%
British & Irish
0.3%
Broadly European
1.1%
Middle Eastern & North African
0.8%
Middle Eastern
0.8%
East Asian & Native American
0.5%
East Asian
0.4%
Japanese
0.3%
Mongolian
< 0.1%
Broadly East Asian & Native American
< 0.1%
Unassigned
0.3%


Father:


European
98.7%
Southern European
97.4%
Balkan
88.2%
Broadly Southern European
9.2%
Northwestern European
0.7%
British & Irish
0.3%
Broadly Northwestern European
0.4%
Broadly European
0.7%
Middle Eastern & North African
1.1%
Middle Eastern
1.1%
South Asian
< 0.1%
Broadly South Asian
< 0.1%
Unassigned
0.1%


How did you tell they were arvanites btw? Most mine were also in northwest Greece. One had the name arvanitakis which is a give away. The rest I couldn't be sure if arvanites or just Greeks.

it is my understanding that Greek last names ending in -akis are native to Crete.

Dibran
04-28-2017, 02:07 PM
it is my understanding that Greek last names ending in -akis are native to Crete.

Interesting. Considering(to my knowledge) arvanite is the Greek term for Albanians.

Maybe an arvanites from the mainland moved to Crete and adjusted the name? Just a thought.

JerryS.
04-29-2017, 12:41 AM
Interesting. Considering(to my knowledge) arvanite is the Greek term for Albanians.

Maybe an arvanites from the mainland moved to Crete and adjusted the name? Just a thought.

the west end of the island from hannia (candia) souda bay area has scads of -akis, particularly on the south shores.....

serena297
08-06-2017, 03:52 AM
Something between the two? Both needs work, but GEDMatch seems to take a wider approach in terms of including many different ethnicities whereas many ethnicities even in the Euro category on 23andme could use some breaking up let alone the fact that outside of the Euro category everything just seems to be lumped together.

Jbr10291967
08-20-2017, 03:38 AM
Same with me and my 23andme experience. Through my paternal Grandmother, we have documented Native American Indian. Mexican and Spanish heritage going back 9 generations in my direct line and yet 23andme barely picked up any of my Native American roots (showing at less that 1%) and although 1.4 % southern European showed up, all available categories (Sardinian, Spanish/Iberian, etc) all showed 0.0%. Also I showed British and Irish (which is it?) and French and German (which is it?). Not very specific although yes, I have been able to trace French on my Paternal Grandfather's side and German on my Maternal Grandfather's side. Gedmatch on the other hand not only picked up my Spanish/Iberian roots but using the Oracle 4 feature, it specifically pinpointed Andalusian populations which - Andalusia, Spain- is where my Paternal Grandmother's line traces to. Can't say I know a lot about this DNA testing but, definitely know over 200 years of my family history.

Dimanto
08-20-2017, 02:01 PM
Same with me and my 23andme experience. Through my paternal Grandmother, we have documented Native American Indian. Mexican and Spanish heritage going back 9 generations in my direct line and yet 23andme barely picked up any of my Native American roots (showing at less that 1%) and although 1.4 % southern European showed up, all available categories (Sardinian, Spanish/Iberian, etc) all showed 0.0%. Also I showed British and Irish (which is it?) and French and German (which is it?). Not very specific although yes, I have been able to trace French on my Paternal Grandfather's side and German on my Maternal Grandfather's side. Gedmatch on the other hand not only picked up my Spanish/Iberian roots but using the Oracle 4 feature, it specifically pinpointed Andalusian populations which - Andalusia, Spain- is where my Paternal Grandmother's line traces to. Can't say I know a lot about this DNA testing but, definitely know over 200 years of my family history.

''going back 9 generations'' :biggrin1:

Batroun
08-20-2017, 04:53 PM
23andMe for Lebanese people and probably middle eastern people in general sucks because it gives you Middle Eastern/North African which is so broad. The only other things I scored were Italian and "Broadly Southern European". Gedmatch broke it down really well for me. Even if I dont understand what West caucasian means or or West Med it still provided much more detail

Jbr10291967
08-20-2017, 06:01 PM
''going back 9 generations'' :biggrin1:

Tried to post my Dodecad info my keep getting a error prompt saying I am not allowed to post links, Anyway,

Yeah the 9 generations and beyond is my paternal Grandmother's direct line (sorry I wasn't more specific. other direct lines rooting off my Paternal Grandmother's direct line go into Mexico and another traces to the Canary Islands. Got lucky with her line that most those people had remained in the same area of Sabine Parish, Louisiana as mentioned, for hundreds of years before (in my families case) moving to Texas. My Direct X6 Great Grandfather was Antonio Gil Y'barbo and Juan Curbello (1680-1760) who came to San Antonio, Texas from Lancerote, Canary Islands, Spain.

Jbr10291967
08-20-2017, 07:30 PM
23andMe for Lebanese people and probably middle eastern people in general sucks because it gives you Middle Eastern/North African which is so broad. The only other things I scored were Italian and "Broadly Southern European". Gedmatch broke it down really well for me. Even if I dont understand what West caucasian means or or West Med it still provided much more detail

Yeah I got that Broadly Southern European as well. Another interesting site that also picked up some of my known Spanish heritage was DNA Land. In addition, it showed specific regions throughout Europe which did support 23andme and their 97% general European. Additional findings on DNA Land were (3) areas of Italy, (2) areas of Spain (specifically the Iberian Peninsula which supports my known ancestry). Unexpected was (3) areas of Italy although Gedmatch Dodecad World9 showed lots of Italian on their Oracle 4 feature and the Eurogenes K36 Calculator also showed close to 10% Italian) as well as my known Iberian roots and Bascue (unspecified Spanish/French) roots which I likely have both.
these make sense as they fill in the blanks that 23ndme had missed. Another difference was that 23andme says no Ashkenazi but yet, both my Heritage and the Eurogenes Jtest also picked up Ashkenazi.
by the way, any opinions on DNA Land or the different results from one company to the other ?

Batroun
08-21-2017, 01:30 AM
Yeah I got that Broadly Southern European as well. Another interesting site that also picked up some of my known Spanish heritage was DNA Land. In addition, it showed specific regions throughout Europe which did support 23andme and their 97% general European. Additional findings on DNA Land were (3) areas of Italy, (2) areas of Spain (specifically the Iberian Peninsula which supports my known ancestry). Unexpected was (3) areas of Italy although Gedmatch Dodecad World9 showed lots of Italian on their Oracle 4 feature and the Eurogenes K36 Calculator also showed close to 10% Italian) as well as my known Iberian roots and Bascue (unspecified Spanish/French) roots which I likely have both.
these make sense as they fill in the blanks that 23ndme had missed. Another difference was that 23andme says no Ashkenazi but yet, both my Heritage and the Eurogenes Jtest also picked up Ashkenazi.
by the way, any opinions on DNA Land or the different results from one company to the other ?


Only ones I have done are 23andMe and im still waiting on my National Geographic one to come in.. I uploaded to Geneplaza, interesting results but they dont make much sense

jortita
08-21-2017, 04:33 AM
I think a fact that people tend to ignore is that you cannot use Gedmatch ancestry calculators, unless you have your raw data from 23andMe/FTDNA/LivingDNA/Nat Geo/Ancestry DNA/My Heritage DNA or others. So there is no argument on which is better as one is dependant on the other

JerryS.
08-21-2017, 12:50 PM
I think a fact that people tend to ignore is that you cannot use Gedmatch ancestry calculators, unless you have your raw data from 23andMe/FTDNA/LivingDNA/Nat Geo/Ancestry DNA/My Heritage DNA or others. So there is no argument on which is better as one is dependant on the other

the data while from the same source is calculated vastly different for many people though. for instance Ancestry gives me less than 1% Iberian, as in damn near non-existent, while Eurogenes K36 gives me 12% Iberian!

pigeonflight
09-06-2017, 06:30 PM
I want to say GedMatch has been accurate for me. So I got my 23andMe results RIGHT after they installed the V5 chip, on August 18th. Most of it was shuffled into the "Broadly" categories of Northwestern and Broadly European. Irish was also lumped in with the British Isles.

Known Ancestry: Back in 2009 my maternal grandparents got their DNA tested with National Geographic as a gift. Grandma always claimed to be blend of Irish and Native American, but he results showed East Poland, Ukrainian and Russian. I know this happens quite a bit, no one wants to call Grandma a liar. I'm sure she was told that her grandmother (who did not speak English, according to Grandma) was a native American rather than a Ukranian immigrant to hide something. This also explain the very unique, non Irish names that she, her sister, her mother and Grandmother all had. It's hard to find a paper trail on my grandmother's family as they just randomly showed up around 1920 after my great-grandmother was married (grandma was born in 1924). My Grandfather showed British Isles (no Ireland) and a touch of Spanish/Italian.

Going into 23andMe, grandma hunk of DNA appeared under the label of Eastern European and "Broadly European". Grandpa's Results were "British/Irish" and "General European".

Unexpected results of me was that not much of my father's documented DNA lineage showed up. I should have been way more eastern European (even allowing for high variance between generations), because my father is 1st generation from Moravian/Czech, but it wasn't there. Instead, Scandinavian shows up. I just sent off a sample to FTDNA where very close relatives of my father have used to see if this is a fluke of not. This isn't TOO unexpected as I always questioned it. I am the only one in my family with curly hair (though my genes all say I should have straight hair), I tan very easily and my skin is a bit more olive (except my lips, to which I always have a slather of lip balm on). Again the SNP's say I have more markers for very pale, easily burned skin (with my other two siblings have). I was expecting Greek or some Italian, Scandinavian was a surprise.

So I run the results through GedMatch Genesis, remember V5 chip with the Eurogenes project as I learned on this website, that it can help sift out more broad categories.

Eurogenes Oracle 4 all pointed to the Polish, Ukranian and Russian aspects of maternal Grandma. Depending on what calculator you used, some would be more Ukranian, others Polish, others Russia. Eurogenes pointed out the big chunk of British DNA of my grandfather, and then put the rest under varying degrees of Norwegian or Danish. My closest squares is using the EUTest with:

1 Cornish + DK + EE + FR at 1.456880
2 DK + EE + English + FR at 1.475913
3 DK + EE + FR + NL at 1.481192
4 Cornish + Cornish + DK + UA at 1.561013
5 DK + DK + FR + PL at 1.671247
6 DK + FR + NO + PL at 1.681202


The k15 V2 Results are:

1 East_German + East_German + Norwegian + Norwegian at 4.181876
2 East_German + East_German + Norwegian + West_Norwegian at 4.220890
3 Norwegian + Norwegian + Southwest_English + Ukrainian at 4.225750
4 French + Norwegian + Ukrainian + West_Norwegian at 4.258610
5 Norwegian + Norwegian + Ukrainian + West_German at 4.290428

Number 4 here seems to be the most accurate citing both Ukraine and English.

CCC
02-02-2018, 12:42 AM
Hi all...

This is my first visit here, stumbling on it from a Google search '23andMe GEDmatch discrepancy'. This is a bit long, but bear with me. I want to be as thorough as possible.

My Y-DNA haplogroup is T, Middle/Near Eastern. My family always believed we are fully of Italian and Sicilian descent (born American). However, our surname is of Norman French origin as far as I can tell, not Italian, and is cognate with two forms of a common British surname. Our surname is concentrated in the Palermo Province region, the seat of the Norman kingdom in southern Italy and Sicily. It is sparse on the Italian mainland, even in the south of the boot. My Countries of Ancestry file from 23andMe shows a good number of UK, Sweden, Norway, Ireland with the DNA matches' 4 grandparents from those countries, predominately UK. I have suspected from that circumstantial evidence that we have Norman ancestry.

My AC from 23andMe shows 70% Italian, but I think that's misleading and not very finely broken down.

70.3 Italian
10.8 Broadly Southern European
1.2 Broadly Northern European
3.2 Broadly European
12.9 Middle Eastern & North African

Eurogenes K15 shows (K13 is not much different):

North_Sea - 9.33
Atlantic - 16.77
Baltic - 2.81
Eastern_Euro - 3.71
West_Med - 18.09
West_Asian - 12.72
East_Med - 29.52
Red_Sea - 4.64
Northeast_African - 0.24
Sub-Saharan - 2.17

The West Asian, East and West Mediterranean, Red Sea, African are no doubt the T haplogroup. The European numbers are probably my mtDNA HV4a haplogroup.

What I want to find, and have not yet been able to do so, is a finer breakdown of the regions of North Sea, Atlantic, and Baltic. Those are the areas where the Germanic tribes held sway: Norse, Anglo-Saxon, Norman, Franks, Vandals, etc. Does anyone know how to get a finer breakdown, and why there's such a discrepancy between 23andMe and GEDmatch?




Could you contact me we have the same haplogroup

Jenny
02-02-2018, 07:41 AM
23andme was totally accurate for my known family tree if you discount that under 3% noise. Gedmatch results were far less accurate almost laughable

JerryS.
02-02-2018, 03:31 PM
23andme was totally accurate for my known family tree if you discount that under 3% noise. Gedmatch results were far less accurate almost laughable

which models were you using on GEDmatch?

greerpalmer
02-02-2018, 09:55 PM
My faith in 23andme has lessened since the launch of V5's and their inability to phase correctly with older kits. I'm all for more "conservative' aka more "accurate" results, but when you receive an ancestry composition with a set of ethnic groups based on your kits being phased with your parents, but your parents do not receive those same groups, it makes me question why I even tested my parents in the first place...

JerryS.
02-02-2018, 10:09 PM
My faith in 23andme has lessened since the launch of V5's and their inability to phase correctly with older kits. I'm all for more "conservative' aka more "accurate" results, but when you receive an ancestry composition with a set of ethnic groups based on your kits being phased with your parents, but your parents do not receive those same groups, it makes me question why I even tested my parents in the first place...

do not rely on the for profit commercial companies to give you an accurate guestimate. simply take the raw data, load it to GEDmatch and run it through the various calculators and sub-models to see what regions/countries regularly show up in similar percentages.

Censored
02-24-2018, 09:33 PM
I don't know honestly. Some companies show me to be totally eastern/southern shifted like WeGene and MyHeritage but 23andme and ftDNA show a western shift as well with Middle Eastern and African(in 23andme's case) which aligns way more closely with what I know of my family so I think as far as major companies go, 23andme is probably the best out of all the ones I tried. Never used Ancestry btw.

However, on a more regional scale Gedmatch is better. They would definitely struggle with minor ancestries though and the dozens of different calculators is another headache.

Kale
03-05-2018, 03:12 AM
I just got my 23andme V5 results and ran them through the Gedmatch Genesis calculators (All of MDLP, Eurogenes, and Dodecad). Not a single one was able to capture known ancestry in the mix-oracles. 23andme actually nailed it spot on, though completely by coincidence.

Flint
05-07-2018, 10:57 PM
Some people say that 23andme predicts ancestry from the last 500 years, and that it is therefore more reliable than GEDmatch.

I don't really agree because I've noticed very peculiar things happen with the 23andme calculator. I've seen families where one parent is Sicilian and the other is Irish. The Sicilian parent scores 15% Middle Eastern, while the child scores 0%. I've yet to see a half North European, half SE European (Cretan, Sicilian, Greek islander) score significant Middle Eastern even though full members of said groups score a significant amount.

Likewise, I suspect 23andme underestimates admixture when it is small. I have seen that in people with under 10% of African, the African is greatly underestimated on 23andme.

What do others think? I think due to its algorithms, 23andme ends up giving misleading results, but it is good for telling if a given ancestral component exists... just not necessarily the amounts.
Hey Sikeliot, how are you doing?:P
I left your site after Loki attempted to bully me with negative reps.

As far as 23andMe ancestry prediction estimates, I believe they're garbage and give people skewed ideas on the origins of most ethnic groups.
Ethnicity isn't tied to geographic location over the past 5 centuries.
Ethnicity is tied to the racial/ethnic groups that blend to form your specific genetic composition.
By 23andMe's garbage methodology, if Pygmies or Aboriginal Australians lived in Europe for 500 years they would be counted as "100% European" on 23andMe.
23andMe and their methodology just helps spread rank ignorance, and at best a handful of 23andMe users even bother to go to sites like GEDmatch to unearth a semblance of sanity about their actual ethnic makeup.

Leto
05-08-2018, 12:21 AM
23andMe and their methodology just helps spread rank ignorance, and at best a handful of 23andMe users even bother to go to sites like GEDmatch to unearth a semblance of sanity about their actual ethnic makeup.
This is very true. It's enough to watch some of the videos on YouTube where people make stupid emoji-inspired faces saying ridiculous things like "I'm not 100% Mexican or British!!!" once they've seen their 23andme AC (30% French & German, 10% Iberian and all that jazz). I think to newbies these tests may bring a lot of confusion.

Flint
05-08-2018, 12:26 AM
This is very true. It's enough to watch some of the videos on YouTube where people make stupid emoji-inspired faces saying ridiculous things like "I'm not 100% Mexican or British!!!" once they've seen their 23andme AC (30% French & German, 10% Iberian and all that jazz). I think to newbies these tests may bring a lot of confusion.
Howdy Leto!:)
How are things at the Apricity?
I got sick of that zoo...
Is Bobby still a member, or has Loki ban-hammered him?

Leto
05-08-2018, 12:30 AM
Howdy Leto!:)
How are things at the Apricity?
I got sick of that zoo...
Is Bobby still a member, or has Loki ban-hammered him?
Do I know you? :confused:
Yeah, it's mostly a shitshow, bar the genetic section which I stick to.
Bobby is still around. Loki usually bans the wrong people.

Flint
05-08-2018, 12:47 AM
Do I know you? :confused:
Yeah, it's mostly a shitshow, bar the genetic section which I stick to.
Bobby is still around. Loki usually bans the wrong people.
We know each other.
We had minor disagreements, but I always considered you a knowledgeable poster.
Loki is a mental midget, and had banned many good posters.
The Apricity is falling apart.
Science based population genetics is maybe 15% of the site, without Sikeliot the Apricity would go under.

msmarjoribanks
05-08-2018, 12:48 AM
23 and Me was reasonably accurate for me, but I'm pretty skeptical about how much detail is really possible when you are trying to determine ancestry based on existing countries of origin, especially for an American (like me) with mixed ancestry from places with similar (or in the case of England especially, similarly mixed) ancestry.

So that I got over 31% broadly NW Europe seemed reasonable. 56% Irish and British (attributed more to British) is accurate (I'm around 50% or more English, plus some Welsh (could be included in the English too), some Irish and Scottish (could be in the English too). Similarly 9.5% French and German (attributed to Germany) is reasonable -- I am a bit French and definitely some German, the number might be a bit high, but could include some Swedish. The miss was 1% Scandinavian when I'm about 12.5%, but that's likely in the general NW category (if not the English and German).

The Gedmatch calculators seem to me to be doing something else entirely -- not trying to predict percentages of recent ancestry, but components within whatever ancestry you have, so comparing the two makes no sense to me.

For example, in Eurogenes K13, I have 7% West Asian and over 12% West Med. Does this mean that 23 and Me (and everyone else) missed that I have immigrant ancestors from Turkey (or the like) and Spain or Southern France? Probably not, and interestingly if you compare with, say, the average person from SW England, you get 15% West Med and 5% West Asian, so clearly there are other explanations for those results than assuming they actually came from those regions anytime in the recent past (or even genealogical time).

Thracian88
05-09-2018, 11:28 AM
Updated 23andMe is much better than Gedmatch.

JerryS.
05-09-2018, 12:11 PM
Updated 23andMe is much better than Gedmatch.

GEDmatch is just a web site. which calculator model(s) are you saying 23 and me is better than?

Flint
05-09-2018, 12:37 PM
Updated 23andMe is much better than Gedmatch.
Please explain and show the updated data that 23andMe gave you?

I closed my 23andMe account last week, out of absolute frustration with their crude and non informative ethnic assignments.
Geographic populations living in Europe for centuries like full Ashkenazi Jews get "100% European" on 23andMe and often score little to NONE middle eastern/southern European or north African.
That is an insane way to measure ethnicity in a population that is at or near hybridized with Levantine/southern European ancestry.
Most people on 23andMe aren't savvy with anthro forums and just take their results at face value, this just spreads ignorance and harms the science of population genetics.
Take this Ashkenazi result from 23andMe for example.
A hybridized middle easterner/Euro is classified as 99.9% European???
WTF is that?
Genetically that is garbage.
98% "Ashkenazi Jewish" and this ethnic group that is strongly influenced by Greco-Roman ancestry, is only scoring only 0.3% in southern European and broadly southern European.:blah:
https://qph.ec.quoracdn.net/main-qimg-1f929c3f9a1d145da1d7519674daef9c-c

Thracian88
05-09-2018, 12:51 PM
GEDmatch is just a web site. which calculator model(s) are you saying 23 and me is better than?

Well. I wasn't talking about a specific calculator. I mean I got much better and reliable scores in 23andMe than almost all calculators in Gedmatch. I really like Gedmatch as well but in my case 23andMe is more accurate.

JerryS.
05-09-2018, 12:55 PM
Well. I wasn't talking about a specific calculator. I mean I got much better and reliable scores in 23andMe than almost all calculators in Gedmatch. I really like Gedmatch as well but in my case 23andMe is more accurate.

im going to guess just by your screen name you get groups from Turkey, Greece, Bulgaria, Romania?

Thracian88
05-09-2018, 01:08 PM
Please explain and show the updated data that 23andMe gave you.

I closed my 23andMe account last week, out of absolute frustration with their crude and non informative ethnic assignments.
Geographic populations living in Europe for centuries like full Ashkenazi Jews get "100% European" on 23andMe and often score little to NONE middle eastern/southern European or north African.
That is an insane way to measure ethnicity in a population that is at or near hybridized with Levantine/southern European ancestry.
Most people on 23andMe aren't savvy with anthro forums and just take their results at face value, this just spreads ignorance and harms the science of population genetics.
Take this Ashkenazi result from 23andMe for example.
A hybridized middle easterner/Euro is classified as 99.9% European???
WTF is that?
Genetically that is garbage.
98% "Ashkenazi Jewish" and this ethnic group that is strongly influenced by Greco-Roman ancestry, is only scoring 0.3% in southern European and broadly southern European.
https://qph.ec.quoracdn.net/main-qimg-1f929c3f9a1d145da1d7519674daef9c-c

Sure, I can explain.

5 dots for Bulgaria
1 dot for Greece, Romania and Turkey

6/8 of my grandparents are from Bulgaria. 2 of them are from Trnova (Romania is just 1 hour from Trnova). I got 5 dots for Bulgaria and 1 dot for Romania are pretty accurate.
1/8 of my grandparents are from Gevgelije, Macedonia(it is located Greek border, it was part of Thessaloniki Eyalet in Ottomans). 1 dot for Greece is accurate as well.
1/8 is most likely from Turkey. And 1 dot for Turkey is accurate for me.

Both single population and mixed modes are not accurate for me in Gedmatch.

Thracian88
05-09-2018, 01:13 PM
im going to guess just by your screen name you get groups from Turkey, Greece, Bulgaria, Romania?

Yes, I do. Somehow I get weird groups as well such as Scandinavians, British Isles, Jews, Persians etc..

Flint
05-09-2018, 01:19 PM
Sure, I can explain.

5 dots for Bulgaria
1 dot for Greece, Romania and Turkey

6/8 of my grandparents are from Bulgaria. 2 of them are from Trnova (Romania is just 1 hour from Trnova). I got 5 dots for Bulgaria and 1 dot for Romania is pretty accurate.
1/8 of my grandparents are from Gevgelije, Macedonia(it is located Greek border, it was part of Thessaloniki Eyalet in Ottomans). 1 dot for Greece is accurate as well.
1/8 is most likely from Turkey. And 1 dot for Turkey is accurate for me.

Both single population and mixed modes are not accurate for me in Gedmatch.
Ok.
Thank you for your explanation.
This "updated info" was already available to me when I closed my account.
It still changes little to nothing about the issue I posted.
Again, I will use Ashkenazi Jews as an example.
None of the expanded populations will give them hits and elucidate any of the ancestry that is in their actual ethnic make-up.
AJ Jews aren't scoring even 1 dot on any of the actual populations that interweave to create their ethnic group.
23andMe is Youtube level in IQ when it comes to assigning ethnic data.
In fact the miscreants on Youtube use 23andMe's bastardized data to spread the most dreadful ignorance worldwide.
How many dots do Southern Italians/Sicilians get of middle eastern or north African on this brave new expanded ethnicity estimate?
I would venture to say NONE...

JerryS.
05-09-2018, 01:24 PM
Yes, I do. Somehow I get weird groups as well such as Scandinavians, British Isles, Jews, Persians etc..

Jews and Persia are not unusual to see for that region when you force a four population group to be shown. The UK and Viking stuff is a little strange though.

Flint
05-09-2018, 01:43 PM
I would also like to point out the 23andMe's Disqus blog has "moderation," and when I posted my criticism in a VERY concise AND respectful way, my posts were never allowed on.
2 months now and my posts are "still awaiting moderation."
In that 2 month period dozens of posts that are spewing imbecile level misinformation, were allowed on their blog....:P

JerryS.
05-09-2018, 02:00 PM
I would also like to point out the 23andMe's Disqus blog has "moderation," and when I posted my criticism in a VERY concise AND respectful way, my posts were never allowed on.
2 months now and my posts are "still awaiting moderation."
In that 2 month period dozens of posts that are spewing imbecile level misinformation, were allowed on their blog....:P

some people have delicate egos, some companies are agenda driven.... and neither allow contrary talk even if it is fact finding in nature. I am the first to say I am a layman and will ask questions that might seem sarcastic but I really am just trying to gain clarification and a better understanding by eliminating a lot of nerd jargon so I can have something explained in the simplest terms; that doesn't sit well with some.

a prime example is when I first joined this forum I thought the Turks had a significant influence on Italian DNA because of the spread of the Ottoman Empire was similar to the Moors. I argued to a point but eventually conceded because I was coming from a view of what seemed plausible from a layman. I thought this because I got Turkish groups and Balkan groups mixed in with Italian on some calculators. I later learned that it was because some models are slanted certain ways and favor certain regions over others....also that having mixed regions as I do can confuse some calculator models..... I learned, I hoped I didn't hurt anyone's feelings, I'm glad I didn't get banned.

I had to hold my tongue and do some reading to figure out this Ashkenazi thing that replaces South Italian. LOL.

msmarjoribanks
05-09-2018, 05:38 PM
Ok.
Thank you for your explanation.
This "updated info" was already available to me when I closed my account.
It still changes little to nothing about the issue I posted.
Again, I will use Ashkenazi Jews as an example.
None of the expanded populations will give them hits and elucidate any of the ancestry that is in their actual ethnic make-up.
AJ Jews aren't scoring even 1 dot on any of the actual populations that interweave to create their ethnic group.
23andMe is Youtube level in IQ when it comes to assigning ethnic data.
In fact the miscreants on Youtube use 23andMe's bastardized data to spread the most dreadful ignorance worldwide.
How many dots do Southern Italians/Sicilians get of middle eastern or north African on this brave new expanded ethnicity estimate?
I would venture to say NONE...

23 and Me is telling people what their more recent ancestry is. That test tells the person he is 98% Ashkenazi, it's not missing the AJ at all.

It's like telling someone they are British/Irish rather than presenting what would be a confusing mishmash of various underlying components. If someone wants to know their ancestry (estimated) within genealogical time, they typically want to know AJ or English or Irish, not "Scandinavian + German + Celtic + a little Western Med" or whatever which might well be confusing.

That's why 23 and Me and Gedmatch are good supplements to each others; they do different things.

(My interest was in family history, since I am and remain skeptical about the ability to identify specific countries of ancestral origin and already knew I was basically all NW European. For that, for an American, Ancestry is probably best even though I believe 23 and Me is probably better with the ethnicity predictions. My main interest with the ethnic origin calculator was that I was curious to see if there'd be a surprise, and there hasn't been.)

ianz91
05-09-2018, 07:35 PM
Please explain and show the updated data that 23andMe gave you?

I closed my 23andMe account last week, out of absolute frustration with their crude and non informative ethnic assignments.
Geographic populations living in Europe for centuries like full Ashkenazi Jews get "100% European" on 23andMe and often score little to NONE middle eastern/southern European or north African.
That is an insane way to measure ethnicity in a population that is at or near hybridized with Levantine/southern European ancestry.
Most people on 23andMe aren't savvy with anthro forums and just take their results at face value, this just spreads ignorance and harms the science of population genetics.
Take this Ashkenazi result from 23andMe for example.
A hybridized middle easterner/Euro is classified as 99.9% European???
WTF is that?
Genetically that is garbage.
98% "Ashkenazi Jewish" and this ethnic group that is strongly influenced by Greco-Roman ancestry, is only scoring only 0.3% in southern European and broadly southern European.:blah:
https://qph.ec.quoracdn.net/main-qimg-1f929c3f9a1d145da1d7519674daef9c-c


I have to admit, I don't like how these testing companies have Jewish ancestry put into the "European" category, it doesn't make sense. I think Jewish people should have their own category, at least FTDNA does this with their "Jewish Diaspora".

ianz91
05-09-2018, 07:39 PM
Ok.
Thank you for your explanation.
This "updated info" was already available to me when I closed my account.
It still changes little to nothing about the issue I posted.
Again, I will use Ashkenazi Jews as an example.
None of the expanded populations will give them hits and elucidate any of the ancestry that is in their actual ethnic make-up.
AJ Jews aren't scoring even 1 dot on any of the actual populations that interweave to create their ethnic group.
23andMe is Youtube level in IQ when it comes to assigning ethnic data.
In fact the miscreants on Youtube use 23andMe's bastardized data to spread the most dreadful ignorance worldwide.
How many dots do Southern Italians/Sicilians get of middle eastern or north African on this brave new expanded ethnicity estimate?
I would venture to say NONE...


You have to keep in mind that 23andme is geared towards people who want to know their health information, Ancestry is just an add on and not their main focus.

msmarjoribanks
05-10-2018, 11:22 AM
I have to admit, I don't like how these testing companies have Jewish ancestry put into the "European" category, it doesn't make sense. I think Jewish people should have their own category, at least FTDNA does this with their "Jewish Diaspora".

I think AJ makes sense as a European category -- AJs are either in Europe or descendants of people who immigrated from Europe. Yes, if you go back far enough you leave Europe (although there's still debate about how much European admix there is, for example that article arguing that some of the most common mtDNA lines for AJs are of European origin), but that's the same for everyone to some degree, so it becomes a debate about where the relevant cutoff date is, and with a population that's been in Europe for 100s of years and is genetically separate from Jewish populations who stayed in the Middle East, then I think it makes sense.

Bigger issue, what difference does it make if the ancestry is identified as AJ, not vague unspecified European? I just don't get what the problem is supposed to be.

Flint
05-10-2018, 12:59 PM
I think AJ makes sense as a European category -- AJs are either in Europe or descendants of people who immigrated from Europe. Yes, if you go back far enough you leave Europe (although there's still debate about how much European admix there is, for example that article arguing that some of the most common mtDNA lines for AJs are of European origin), but that's the same for everyone to some degree, so it becomes a debate about where the relevant cutoff date is, and with a population that's been in Europe for 100s of years and is genetically separate from Jewish populations who stayed in the Middle East, then I think it makes sense.

Bigger issue, what difference does it make if the ancestry is identified as AJ, not vague unspecified European? I just don't get what the problem is supposed to be.
Placing Ashkenazi Jews in the 100% European category makes no sense genetically; nor does claiming that AJ have little to often NO southern European/middle eastern.
All of that is science fiction and not part of the science of population genetics.
Spreading complete ignorance of the ethnic origins of AJ is political in nature, and not the job of an autosomal DNA testing company.

Flint
05-10-2018, 01:03 PM
You have to keep in mind that 23andme is geared towards people who want to know their health information, Ancestry is just an add on and not their main focus.
People are paying for separate health results, and the ancestry only results are on discussion here.
Not sure how your post is relevant to this specific discussion of 23andMe's methodology of reporting a hybridized middle eastern/European group, as "100% European."
How can 23andMe also validly claim that Ashkenazi Jews have little to no ancestry in the southern European or Middle eastern category?
This has nothing at all to do with health testing.

Flint
05-10-2018, 01:12 PM
23 and Me is telling people what their more recent ancestry is. That test tells the person he is 98% Ashkenazi, it's not missing the AJ at all.

It's like telling someone they are British/Irish rather than presenting what would be a confusing mishmash of various underlying components. If someone wants to know their ancestry (estimated) within genealogical time, they typically want to know AJ or English or Irish, not "Scandinavian + German + Celtic + a little Western Med" or whatever which might well be confusing.

That's why 23 and Me and Gedmatch are good supplements to each others; they do different things.

(My interest was in family history, since I am and remain skeptical about the ability to identify specific countries of ancestral origin and already knew I was basically all NW European. For that, for an American, Ancestry is probably best even though I believe 23 and Me is probably better with the ethnicity predictions. My main interest with the ethnic origin calculator was that I was curious to see if there'd be a surprise, and there hasn't been.)




23 and Me is telling people what their more recent ancestry is. That test tells the person he is 98% Ashkenazi, it's not missing the AJ at all.
You completely failed to comprehend my point, that I clearly stated in multiple posts.
Nobody is claiming that 23andMe is "missing" Ashkenazi ancestry.
23andMe is grossly MISINTERPRETING the ethnic origins of Ashkenazi ancestry...

The issues are that 23andMe defines Ashkenazi Jews as "100% European" and this is supposed to be a GENETIC testing site and not a NYT opinion piece.
Also, absurdly so, 23andMe assigns basically ZERO to minuscule amounts of southern European/Middle eastern/North African to Ashkenazi Jews.


It's like telling someone they are British/Irish rather than presenting what would be a confusing mishmash of various underlying component.
Sorry, but giving people a totally misleading and scientifically absurd genetic result of their ethnic make-up borders on fraud.

Flint
05-10-2018, 02:18 PM
I have to admit, I don't like how these testing companies have Jewish ancestry put into the "European" category, it doesn't make sense. I think Jewish people should have their own category, at least FTDNA does this with their "Jewish Diaspora".
Not even so much about the 100% European category, as much as failing to correctly put in their percentages of southern European at the very least.
Many Ashkenazi Jews get minuscule to no southern euro in their results, that is a fraudulent methodology.
No middle eastern?
That is also a fraudulent methodology.

Flint
05-10-2018, 02:35 PM
Here is the direct results of 23andMe's and many other companies flawed methodology.
This video alone has over 18,000 views and there are many more like it, and websites and blogs, too.

ASHKENAZI DNA is 99% EUROPEAN with no trace of Middle East

http://www.youtube.com/watch?v=7FMBtZiYysI
Published on Oct 8, 2017


The Ashkenazi are one of the purest of any European groups at 99% European with no trace of Middle East ancestry. That is very surprising for a people who claim to be descended from Abraham, Isaac, Jacob. The entire claim to Palestine is based on a fictitious ancestry link to the region that DNA has now settled. The Asheknazi on average score 96% in what they call "Ashkenazi" or "European Jewish" that traces their ancestry to Poland, Hungry, Austria, Lithuania ...

msmarjoribanks
05-10-2018, 03:01 PM
Placing Ashkenazi Jews in the 100% European category makes no sense genetically; nor does claiming that AJ have little to often NO southern European/middle eastern.
All of that is science fiction and not part of the science of population genetics.
Spreading complete ignorance of the ethnic origins of AJ is political in nature, and not the job of an autosomal DNA testing company.

But again the way that the major testing companies work is to identify the basic ethnicity/country of origin and not the components underlying it. Someone is not looking to see the make up of AJ (or English or Swedish or whatever), but to know that they ARE, say, 50% AJ, 50% whatever else. Gedmatch will show that someone from SE England is likely to have 5% Western Asian (as that component is called in K13), but I wouldn't expect 23 & Me to identify someone who was from SE England as 5% West Asian, but 100% English/Irish (with 5 stars for England) -- if that was how they tested, of course.

I get about 7% West Asian on Gedmatch, but I don't think it's wrong that I don't get that on Ancestry or the like. I get somewhere between 0 and 2%, instead.

msmarjoribanks
05-10-2018, 03:06 PM
Where is that video from? You aren't claiming it is 23 and Me, are you? I can't imagine they'd use such stupid language as "purest." I didn't watch it, but just the beginning distorts how the tests and results even work.

Archimedes
05-10-2018, 03:13 PM
Here is the direct results of 23andMe's and many other companies flawed methodology.
This video alone has over 18,000 views and there are many more like it, and websites and blogs, too.

ASHKENAZI DNA is 99% EUROPEAN with no trace of Middle East

http://www.youtube.com/watch?v=7FMBtZiYysI
Published on Oct 8, 2017

That video shows Kushner's results as being almost 100% Ashkenazi... Seems very accurate. Yes, Ashkenazi is listed as a sub ethnicity of Europe, which it is

Ashkenazim do have some roots in the middle east, but that ancestry is not recent. Ashkenazim are most similar to ..... Ashkenazim, which happens to be a unique population, which is located in......Europe. I don't really see the issue. Ashkenazim are a mix, but this mix was created in Europe, and deserves to be classed as a unique sub ethnicity of Europe. We can disagree, but the methodology is reasonable

Flint
05-10-2018, 03:14 PM
But again the way that the major testing companies work is to identify the basic ethnicity/country of origin and not the components underlying it. Someone is not looking to see the make up of AJ (or English or Swedish or whatever), but to know that they ARE 50% AJ. Gedmatch will show that someone from SE England is likely to have 5% Western Asian (as that component is called in K13), but I wouldn't expect 23 & Me to identify someone who was from SE England as 5% West Asian, but 100% English/Irish (with 5 stars for England) -- if that was how they tested, of course.

I get about 7% West Asian on Gedmatch, but I don't think it's wrong that I don't get that on Ancestry or the like. I get somewhere between 0 and 2%, instead.
You just highlighted a major problem that needs to be addressed.
Can't see how they're are doing a legit service, when a group like Ashkenazi Jews currently reside in the nation of Israel.
By calling them "100% European" sites like 23andMe are fueling rampant antisemitism and spreading fodder for extremists to delegitimatize their claim of Israel as their ancestral homeland.

Someone is not looking to see the make up of AJ (or English or Swedish or whatever)
Speak for yourself, many people have serious issues with the data methodology from 23andMe.
A crude ethnic designation like "Ashkenazi Jewish" in conjunction with NO south euro or middle eastern component is next to worthless.
GEDmatch is FREE and a BILLION dollar company like 23andMe can't even offer a similar set of tools?
And this is acceptable to you?

msmarjoribanks
05-10-2018, 03:20 PM
Average AJ result from Eurogenes K13:

N Atl 15.32
Baltic 10.16
W Med 18.44
W Asian 10.29
E Med 34.61
Red Sea 5.98
Rest are small percentages so excluded

Average Italian-Abruzzo
N Atl 22.21
Baltic 8.65
W Med 20.33
W Asian 15.04
E Med 27.45
Red Sea 4.90

For either group, if you tell them they have recent ancestry from, say, the Middle East or Asia (or the North Atlantic, for that matter) it's confusing and misleading. If you tell them they are Italian or AJ, they can explore further what the components that let to that estimate are.

Anyone sensible knows what AJ means and can explore the history, the diaspora, so on.

msmarjoribanks
05-10-2018, 03:32 PM
A crude ethnic designation like "Ashkenazi Jewish" in conjunction with NO south euro or middle eastern component is next to worthless.
GEDmatch is FREE and a BILLION dollar company like 23andMe can't even offer a similar set of tools?
And this is acceptable to you?

Again, you ignore that people generally use 23 and Me and the like to determine their own ethnic/ancestral origin, not deep ancestry. Gedmatch is doing something different. (If someone has no clue of ancestry and discovers they are 25% AJ, that's way more informative than just telling them components if they are using the test for genealogy.)

I think it's good that both exist, but even with the little "noise" percentages like, say, 2% West Asian from 23 and Me or Ancestry some will think that means they have some identifiable ancestor from Turkey or whatever, and that's not correct, that's not what it means in most cases. Breaking it down into the components rather than telling someone Italian or AJ would not be what most who use the service are seeking, IMO (and apparently in the opinion of those who run the service). Those who ALSO (or solely) are interested in deep ancestry can use Gedmatch, and frankly all of them are extremely imperfect, but complaining that 23 and Me doesn't do something it doesn't claim to do or set out to do seems to me unfair.

That the category you place AJ under seems politicized and likely to be used in an unpleasant and misleading way wherever it is placed is unfortunate, but not 23 and Me's fault, and suggesting that AJs are/were not real Europeans strikes me as potentially problematic also (especially if extended to various other groups with substantial admixes). It's the worst thing about these kinds of discussions.

Flint
05-10-2018, 03:36 PM
Average AJ result from Eurogenes K13:

N Atl 15.32
Baltic 10.16
W Med 18.44
W Asian 10.29
E Med 34.61
Red Sea 5.98
Rest are small percentages so excluded

Average Italian-Abruzzo
N Atl 22.21
Baltic 8.65
W Med 20.33
W Asian 15.04
E Med 27.45
Red Sea 4.90

For either group, if you tell them they have recent ancestry from, say, the Middle East or Asia (or the North Atlantic, for that matter) it's confusing and misleading. If you tell them they are Italian or AJ, they can explore further what the components that let to that estimate are.

Anyone sensible knows what AJ means and can explore the history, the diaspora, so on.
You have yet again failed to grasp the specific implications of calling Ashkenazi Jews "100% European."
This particular group is a massive target of bigotry and a systemic effort to delegitimize their presence in the state of Israel.

Posting Italian-Abruzzo comparisons with Ashkenazi Jews just buttresses my argument that 23andMe leaving out southern euro and Italian in their ancestry is laughable.
Tell me how the hell can 23andMe say Ashkenazi Jews have 0% to minuscule Italian, or even for that matter no southern European ancestry?

BTW, Abruzzo Italians are NOT the closest average match to AJ from Italy...

Flint
05-10-2018, 03:42 PM
Again, you ignore that people generally use 23 and Me and the like to determine their own ethnic/ancestral origin, not deep ancestry. Gedmatch is doing something different. (If someone has no clue of ancestry and discovers they are 25% AJ, that's way more informative than just telling them components if they are using the test for genealogy.)

I think it's good that both exist, but even with the little "noise" percentages like, say, 2% West Asian from 23 and Me or Ancestry some will think that means they have some identifiable ancestor from Turkey or whatever, and that's not correct, that's not what it means in most cases. Breaking it down into the components rather than telling someone Italian or AJ would not be what most who use the service are seeking, IMO (and apparently in the opinion of those who run the service). Those who ALSO (or solely) are interested in deep ancestry can use Gedmatch, and frankly all of them are extremely imperfect, but complaining that 23 and Me doesn't do something it doesn't claim to do or set out to do seems to me unfair.

That the category you place AJ under seems politicized and likely to be used in an unpleasant and misleading way wherever it is placed is unfortunate, but not 23 and Me's fault, and suggesting that AJs are/were not real Europeans strikes me as potentially problematic also (especially if extended to various other groups with substantial admixes). It's the worst thing about these kinds of discussions.
Worst post yet.
You clearly have no reading comprehension and have literally spun my post like a fake news monger from one of the failing networks.
So far on this site you're the most ignorant and failed poster, that I have attempted to engage with.
Thankfully there are many reasonable posters here.
I have tried to reason with you, and you have quite disrespectfully disregarded my posts entirely and farted out trollish replies.
I'm done with you.
Blocked.

Flint
05-10-2018, 03:56 PM
Having to explain to posters the folly of calling a HALF LEVANTINE group "100% European," in a normal world would be easy.......

Having to explain to posters the folly of calling a group that heavily overlaps with Sicilians and south Italians "0% southern European/Italian," in a normal world would be easy.....

Having to explain explain the folly of calling a group that is half middle eastern on average "0% middle eastern," in a normal world would be easy.

Anthro forums aren't a normal world.
I think I might just make a blog with a comments section, and help spread some reality on the web...

digital_noise
05-10-2018, 04:30 PM
Both 23 and Me and Ancestry DNA were accurate for me when it comes to northern, northwestern estimates but they both utterly failed with southern european/Italian for me. Not sure if the 25% south Italian genes I have confused them or if that DNA is so admixed they could only pluck out the Italian?? Who knows but My Heritage got it right as do most Gedmatch calcs. My Heritage kid of just lumps Scandinavian, western Europe into Great Britian for me though so I have to look at things objectively to make any reasonable sense out of it.

JerryS.
05-10-2018, 04:36 PM
while I am not as well versed as many here, even I can see the problem with categorizing AJ as 100% European, when in fact it is a mix of MENA and Southern European with a little Eastern European as far as I can tell (and that's over simplifying it).

JerryS.
05-10-2018, 04:40 PM
Both 23 and Me and Ancestry DNA were accurate for me when it comes to northern, northwestern estimates but they both utterly failed with southern european/Italian for me. Not sure if the 25% south Italian genes I have confused them or if that DNA is so admixed they could only pluck out the Italian?? Who knows but My Heritage got it right as do most Gedmatch calcs. My Heritage kid of just lumps Scandinavian, western Europe into Great Britian for me though so I have to look at things objectively to make any reasonable sense out of it.

I think the for profit DNA companies are geared more for the European-American populous which is mostly N./N.W. European based. because of this unless there is substantial (50% or more) Italian Greek they seem to be hit or miss many times on showing it or showing it quite low than it actually is.

digital_noise
05-10-2018, 04:42 PM
Regarding the categorization of Ashkenazi, what you all are saying makes sense when viewed from a DNA enthusiasts perspective. However, I'm pretty sure the average 23 and Me customer doesnt care that Ashkenazi is a mix. They just want to see a % next to a country that they recognize. Right or wrong, I cannot say but if they really broke Ashkenazi down to what it truly is, it would probably create more questions in the long run.

Leto
05-10-2018, 04:54 PM
Regarding the categorization of Ashkenazi, what you all are saying makes sense when viewed from a DNA enthusiasts perspective. However, I'm pretty sure the average 23 and Me customer doesnt care that Ashkenazi is a mix. They just want to see a % next to a country that they recognize. Right or wrong, I cannot say but if they really broke Ashkenazi down to what it truly is, it would probably create more questions in the long run.
Many people don't even know that Jewish is not just a religion. They think they are white people that practice the Jewish religion. That's the case at least in North America, definitely not in Eastern Europe.

I personally think that a geographica breakdown is much better than a list of countries. It's better to say Western Europe or East Asia than France & Germany and Korea, because a lot of newbies and ignorant people would automatically assume they have actual ancestry from those countries and therefore aren't purely this or that. That does bring a lot of confusion. These are the results of a Spanish woman (I think Eastern Spain or something like that):
https://i.imgur.com/KbrvYJk.png
A layman would go like 'wow, she's only half Spanish and 1/5 Italian! So mixed, after all we're all a mix of all sorts of things, all are originally from Africa" :lol: But the truth is that she might have no recent ancestry from outside Iberia.

Flint
05-10-2018, 05:24 PM
Many people don't even know that Jewish is not just a religion. They think they are white people that practice the Jewish religion. That's the case at least in North America, definitely not in Eastern Europe.
.
That is true, and brings in the whole question of whiteness and the extent that misleading genetic results can play in it.

I used Ashkenazi Jews as a generic example.
GENETICALLY you can make the same exact case with Sicilians and Greek Islanders, and mainland Greece goes from a Cline of heavy non euro admix in the Peloponnese, to a more mildly admixed level in Macedonia.
Do I think Ashkenazi Jews or Sicilians are "white?"
I consider both of them them "white," but as far as strictly on a genetic level, nether are Europeans.
23andMe has no data for "cultural whiteness" or religion, and is supposed to be STRICTLY testing and crunching genetic data.
I need to be crystal clear that I was only using Ashkenazi Jews as an example for deep southern euro populations, and how 23andMe gets them all wrong.
Eastern euros and Slavs in particular have a long history of antisemitism, and I don't subscribe to their ideology, I'm strictly talking the science of population genetics.

We can discuss the concept of cultural whiteness, but on this thread that would be off topic. :P

Ruderico
05-10-2018, 05:40 PM
I consider both of them them "white," but as far as strictly on a genetic level, neither are Europeans.
23andMe has no data for "cultural whiteness" or religion, and is supposed to be STRICTLY testing and crunching genetic data.

Silly things like "white" aside (or "cultural whiteness" which I've never even heard of, maybe it's an American thing) what is a European genetically? Please define it because I'm having a real hard time seeing Sicilians (or South Italians) as non-European just because they are not exactly your cookie-cutter Central/North European

Flint
05-10-2018, 05:50 PM
Silly things like "white" aside (or "cultural whiteness" which I've never even heard of, maybe it's an American thing) what is a European genetically? Please define it because I'm having a real hard time seeing Sicilians (or South Italians) as non-European just because they are not exactly your cookie-cutter Central/North European
First off Ruderico, I put white and cultural whiteness in quotes for a reason.
Clearly I consider both to be shady and not valid concepts.
I'm sorry that went over your head.


what is a European genetically?
That is the question we're debating on this thread.
Sicilians can be modeled accurately as half Lebanese Christian and half northern Italian.
Is that genetic admix a "European?"
You tell me...

msmarjoribanks
05-10-2018, 05:51 PM
You have yet again failed to grasp the specific implications of calling Ashkenazi Jews "100% European."
This particular group is a massive target of bigotry and a systemic effort to delegitimize their presence in the state of Israel.

Posting Italian-Abruzzo comparisons with Ashkenazi Jews just buttresses my argument that 23andMe leaving out southern euro and Italian in their ancestry is laughable.
Tell me how the hell can 23andMe say Ashkenazi Jews have 0% to minuscule Italian, or even for that matter no southern European ancestry?

BTW, Abruzzo Italians are NOT the closest average match to AJ from Italy...

Who said they were?

Again, you (1) fail to note what 23 and Me is doing (AJ is a distinct group, there is no need to go into the underlying ancestry); and (2) ignore that the label could and would be used in a bigoted way if they defined them as non European too, thereby erasing years of history. 23 and Me can't win there.

Flint
05-10-2018, 06:08 PM
) what is a European genetically?
Here are mixed mode results for a Sicilian sample on GEDmatch.
You tell me Ruderico, if these combos= European genetically.

# Primary Population (source) Secondary Population (source) Distance
1 58.8% Spanish_Andalucia + 41.2% Lebanese_Druze @ 2.85
2 56.3% Spanish_Castilla_La_Mancha + 43.7% Lebanese_Druze @ 3
3 56% Lebanese_Druze + 44% French_Basque @ 3.04
4 60.9% Spanish_Murcia + 39.1% Kurdish_Jewish @ 3.28
5 59.7% Spanish_Valencia + 40.3% Kurdish_Jewish @ 3.33
6 50.1% Lebanese_Muslim + 49.9% Spanish_Aragon @ 3.34
7 53.6% Spanish_Valencia + 46.4% Lebanese_Muslim @ 3.38
8 53.7% Spanish_Aragon + 46.3% Lebanese_Druze @ 3.4
9 56.1% Spanish_Aragon + 43.9% Kurdish_Jewish @ 3.43
10 53.7% Spanish_Cantabria + 46.3% Lebanese_Druze @ 3.43
11 54.8% Spanish_Murcia + 45.2% Lebanese_Muslim @ 3.48
12 61.2% Spanish_Andalucia + 38.8% Kurdish_Jewish @ 3.52
13 55.2% Spanish_Andalucia + 44.8% Lebanese_Muslim @ 3.53
14 59.6% Spanish_Extremadura + 40.4% Lebanese_Druze @ 3.64
15 56.1% Spanish_Aragon + 43.9% Iranian_Jewish @ 3.64
16 60.2% Spanish_Murcia + 39.8% Assyrian @ 3.64
17 52.6% Spanish_Castilla_La_Mancha + 47.4% Lebanese_Muslim @ 3.66
18 61.2% Spanish_Andalucia + 38.8% Iranian_Jewish @ 3.67
19 60.9% Spanish_Murcia + 39.1% Iranian_Jewish @ 3.69
20 58.7% Spanish_Castilla_La_Mancha + 41.3% Kurdish_Jewish @ 3.72

Ruderico
05-10-2018, 06:12 PM
I'll ignore the first part of your reply because it's pointless


Sicilians can be modeled accurately as half Lebanese Christian and half northern Italian.
Is that genetic admix a "European?"
You tell me...

There is nothing special about Europe or its people, they have not been separated from the rest of the world to make its genetic profile clear and easy to define. The eastern Mediterranean has been a region with significant trade and contact between all the peoples that inhabited it for over 4000 years, it's only natural that there is a degree of overlap and admixture between them.

Modern populations being modelled with non-Euro admixture doesn't rule them out from being "genetically European". Sicilians live in Europe, and the overwhelming majority of their ancestors up to 2000 years ago did so too - it is where they were genetically formed. In fact, even populations of ancient SE Europe showed a similar autossomal profile to modern ones, they plot pretty close in a PCA (https://1.bp.blogspot.com/-sj4uKA5Cpb4/WZXFlsW7cPI/AAAAAAAAGZw/zdFzz4-9UvAWwFGUXAaN9dqYslrduYoPACLcBGAs/s1600/Minoans_%2526_Mycenaeans.png).

Are Finns European? Are Iberians European? Because all of them can be modelled with extra-European admixture, so its seems like a rather bad way to define such a thing. But that's to be expected, a "genetic European" is something that is kind of made up, particularly because all Europeans and their neighbours are West Eurasians, and have basically been formed with different proportions of the same peoples (with a few exceptions who have a little bit of exotic in them).


So to answer your question. Yes, Sicilians are European genetically, it's just that Europeans are not an homogenous and easy-to-define thing. Don't be so dogmatic

Flint
05-10-2018, 06:20 PM
I'll ignore the first part of your reply because it's pointless



There is nothing special about Europe or its people, they have not been separated from the rest of the world to make its genetic profile clear and easy to define. The eastern Mediterranean has been a region with significant trade and contact between all the peoples that inhabited it for over 4000 years, it's only natural that there is a degree of overlap and admixture between them.

Modern populations being modelled with non-Euro admixture doesn't rule them out from being "genetically European". Sicilians live in Europe, and the overwhelming majority of their ancestors up to 2000 years ago did so too - it is where they were genetically formed. In fact, even populations of ancient SE Europe showed a similar autossomal profile to modern ones, they plot pretty close in a PCA (https://1.bp.blogspot.com/-sj4uKA5Cpb4/WZXFlsW7cPI/AAAAAAAAGZw/zdFzz4-9UvAWwFGUXAaN9dqYslrduYoPACLcBGAs/s1600/Minoans_%2526_Mycenaeans.png).

Are Finns European? Are Iberians European? Because all of them can be modelled with extra-European admixture, so its seems like a rather bad way to define such a thing. But that's to be expected, a "genetic European" is something that is kind of made up, particularly because all Europeans and their neighbours are West Eurasians, and have basically been formed with different proportions of the same peoples (with a few exceptions who have a little bit of exotic in them).
Clearly you're ignoring the conversation on this thread and attempting to spin.
Calling a hybridized Levantine/southern European group "100% European" basically nullfies all the work of population genetics.
Attempting (poorly so), to relegate the established concept of European genetic groupings to the dust bin, is laughable and the height of trollish endeavors.


Are Iberian European? Because all of them can be modelled with extra-European admixture
Complete proof positive that you're a troll.
Iberians are OVERWHELMINGLY European genetically, and you have attempted and failed to conflate them in a discussion about HYBRIDIZED groups that are 50% Levantine.

Tz85
05-10-2018, 06:26 PM
I'll ignore the first part of your reply because it's pointless



There is nothing special about Europe or its people, they have not been separated from the rest of the world to make its genetic profile clear and easy to define. The eastern Mediterranean has been a region with significant trade and contact between all the peoples that inhabited it for over 4000 years, it's only natural that there is a degree of overlap and admixture between them.

Modern populations being modelled with non-Euro admixture doesn't rule them out from being "genetically European". Sicilians live in Europe, and the overwhelming majority of their ancestors up to 2000 years ago did so too - it is where they were genetically formed. In fact, even populations of ancient SE Europe showed a similar autossomal profile to modern ones, they plot pretty close in a PCA (https://1.bp.blogspot.com/-sj4uKA5Cpb4/WZXFlsW7cPI/AAAAAAAAGZw/zdFzz4-9UvAWwFGUXAaN9dqYslrduYoPACLcBGAs/s1600/Minoans_%2526_Mycenaeans.png).

Are Finns European? Are Iberians European? Because all of them can be modelled with extra-European admixture, so its seems like a rather bad way to define such a thing. But that's to be expected, a "genetic European" is something that is kind of made up, particularly because all Europeans and their neighbours are West Eurasians, and have basically been formed with different proportions of the same peoples (with a few exceptions who have a little bit of exotic in them).


So to answer your question. Yes, Sicilians are European genetically, it's just that Europeans are not an homogenous and easy-to-define thing. Don't be so dogmatic

Is this individual fully European??


Using 2 populations approximation:
1 50% Ashkenazi +50% Moldavian @ 6.243006


Using 3 populations approximation:
1 50% East_Sicilian +25% Greek_Thessaly +25% La_Brana-1 @ 5.156192


Using 4 populations approximation:
++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++ +++++++++++++
1 Algerian_Jewish + Bulgarian + East_Sicilian + La_Brana-1 @ 5.150026
2 East_Sicilian + East_Sicilian + Greek_Thessaly + La_Brana-1 @ 5.156192
3 Algerian_Jewish + Bulgarian + Central_Greek + La_Brana-1 @ 5.163017
4 Algerian_Jewish + Bulgarian + La_Brana-1 + South_Italian @ 5.187834
5 Central_Greek + East_Sicilian + Greek_Thessaly + La_Brana-1 @ 5.229041
6 Algerian_Jewish + Greek_Thessaly + Greek_Thessaly + La_Brana-1 @ 5.252636
7 Bulgarian + La_Brana-1 + South_Italian + South_Italian @ 5.281569
8 Algerian_Jewish + La_Brana-1 + Romanian + South_Italian @ 5.297127
9 Greek_Thessaly + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.302997
10 Bulgarian + East_Sicilian + La_Brana-1 + South_Italian @ 5.311954
11 East_Sicilian + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.313483
12 Algerian_Jewish + East_Sicilian + La_Brana-1 + Romanian @ 5.332044
13 Algerian_Jewish + Central_Greek + La_Brana-1 + Romanian @ 5.362298
14 Central_Greek + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.386769
15 Central_Greek + Central_Greek + Greek_Thessaly + La_Brana-1 @ 5.410214
16 Iranian_Jewish + La_Brana-1 + Moldavian + Sardinian @ 5.488004
17 Bulgarian + Central_Greek + La_Brana-1 + Libyan_Jewish @ 5.491062
18 East_Sicilian + Greek_Thessaly + Greek_Thessaly + La_Brana-1 @ 5.491326
19 La_Brana-1 + Libyan_Jewish + Moldavian + South_Italian @ 5.506254
20 Algerian_Jewish + La_Brana-1 + Moldavian + South_Italian @ 5.508245

Flint
05-10-2018, 06:35 PM
Is this individual fully European??


Using 2 populations approximation:
1 50% Ashkenazi +50% Moldavian @ 6.243006


Using 3 populations approximation:
1 50% East_Sicilian +25% Greek_Thessaly +25% La_Brana-1 @ 5.156192


Using 4 populations approximation:
++++++++++++++++++++++++++++++++++++++++++++++++++ ++++++++++++++++++++++++++++++++++++++++++++++++++ +++++++++++++
1 Algerian_Jewish + Bulgarian + East_Sicilian + La_Brana-1 @ 5.150026
2 East_Sicilian + East_Sicilian + Greek_Thessaly + La_Brana-1 @ 5.156192
3 Algerian_Jewish + Bulgarian + Central_Greek + La_Brana-1 @ 5.163017
4 Algerian_Jewish + Bulgarian + La_Brana-1 + South_Italian @ 5.187834
5 Central_Greek + East_Sicilian + Greek_Thessaly + La_Brana-1 @ 5.229041
6 Algerian_Jewish + Greek_Thessaly + Greek_Thessaly + La_Brana-1 @ 5.252636
7 Bulgarian + La_Brana-1 + South_Italian + South_Italian @ 5.281569
8 Algerian_Jewish + La_Brana-1 + Romanian + South_Italian @ 5.297127
9 Greek_Thessaly + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.302997
10 Bulgarian + East_Sicilian + La_Brana-1 + South_Italian @ 5.311954
11 East_Sicilian + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.313483
12 Algerian_Jewish + East_Sicilian + La_Brana-1 + Romanian @ 5.332044
13 Algerian_Jewish + Central_Greek + La_Brana-1 + Romanian @ 5.362298
14 Central_Greek + Greek_Thessaly + La_Brana-1 + South_Italian @ 5.386769
15 Central_Greek + Central_Greek + Greek_Thessaly + La_Brana-1 @ 5.410214
16 Iranian_Jewish + La_Brana-1 + Moldavian + Sardinian @ 5.488004
17 Bulgarian + Central_Greek + La_Brana-1 + Libyan_Jewish @ 5.491062
18 East_Sicilian + Greek_Thessaly + Greek_Thessaly + La_Brana-1 @ 5.491326
19 La_Brana-1 + Libyan_Jewish + Moldavian + South_Italian @ 5.506254
20 Algerian_Jewish + La_Brana-1 + Moldavian + South_Italian @ 5.508245

Good question.
Glad one poster is on subject and sincerely trying to have a conversation.
I will await to see the trolls reply to you.

Ruderico
05-10-2018, 06:36 PM
Clearly you're ignoring the conversation on this thread and attempting to spin.
Calling a hybridized Levantine/southern European group "100% European" basically nullfies all the work of population genetics.
Attempting (poorly so), to relegate the established concept of European genetic groupings to the dust bin, is laughable and the height of trollish endeavors.


Complete proof positive that you're a troll.
Iberians are OVERWHELMINGLY European genetically, and you have attempted and failed to conflate them in a discussion about HYBRIDIZED groups that are 50% Levantine.

Iberians, particularly western ones, have 10% real north African ancestry. Sicilians are not "hybridised" anymore than any other population is, we're all mixed, I'll bet that if you find samples from Sicily from 2000 years ago they will look rather similar to modern ones. You made no effort to deconstruct the argument, you just showed a crap 2pop model while ignoring the historical and genetic makeup that has been in place (in SE Europe) for thousands of years.

And you, new guy with a rude and aggressive posture, call others trolls. I don't generally do this, but kindly sod off back to TA. I hope I had banned you back there, so you'd have a reason to have that attitute

Leto
05-10-2018, 07:18 PM
I'll ignore the first part of your reply because it's pointless



There is nothing special about Europe or its people, they have not been separated from the rest of the world to make its genetic profile clear and easy to define. The eastern Mediterranean has been a region with significant trade and contact between all the peoples that inhabited it for over 4000 years, it's only natural that there is a degree of overlap and admixture between them.

Modern populations being modelled with non-Euro admixture doesn't rule them out from being "genetically European". Sicilians live in Europe, and the overwhelming majority of their ancestors up to 2000 years ago did so too - it is where they were genetically formed. In fact, even populations of ancient SE Europe showed a similar autossomal profile to modern ones, they plot pretty close in a PCA (https://1.bp.blogspot.com/-sj4uKA5Cpb4/WZXFlsW7cPI/AAAAAAAAGZw/zdFzz4-9UvAWwFGUXAaN9dqYslrduYoPACLcBGAs/s1600/Minoans_%2526_Mycenaeans.png).

Are Finns European? Are Iberians European? Because all of them can be modelled with extra-European admixture, so its seems like a rather bad way to define such a thing. But that's to be expected, a "genetic European" is something that is kind of made up, particularly because all Europeans and their neighbours are West Eurasians, and have basically been formed with different proportions of the same peoples (with a few exceptions who have a little bit of exotic in them).


So to answer your question. Yes, Sicilians are European genetically, it's just that Europeans are not an homogenous and easy-to-define thing. Don't be so dogmatic
Whites don't exist, Europeans don't exist, hence there is no such thing as race. What a familiar narrative. For some reason I don't see the same thing being said about other regions and continents this often...

Leto
05-10-2018, 07:24 PM
Genetic Europeans are those that cluster in Europe, with European populations. It's not that difficult see that Arabians, Africans, Indians and Chinese do NOT plot in Europe.

Flint
05-10-2018, 07:35 PM
Whites don't exist, Europeans don't exist, hence there is no such thing as race. What a familiar narrative. For some reason I don't see the same thing being said about other regions and continents this often...
Spot on and well said, Leto.
Interesting mental gymnastics that trolls will engage in, to hammer home their failed narratives.

Leto
05-10-2018, 07:48 PM
Spot on and well said, Leto.
Interesting mental gymnastics that trolls will engage in, to hammer home their failed narratives.
Well, I really don't want to discuss anything related to politics and agendas in any way. Actually, I don't think such discussions are encouraged here either. So let's not delve into that stuff even if we may have strong views on that subject or other subjects.

Ruderico
05-10-2018, 07:53 PM
Whites don't exist, Europeans don't exist, hence there is no such thing as race. What a familiar narrative. For some reason I don't see the same thing being said about other regions and continents this often...

The same thing applies to all populations except very isolated ones. If you want to politicise the issue go ahead and knock your self out

Flint
05-10-2018, 07:58 PM
Well, I really don't want to discuss anything related to politics and agendas in any way. Actually, I don't think such discussions are encouraged here either. So let's not delve into that stuff even if we may have strong views on that subject or other subjects.
I was swarmed by trolls, and was posting on the thread topic of 23andMe and why I feel their ancestry composition is fatally flawed in regards to some populations.
I believe any review of this thread will reveal that.
Also Leto, it's actually 23andMe that is injecting THEIR political narrative into their ancestry composition.
The company is run by an Ashknazi Jewish woman (Anne Wojcicki)...
Their definition of Ashkenazi Jews as "100% European" and ZERO middle eastern is highly political and that can't be overlooked.

Tz85
05-10-2018, 11:54 PM
Iberians, particularly western ones, have 10% real north African ancestry. Sicilians are not "hybridised" anymore than any other population is, we're all mixed, I'll bet that if you find samples from Sicily from 2000 years ago they will look rather similar to modern ones. You made no effort to deconstruct the argument, you just showed a crap 2pop model while ignoring the historical and genetic makeup that has been in place (in SE Europe) for thousands of years.

And you, new guy with a rude and aggressive posture, call others trolls. I don't generally do this, but kindly sod off back to TA. I hope I had banned you back there, so you'd have a reason to have that attitute

Iberians are more European than Sicilians, and Southern Italians. Fact.

shudra
05-11-2018, 12:04 AM
23andme just groups all South Asian countries as one unlike Europe, where it shows specific countries where you're descent is from, and so I don't think it would be very useful to a South Asian than GEDmatch would be with its detailed admix breakdown.

Archimedes
05-11-2018, 12:05 AM
Iberians, particularly western ones, have 10% real north African ancestry. Sicilians are not "hybridised" anymore than any other population is, we're all mixed, I'll bet that if you find samples from Sicily from 2000 years ago they will look rather similar to modern ones. You made no effort to deconstruct the argument, you just showed a crap 2pop model while ignoring the historical and genetic makeup that has been in place (in SE Europe) for thousands of years.

And you, new guy with a rude and aggressive posture, call others trolls. I don't generally do this, but kindly sod off back to TA. I hope I had banned you back there, so you'd have a reason to have that attitute

Please cite some studies that show that the avg/median Spaniard's genetic makeup is 10% north African

Ruderico
05-11-2018, 12:23 AM
Please cite some studies that show that the avg/median Spaniard's genetic makeup is 10% north African

I said "particularly western ones", not "average Spanish". There are a gazillion models online, but if you want a recent study there's this (https://www.biorxiv.org/content/biorxiv/early/2018/03/12/250191.full.pdf). Here (https://i.imgur.com/BwjEwnP.png)'s an image that illustrates. Yet no one in their right mind would say they are not European