italianguy
11-04-2019, 12:01 AM
Using two sisters to find hidden ancestries.
My mother, my aunt and I have all taken DNA tests. However, my mother and I took 23andMe, while my aunt took an Ancestry DNA test. I have uploaded all of our DNA to GEDmatch and I am curious as to the differences I see in my mother and aunt using the Eurogenes calculators. I am especially interested in the common ancestries they are showing for which the family has no paper trail.
Five ancestries show up repeatedly for both my mother and my aunt for which I have never found a paper trail. These are Lebanese Christian, Lebanese Muslim, Orcadian, variants of Sephardic Jewish and variants of Spanish.
Orcadian shows up A LOT, both for my mother and my aunt. My theory is that this must come from their father’s side because I have a paper trail on part of his family going WAY BACK to medieval England.
The Lebanese is weird in that Lebanese Christian shows up for my mother while Lebanese Muslim shows up for my aunt. Keep in mind that we are Sicilian on my maternal grandmother’s side and I have never found a paper trail there that leads out of Sicily, so I don’t know if the Lebanese and Jewish DNA, for instance, could be distant Middle Eastern DNA (due to the extremely diverse history of Sicily) pinging these populations as a proxy.
As for Sephardic Jewish, it shows up a lot in my mother’s K13 Oracle 2:
1 59.3% Southwest_English + 40.7% Sephardic_Jewish @ 2.93
2 57.1% West_Scottish + 42.9% Sephardic_Jewish @ 3.04
3 57.3% Irish + 42.7% Sephardic_Jewish @ 3.25
4 62.7% Southwest_English + 37.3% Tunisian_Jewish @ 3.57
5 69.6% Southwest_English + 30.4% Lebanese_Christian @ 3.76
6 60.6% West_Scottish + 39.4% Tunisian_Jewish @ 3.84
7 69.1% Southwest_English + 30.9% Samaritan @ 3.86
8 66% Southwest_English + 34% Cyprian @ 3.91
9 52.2% Italian_Abruzzo + 47.8% Southwest_English @ 3.94
10 58% Orcadian + 42% Sephardic_Jewish @ 3.99
11 83.5% French + 16.5% Kurdish_Jewish @ 3.99
12 57.7% West_Scottish + 42.3% Italian_Jewish @ 4
13 54.5% Italian_Abruzzo + 45.5% West_Scottish @ 4
14 83.4% French + 16.6% Iranian_Jewish @ 4.03
15 60.8% Irish + 39.2% Tunisian_Jewish @ 4.04
16 66.8% Southwest_English + 33.2% Lebanese_Muslim @ 4.06
17 63.1% Southwest_English + 36.9% Libyan_Jewish @ 4.06
18 60.7% Southeast_English + 39.3% Sephardic_Jewish @ 4.09
19 53.5% West_Scottish + 46.5% South_Italian @ 4.09
20 60% Southwest_English + 40% Italian_Jewish @ 4.13
And Kurdish Jewish shows up in the top five results for her K13 Oracle 4:
1 Kurdish_Jewish + Southwest_English + Southwest_English + Spanish_Aragon @ 2.919453
2 Kurdish_Jewish + Southwest_English + Spanish_Aragon + West_Scottish @ 2.950162
3 Kurdish_Jewish + Southeast_English + Southwest_English + Spanish_Aragon @ 3.040684
4 Kurdish_Jewish + Southwest_English + Southwest_English + Spanish_Murcia @ 3.072579
5 Kurdish_Jewish + Southeast_English + Spanish_Aragon + West_Scottish @ 3.077009
However, as you can tell, all of these have distances of more than 2.00.
For my aunt, her K13 Oracle 2 results don’t have anywhere near the presence of Jewish population groups like those of my mother. Only three of my aunt’s K13 Oracle 2 combos include Jewish. However, she still has 7/20 of the groupings including a variant of Jewish in K15 Oracle 2. What may be even more convincing however is that pesky Orcadian DNA. It shows up again in my aunt’s K15 Oracle 2.
For me, Lebanese Muslim barely shows up in my K13 Oracle 2 results. It is included as a small percentage in the last result. In fact, my K13 Oracle 2 results seem to be all over the place:
1 66% Southwest_English + 34% Bulgarian @ 2.07
2 62.5% Southwest_English + 37.5% Romanian @ 2.11
3 68.8% Irish + 31.2% Greek_Thessaly @ 2.12
4 96.1% West_German + 3.9% Moroccan @ 2.25
5 97.1% West_German + 2.9% Saudi @ 2.26
6 96.8% West_German + 3.2% Bedouin @ 2.3
7 66.7% North_German + 33.3% North_Italian @ 2.35
8 96.4% West_German + 3.6% Mozabite_Berber @ 2.35
9 96.4% West_German + 3.6% Algerian @ 2.37
10 96.9% West_German + 3.1% Syrian @ 2.39
11 96.6% West_German + 3.4% Tunisian @ 2.4
12 97.1% West_German + 2.9% Egyptian @ 2.42
13 67.8% West_Scottish + 32.2% Greek_Thessaly @ 2.42
14 97.1% West_German + 2.9% Jordanian @ 2.43
15 95.5% West_German + 4.5% East_Sicilian @ 2.43
16 95.4% West_German + 4.6% Central_Greek @ 2.44
17 94.6% South_Dutch + 5.4% Georgian_Jewish @ 2.45
18 96.7% West_German + 3.3% Turkish @ 2.45
19 97.3% West_German + 2.7% Samaritan @ 2.45
20 97.1% West_German + 2.9% Lebanese_Muslim @ 2.46
Spaniard doesn’t show up at all in mine except in single population sharing at distances of more than 14 and the only Jewish suggestion is Georgian Jewish at 5.4% in Result 17.
This raises a number of questions:
Does Eurogenes struggle with differentiating between Sicilian DNA and Sephardic Jewish DNA? If so, how long has this been going on and why is it happening?
To this day, the Orkney Islands have a population of less than 30,000. I would think this small population would have some genetic markers. Given this, do the Orcadian results indicate a high likelihood of ancestry originating from Orkney?
Why would two sisters have such different results? Could the difference between 23andme and Ancestry have something to do with it?
Are these Jewish and Lebanese population groups often “proxies” for other Middle Eastern ethnicities? In other words, are the real indications actually broader than the labels given here?
Was the Jewish population in Sicily before the expulsion of 1493 genetically distant enough from the rest of the island’s population to differentiate it in a person’s DNA 500 years later?
Is it a common problem that Sicilians show false positive results for admixtures of Jewish origin?
Are Spanish results common in Oracle for Sicilian customers?
How big are these reference population datasets and are they even reliable in the first place when it comes to GEDmatch admixtures? Is there somewhere I can get this technical information?
I hope some of you know a little more about these things and how to make actual meaning of these results than I do.
Thank you.
P.S. I didn’t post all of the results here because I was afraid it would be confusing and create a major TL;DR situation. If you’re interested or think they would help you answer my questions, just ask.
My mother, my aunt and I have all taken DNA tests. However, my mother and I took 23andMe, while my aunt took an Ancestry DNA test. I have uploaded all of our DNA to GEDmatch and I am curious as to the differences I see in my mother and aunt using the Eurogenes calculators. I am especially interested in the common ancestries they are showing for which the family has no paper trail.
Five ancestries show up repeatedly for both my mother and my aunt for which I have never found a paper trail. These are Lebanese Christian, Lebanese Muslim, Orcadian, variants of Sephardic Jewish and variants of Spanish.
Orcadian shows up A LOT, both for my mother and my aunt. My theory is that this must come from their father’s side because I have a paper trail on part of his family going WAY BACK to medieval England.
The Lebanese is weird in that Lebanese Christian shows up for my mother while Lebanese Muslim shows up for my aunt. Keep in mind that we are Sicilian on my maternal grandmother’s side and I have never found a paper trail there that leads out of Sicily, so I don’t know if the Lebanese and Jewish DNA, for instance, could be distant Middle Eastern DNA (due to the extremely diverse history of Sicily) pinging these populations as a proxy.
As for Sephardic Jewish, it shows up a lot in my mother’s K13 Oracle 2:
1 59.3% Southwest_English + 40.7% Sephardic_Jewish @ 2.93
2 57.1% West_Scottish + 42.9% Sephardic_Jewish @ 3.04
3 57.3% Irish + 42.7% Sephardic_Jewish @ 3.25
4 62.7% Southwest_English + 37.3% Tunisian_Jewish @ 3.57
5 69.6% Southwest_English + 30.4% Lebanese_Christian @ 3.76
6 60.6% West_Scottish + 39.4% Tunisian_Jewish @ 3.84
7 69.1% Southwest_English + 30.9% Samaritan @ 3.86
8 66% Southwest_English + 34% Cyprian @ 3.91
9 52.2% Italian_Abruzzo + 47.8% Southwest_English @ 3.94
10 58% Orcadian + 42% Sephardic_Jewish @ 3.99
11 83.5% French + 16.5% Kurdish_Jewish @ 3.99
12 57.7% West_Scottish + 42.3% Italian_Jewish @ 4
13 54.5% Italian_Abruzzo + 45.5% West_Scottish @ 4
14 83.4% French + 16.6% Iranian_Jewish @ 4.03
15 60.8% Irish + 39.2% Tunisian_Jewish @ 4.04
16 66.8% Southwest_English + 33.2% Lebanese_Muslim @ 4.06
17 63.1% Southwest_English + 36.9% Libyan_Jewish @ 4.06
18 60.7% Southeast_English + 39.3% Sephardic_Jewish @ 4.09
19 53.5% West_Scottish + 46.5% South_Italian @ 4.09
20 60% Southwest_English + 40% Italian_Jewish @ 4.13
And Kurdish Jewish shows up in the top five results for her K13 Oracle 4:
1 Kurdish_Jewish + Southwest_English + Southwest_English + Spanish_Aragon @ 2.919453
2 Kurdish_Jewish + Southwest_English + Spanish_Aragon + West_Scottish @ 2.950162
3 Kurdish_Jewish + Southeast_English + Southwest_English + Spanish_Aragon @ 3.040684
4 Kurdish_Jewish + Southwest_English + Southwest_English + Spanish_Murcia @ 3.072579
5 Kurdish_Jewish + Southeast_English + Spanish_Aragon + West_Scottish @ 3.077009
However, as you can tell, all of these have distances of more than 2.00.
For my aunt, her K13 Oracle 2 results don’t have anywhere near the presence of Jewish population groups like those of my mother. Only three of my aunt’s K13 Oracle 2 combos include Jewish. However, she still has 7/20 of the groupings including a variant of Jewish in K15 Oracle 2. What may be even more convincing however is that pesky Orcadian DNA. It shows up again in my aunt’s K15 Oracle 2.
For me, Lebanese Muslim barely shows up in my K13 Oracle 2 results. It is included as a small percentage in the last result. In fact, my K13 Oracle 2 results seem to be all over the place:
1 66% Southwest_English + 34% Bulgarian @ 2.07
2 62.5% Southwest_English + 37.5% Romanian @ 2.11
3 68.8% Irish + 31.2% Greek_Thessaly @ 2.12
4 96.1% West_German + 3.9% Moroccan @ 2.25
5 97.1% West_German + 2.9% Saudi @ 2.26
6 96.8% West_German + 3.2% Bedouin @ 2.3
7 66.7% North_German + 33.3% North_Italian @ 2.35
8 96.4% West_German + 3.6% Mozabite_Berber @ 2.35
9 96.4% West_German + 3.6% Algerian @ 2.37
10 96.9% West_German + 3.1% Syrian @ 2.39
11 96.6% West_German + 3.4% Tunisian @ 2.4
12 97.1% West_German + 2.9% Egyptian @ 2.42
13 67.8% West_Scottish + 32.2% Greek_Thessaly @ 2.42
14 97.1% West_German + 2.9% Jordanian @ 2.43
15 95.5% West_German + 4.5% East_Sicilian @ 2.43
16 95.4% West_German + 4.6% Central_Greek @ 2.44
17 94.6% South_Dutch + 5.4% Georgian_Jewish @ 2.45
18 96.7% West_German + 3.3% Turkish @ 2.45
19 97.3% West_German + 2.7% Samaritan @ 2.45
20 97.1% West_German + 2.9% Lebanese_Muslim @ 2.46
Spaniard doesn’t show up at all in mine except in single population sharing at distances of more than 14 and the only Jewish suggestion is Georgian Jewish at 5.4% in Result 17.
This raises a number of questions:
Does Eurogenes struggle with differentiating between Sicilian DNA and Sephardic Jewish DNA? If so, how long has this been going on and why is it happening?
To this day, the Orkney Islands have a population of less than 30,000. I would think this small population would have some genetic markers. Given this, do the Orcadian results indicate a high likelihood of ancestry originating from Orkney?
Why would two sisters have such different results? Could the difference between 23andme and Ancestry have something to do with it?
Are these Jewish and Lebanese population groups often “proxies” for other Middle Eastern ethnicities? In other words, are the real indications actually broader than the labels given here?
Was the Jewish population in Sicily before the expulsion of 1493 genetically distant enough from the rest of the island’s population to differentiate it in a person’s DNA 500 years later?
Is it a common problem that Sicilians show false positive results for admixtures of Jewish origin?
Are Spanish results common in Oracle for Sicilian customers?
How big are these reference population datasets and are they even reliable in the first place when it comes to GEDmatch admixtures? Is there somewhere I can get this technical information?
I hope some of you know a little more about these things and how to make actual meaning of these results than I do.
Thank you.
P.S. I didn’t post all of the results here because I was afraid it would be confusing and create a major TL;DR situation. If you’re interested or think they would help you answer my questions, just ask.