An experiment with modern averages only.

[JerryS.]

This is just for fun.

Because I cannot paper trail all of my ancestry back to documents coming from Europe, I decided to try to model myself with modern averages related to my known ancestry (England, Italy, Germany, Scotland). I found that none of my known ancestry fits by the percentage with any of the population averages on an autosomal scale. (However, I can model myself within reason using modern individuals, but that's for another thread). In doing this with modern averages, I only added to the group, nothing was moved or replaced with something else.

all English, all Scottish, all German, and all Italian modern averages.

Target: JerryS._scaled
Distance: 1.8760% / 0.01876041
56.6 Scottish
25.4 German
18.0 Italian_Aosta_Valley

adding all Greek modern averages because it is similar to my one traced South Italian, and all Danish because it borders North Germany where I have all of my German ancestry paper trailed.

Target: JerryS._scaled
Distance: 1.6426% / 0.01642647
73.4 Danish
24.0 Italian_Aosta_Valley
2.6 Greek_Trabzon

adding all Irish because it is similar to Scottish (and there are a few Irish surnames in that line), and adding all Welsh because it borders England. (This now covers all of the British Isles/Ireland where I have a little over half of my purported ancestry).

Target: JerryS._scaled
Distance: 1.6094% / 0.01609363
52.0 Danish
25.0 Irish
17.8 Italian_Aosta_Valley
3.6 Greek_Laconia
1.6 Greek_Trabzon


This doesn't really solve anything other than showing that for me, my English ancestry (which percentage wise on paper is the highest of all my ancestry) gets completely absorbed by other regional averages to the point of not showing on its own in any run.

*Just for fun (all of this is actually) I removed only the Italian Aosta Valley average from the last run. Only a slight bit of Sardinian left from the peninsula. This is the only time in this experiment that I removed any modern average.

Target: JerryS._scaled
Distance: 1.6303% / 0.01630261
45.6 Danish
40.0 Irish
12.6 Greek_Laconia
1.8 Sardinian

This last part has me wondering, with so many averages for just one country (Italy and Greece alike) what would their national averages be on a regional level. For instance, just using geographical locations would all of Italy average out in Umbria, and all of Greece average out in Thessaly?

Just my musings on a rainy day.

[digital_noise]

What is your goal here? Fun or facts?

I don’t have much to add that I have not already said in previous posts.

[lana6765]

If I try modelling my confirmed ancestry, which is Irish and English (I'm English with an Irish grandad and 3rd great grandad), this is how it goes:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

So I can get under 3% using these averages, which makes sense because of where I live and the research I've done so far. I should be mostly English and Irish. But it isn't the best fit and seems disproportionately Irish.

Here's what happens if I add Scottish (I have no known Scottish ancestry, but it's possible):

Target: Lana_scaled
Distance: 2.1600% / 0.02159950
54.4 Scottish
45.6 Irish

Uhhh suddenly I'm very Scottish.

Now I have some matches in Norway and AncestryDNA originally gave me 2% Norway. This is what happens if I add that:

Target: Lana_scaled
Distance: 1.9895% / 0.01989534
49.8 Norwegian
25.4 Irish
24.8 English

Wow. Umm.

Okay, and what about Spanish or Portuguese? This popped up on 23andme and mum swears a 3rd great grandmother had a Portuguese maiden name. I added Spanish_Andalucia:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

Oh...

Then I added different Spanish averages:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

Oh. Okay, now I'm going to add Finland. I have some interesting matches from Finland:

Target: Lana_scaled
Distance: 1.9926% / 0.01992603
60.6 Irish
28.4 English
11.0 Finnish

Wahay! Finland pops up. But it doesn't on AncestryDNA or 23andme. Somehow I think that would be fairly distinctive.

So what can I make of this? Somehow I think records are the only thing which can give me clues, and records can be wrong or misinterpreted. I've got back to the 1850s, but my family are cursed with having common names and upping and moving. I'm also not an expert and have made many false conclusions which I've had to later remove.

[lana6765]

This doesn't really solve anything other than showing that for me, my English ancestry (which percentage wise on paper is the highest of all my ancestry) gets completely absorbed by other regional averages to the point of not showing on its own in any run.

I think G25 is currently lacking in English samples. I don't know where English was taken from, but it doesn't seem to cover my English ancestry either.

I think they're possibly lacking in samples from the East and North. I'm not sure what London DNA is like. It could be all over the place.

[JerryS.]

I think G25 is currently lacking in English samples. I don't know where English was taken from, but it doesn't seem to cover my English ancestry either.

I think they're possibly lacking in samples from the East and North. I'm not sure what London DNA is like. It could be all over the place.

There are a lot of individual modern samples, however the average of them doesn't work for me; neither does the German average when I model myself to try to show my English with the rest of my autosomal data.

[JerryS.]

 

If I try modelling my confirmed ancestry, which is Irish and English (I'm English with an Irish grandad and 3rd great grandad), this is how it goes:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

So I can get under 3% using these averages, which makes sense because of where I live and the research I've done so far. I should be mostly English and Irish. But it isn't the best fit and seems disproportionately Irish.

Here's what happens if I add Scottish (I have no known Scottish ancestry, but it's possible):

Target: Lana_scaled
Distance: 2.1600% / 0.02159950
54.4 Scottish
45.6 Irish

Uhhh suddenly I'm very Scottish.

Now I have some matches in Norway and AncestryDNA originally gave me 2% Norway. This is what happens if I add that:

Target: Lana_scaled
Distance: 1.9895% / 0.01989534
49.8 Norwegian
25.4 Irish
24.8 English

Wow. Umm.

Okay, and what about Spanish or Portuguese? This popped up on 23andme and mum swears a 3rd great grandmother had a Portuguese maiden name. I added Spanish_Andalucia:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

Oh...

Then I added different Spanish samples:

Target: Lana_scaled
Distance: 2.1738% / 0.02173768
73.8 Irish
26.2 English

Oh. Okay, now I'm going to add Finland. I have some interesting matches from Finland:

Target: Lana_scaled
Distance: 1.9926% / 0.01992603
60.6 Irish
28.4 English
11.0 Finnish

Wahay! Finland pops up. But it doesn't on AncestryDNA or 23andme. Somehow I think that would be fairly distinctive.

So what can I make of this? Somehow I think records are the only thing which can give me clues, and records can be wrong or misinterpreted. I've got back to the 1850s, but my family are cursed with having common names and upping and moving. I'm also not an expert and have made many false conclusions which I've had to later remove.

try using all the averages for a particular country instead of only some. don't mix averages with individuals, and don't leave some averages out, use them all. there are over two dozen English/English and English/Cornwall modern individual samples, but only two English modern averages. This makes me think for the G25 data England is fairly homogenous. I can't speak for Spain, but I would use all the modern Spanish and Portuguese averages instead of just one or two regions.

[lana6765]

 

try using all the averages for a particular country instead of only some. don't mix averages with individuals, and don't leave some averages out, use them all. there are over two dozen English/English and English/Cornwall modern individual samples, but only two English modern averages. This makes me think for the G25 data England is fairly homogenous. I can't speak for Spain, but I would use all the modern Spanish and Portuguese averages instead of just one or two regions.

I only used averages, no samples. Sometimes I referred to the averages as samples for some reason. But yeah I'll incorporate England Cornwall in future.

[JerryS.]

I only used averages, no samples. Sometimes I referred to the averages as samples for some reason. But yeah I'll incorporate England Cornwall in future.

what I mean is there are only 2 modern (national?) English samples (English-English and English-Cornwall) but there are 20 something English-E and English-C modern individual samples. what I meant about mixing is using English modern(national?) samples with say German modern individual samples-don't do that. use only things from the same source or chart.

[Magnetic]

i recommend also to try mixing moderns and ancients a bit . it's pretty cool

for example for me :

Target: Magnetic_scaled
Distance: 2.9185% / 0.02918485

44.8 Samaritan
42.2 Northeast_Caucasus
13.0 Iran_Neolithic

-----

Distance: 2.8762% / 0.02876154

50.4 Palestinian_Beit_Sahour
35.8 Northeast_Caucasus
13.8 Iran_Neolithic

------

Distance: 2.7686% / 0.02768586

44.2 Greek_Crete
27.6 Greek_Central_Anatolia
25.0 IRN_Wezmeh_N
3.2 Oroqen


---

Distance: 2.7005% / 0.02700530

65.8 Lebanese_Druze
18.0 IRN_Ganj_Dareh_N
10.4 Irish
3.0 GEO_CHG
2.8 Oroqen

----

[JerryS.]

i recommend also to try mixing moderns and ancients a bit . it's pretty cool

for example for me :

Target: Magnetic_scaled
Distance: 2.9185% / 0.02918485

44.8 Samaritan
42.2 Northeast_Caucasus
13.0 Iran_Neolithic

-----

Distance: 2.8762% / 0.02876154

50.4 Palestinian_Beit_Sahour
35.8 Northeast_Caucasus
13.8 Iran_Neolithic

------

Distance: 2.7686% / 0.02768586

44.2 Greek_Crete
27.6 Greek_Central_Anatolia
25.0 IRN_Wezmeh_N
3.2 Oroqen


---

Distance: 2.7005% / 0.02700530

65.8 Lebanese_Druze
18.0 IRN_Ganj_Dareh_N
10.4 Irish
3.0 GEO_CHG
2.8 Oroqen

----

that's something I will have to play with later, but how did you get that?!