Decoding values of G25 coordinates

1 positive: Atlantic-Baltic
1 negative: Sub-Saharan African

2 positive: West Med
2 negative: East Asian

3 positive: Mesolithic European
3 negative: South Asian

4 positive: Mesolithic Eurasian
4 negative: East Med

5 positive: Paleolithic Eurasian
5 negative: East Eurasian

6 positive: East Eurasian
6 negative: Oceanian

7 positive: Pygmy
7 negative: Amerindian

8 positive: East Asian
8 negative: Amerindian

9 positive: Northeast African
9 negative: West Asian

10 positive: West Med
10 negative: Mesolithic Eurasian

11 positive: Northeast Asian
11 negative: East Eurasian

12 positive: East Asian
12 negative: Red Sea

13 positive: Red Sea
13 negative: Caucasus

14 positive: Baltic
14 negative: East Asian

15 positive: Paleolithic Eurasian
15 negative: West Med

16 positive: South Asian
16 negative: Caucasus

17 positive: West Med
17 negative: South Asian

18 positive: Khoisan
18 negative: Paleolithic Eurasian

19 positive: East Asian
19 negative: North African

20 positive: West Asian
20 negative: South Asian

21 positive: Mesolithic European
21 negative: East Asian

22 positive: Mesolithic Eurasian
22 negative: North African

23 positive: Western Non-African
23 negative: Eastern Non-African

24 positive: Mesolithic Eurasian
24 negative: Mesolithic European

25 positive: East Eurasian
25 negative: Maori

:)

The first 5 values is the bulk of what Europeans score, the next 20 are close to zero. For all Europeans the sum of all values is always positive (above zero). The sum of all values never reaches 1 or -1.

Many have wondered why people with similar Yamnaya/ENF/WHG proportions can have a very large distance to each other, up to 0.07 in some cases. Some were more drifted from that Yamnaya/ENF/WHG calculator, some were less. It happents due to scattering of the same components into different positions in a chain.


https://sun9-16.userapi.com/impg/WVh...9b0&type=album


The pairings are:

8 and 21 with more East Asian for drifted, 12 with less East Asian for drifted.

9 with more West Asian for drifted, 20 with less West Asian for drifted.

10 with less West Med for drifted, 15 with more West Med for drifted.

10 with more Mesolithic Eurasian for drifted, 22 with less Mesolithic Eurasian for drifted.

13 with less Caucasus for drifted, 16 with more Caucasus for drifted.

15 with less Paleolithic Eurasian for drifted, 18 with more Paleolithic Eurasian for drifted.

16 with less South Asian for drifted, 20 with more South Asian for drifted.


The elevated distance between Europeans with a similar Yamnaya/ENF/WHG ratio is produced by a noise in coordinates.

Some of the choices for contrasting are very arbitrary, for example "Sub-Saharan African" is contrasted with a neutral European component which is widely present in all European populations and peaks in some Northern French population (French_Seine-Maritime). It is for a very arbitrary reason much lower in another Northern French population (French_Pas-de-Calais) and as a result a pairing of the first Northern French population with SSA would make another Northern French population score a full percent of SSA, but that is misleading because Davidski's choice to contrast SSA with some component that peaks in Central Europeans is completely baseless. The algorithm is simply searching how to neutralize the elevated "Atlantic-Baltic" component in one Northern French population in relation to a neighboring Northern French population by using a population with a negative value in the same position, in this case, a Sub-Saharan African. It would be a huge improvement if the system of mutual dependencies between components would be one-to-all, not one-to-one. It would eliminate much of the noise and neutralize the scattering effect.

Quote:

---

Originally Posted by Russki

Some of the choices for contrasting are very arbitrary, for example "Sub-Saharan African" is contrasted with a neutral European component which is widely present in all European populations and peaks in some Northern French population (French_Seine-Maritime). It is for a very arbitrary reason much lower in another Northern French population (French_Pas-de-Calais) and as a result a pairing of the first Northern French population with SSA would make another Northern French population score a full percent of SSA, but that is misleading because Davidski's choice to contrast SSA with some component that peaks in Central Europeans is completely baseless. The algorithm is simply searching how to neutralize the elevated "Atlantic-Baltic" component in one Northern French population in relation to a neighboring Northern French population by using a population with a negative value in the same position, in this case, a Sub-Saharan African. It would be a huge improvement if the system of mutual dependencies between components would be one-to-all, not one-to-one. It would eliminate much of the noise and neutralize the scattering effect.

---

IMO this also depends on those two averages (French Seine-Maritime and French_Pas-de-Calais) being made up of only a few individuals, usually less than 4. When I use averages on G25, I usually delete all those consisting of less than 5 individuals.

Quote:

---

Originally Posted by Ajeje Brazorf

IMO this also depends on those two averages (French Seine-Maritime and French_Pas-de-Calais) being made up of only a few individuals, usually less than 4. When I use averages on G25, I usually delete all those consisting of less than 5 individuals.

---

Another thing which also has a place in G25 is a scattering between individuals (which essentially prevents people with equal admixture proportions from being roughly identical and having ~0 distance - on a real test the distinction would be IBD sharing). While it is completely plausible that even two genetically similar Northern French individuals might have different proportions of various West Eurasian components, they both do not carry any SSA admixture, but because of a completely arbitrary decision to contrast SSA with some component peaking in Central Europeans the algorithm will produce a non-existent admixture. This phenomenon occurs in all 25 coordinates of every single individual and random spikes like the example above get leveled out when an average is calculated and the ability of G25 to produce non-existent admixtures when using population averages as references decreases, but remains.

Quote:

---

Originally Posted by Russki

Another thing which also has a place in G25 is a scattering between individuals (which essentially prevents people with equal admixture proportions from being roughly identical and having ~0 distance - on a real test the distinction would be IBD sharing). While it is completely plausible that even two genetically similar Northern French individuals might have different proportions of various West Eurasian components, they both do not carry any SSA admixture, but because of a completely arbitrary decision to contrast SSA with some component peaking in Central Europeans the algorithm will produce a non-existent admixture. This phenomenon occurs in all 25 coordinates of every single individual and random spikes like the example above get leveled out when an average is calculated and the ability of G25 to produce non-existent admixtures when using population averages as references decreases, but remains.

---

Is this why many western Europeans without roots in America are showing Amerindian DNA?