Eurogenes ANE K7 - post your results

Quote:

---

Originally Posted by Strasser-ue

Does there exist a better test for determining ANE?

---

There is only one legitimate way right now to find out how much ANE(that is, specific, ANE peaking genetics, as ANE also had WHG and Amerindian admixture) anyone has right now, and that is the Eastern_Euro component on Eurogenes K15.

All these low K calcs, and the typical non-sense that "Karitiana Indians, Kalash or North Caucasus people are the most ANE" you get from those low K calcs are not accurate at all. PCAs prove this as well.

K7 is overall a useless calculator, the new Basal-rich K7 will also be useless as it will be based on the direct AG3-MA1(especially ridiculous with MA1 since it's more than 24k years old, which won't account for later drift among populations and you're comparing it with modern newer 6-10k year old genomes without the archaic shift towards Andamanese unlike MA1) genome rather than artificially split up components based on peak, it still won't split up the WHG and Amerindian from it(if you think it had no WHG, how the heck could it have haplogroup U? WHGs also had ANE, both is from shared ancestry from earlier on, like in the paleolithic).

The most ANE populations, in order, are:

Afontova Gora 2(50%)
Karelia HG
Samara HG
Khvalynsk
Mari(38%)
Chuvash
Mal'ta
Yamnaya
Erzya
Selkup
Finnish Sami
SHGs
Kargopol Russians
East Finns
Ket
Shors
South-West Russians
Tatars
Belorusians
Lithuanians(21%)

Others:
Irish: 8%
Tabassaran: 14%
Burusho: 15%
West Greenland: 10%
Karitiana: 0%
South Indians/Tamils: 4%

K7 gives 24% to Mari, 27% to Tabassarans and other North Caucasus people, 30% to South Indians, 35% to certain Pakistani groups like the Burusho, and 44% to Karitiana, and differences between NE Europeans and NW Europeans in ANE are negligible.

Pretty obvious which is more accurate based on where ANE geographically actually was.

Quote:

---

Originally Posted by XenophobicPrussian

There is only one legitimate way right now to find out how much ANE(that is, specific, ANE peaking genetics, as ANE also had WHG and Amerindian admixture) anyone has right now, and that is the Eastern_Euro component on Eurogenes K15.

All these low K calcs, and the typical non-sense that "Karitiana Indians, Kalash or North Caucasus people are the most ANE" you get from those low K calcs are not accurate at all. PCAs prove this as well.

K7 is overall a useless calculator, the new Basal-rich K7 will also be useless as it will be based on the direct AG3-MA1(especially ridiculous with MA1 since it's more than 24k years old, which won't account for later drift among populations and you're comparing it with modern newer 6-10k year old genomes without the archaic shift towards Andamanese unlike MA1) genome rather than artificially split up components based on peak, it still won't split up the WHG and Amerindian from it(if you think it had no WHG, how the heck could it have haplogroup U? WHGs also had ANE, both is from shared ancestry from earlier on, like in the paleolithic).

The most ANE populations, in order, are:

Afontova Gora 2(50%)
Karelia HG
Samara HG
Khvalynsk
Mari(38%)
Chuvash
Mal'ta
Yamnaya
Erzya
Selkup
Finnish Sami
SHGs
Kargopol Russians
East Finns
Ket
Shors
South-West Russians
Tatars
Belorusians
Lithuanians(21%)

Others:
Irish: 8%
Tabassaran: 14%
Burusho: 15%
West Greenland: 10%
Karitiana: 0%
South Indians/Tamils: 4%

K7 gives 24% to Mari, 27% to Tabassarans and other North Caucasus people, 30% to South Indians, 35% to certain Pakistani groups like the Burusho, and 44% to Karitiana, and differences between NE Europeans and NW Europeans in ANE are negligible.

Pretty obvious which is more accurate based on where ANE geographically actually was.

---

In this case, I am 9.3% ANE.

Quote:

---

Originally Posted by XenophobicPrussian

There is only one legitimate way right now to find out how much ANE(that is, specific, ANE peaking genetics, as ANE also had WHG and Amerindian admixture) anyone has right now, and that is the Eastern_Euro component on Eurogenes K15.

The most ANE populations, in order, are:

Afontova Gora 2(50%)
Karelia HG
Samara HG
Khvalynsk
Mari(38%)
Chuvash
Mal'ta
Yamnaya

---

i agree with you. however eastern euro component of k15 was not purely ane at all. more like ehg and many other stuff. and ehg is 75 % ane and 25 % whg -- lazaridis approximation. just like overall whgs are 6,7 % ehg according to him. though swiss whg didnt have any extra amount of ehg and so ane. the only relationship it has with ane is having a very old common ancestor as every other have. so nothing special. also ane dont have whg -- you mistake it for its child (ehg) which has whg (~25 % as i said).

so modern chuvash definitely doesnt have more ane than mal'ta boy ancient genome has which it self along with ag2/3 is the reference for ane. :D its more than impossible -- technically. its not good idea to calculate ancients for ones thats made for moderns. it should be opposite. ancients can only be used to proxy to calculate modern ones. so better not to post proportions of ancient genomes using modern blogger calculators.

as to some incredibly high ane in some north indians, i also believe its mistake of calculator of low k that cant distinguish. for instance eurogenes upcoming k7 has a ag3-ma1 component but it has basically two very different stuff that have to be differentiated. (when saying two, im not saying ag3 and ma1 different, its same cluster, what im saying those calculator component has some thing not really ane but similar included).

similarly basal component and villabruna of that calculator.. villabruna component there is mix of more than 3 very very different thing, it has already have whg basal ane and many others in very mixed proportions for different populations and basal have whats actually not basal. so ane is not the only problematic thing with that calculator. for every component, he has to group some together so that he can have clear picture for some europeans which is the aim. if he doesnt do so and attempt to distinguish every thing as possible he can then the result would lead a huge diversity and distance among some europeans.

and one other thing.. in that calculator, ane in some pops may not be not correct for some but for other pops the levels are rather okay. and this is interesting. there is actually nothing wrong with mari and tabassaran s high level of ane as you said, they actually have even more than 24 and 27 if we pick all. but its true on other hand that,indians actually have much lower than north east europeans who have their ane lumped with other stuff on that calculator. i read very soon some 5000 year old ancient genomes coming from n india too though. and some info is already leaked and its nothing ane.

Quote:

---

Originally Posted by itsalex

i agree with you. however eastern euro component of k15 was not purely ane at all. more like ehg and many other stuff. and ehg is 75 % ane and 25 % whg -- lazaridis approximation. just like overall whgs are 6,7 % ehg according to him. though swiss whg didnt have any extra amount of ehg and so ane. the only relationship it has with ane is having a very old common ancestor as every other have. so nothing special. also ane dont have whg -- you mistake it for its child (ehg) which has whg (~25 % as i said).

so modern chuvash definitely doesnt have more ane than mal'ta boy ancient genome has which it self along with ag2/3 is the reference for ane

---

In short, ANE had no WHG, because it had EHG instead, but EHG had 25% WHG. Makes sense.

You're wrong.

Modern Chuvash absolutely has a better case for having more ANE than ANE itself, just like WHGs like La Brana, Loshbour, Villabruna, KO1 are more WHG than paleolithic WHGs such as El Miron, Vestonice, etc(and these are not different populations, Vestonice/etc wasn't replaced, these are populations after genetic drift)

Mal'ta should not be the reference for ANE, it is 24k years old, there has been too much drift genetic between then and now, which is why we also use mesolithic WHGs over paleolithic Europeans. It may as well be we don't have any admixture from Mal'ta at all(we do, but post-drift), if we had direct admixture from Mal'ta without drift going by how much ANE calcs give us we'd all have 5% Andamanese given Mal'ta had 25%, yet we don't. Afontova Gora is better(from 17k BC) but it's still not perfect. Until there is a sample from that region similar to Afontova Gora from 6-12k BC basing ANE on anything other than modern components is ridiculous(btw, you can see on PCAs, Afontova Gora plots near the populations I mentioned while Mal'ta is waaay south going towards Andamanese).

Mal'ta is a step between OoA archaic humans and evolving/drifting into a homogenous race, it's basically a mutt. The homogenous(the most far out on PCA plots) races are(although again, not totally homogenous, but atleast 80-90%) WHG, Natufian, Han Chinese, Amerindian, CHG, Melanesian, Ethiopian Ari/Hadza, Yoruba, Pygmy, San(and Anatolian/Balkan farmers as well as ANE to a far lesser extent).

Quote:

---

its not good idea to calculate ancients for ones thats made for moderns. it should be opposite. ancients can only be used to proxy to calculate modern ones. so better not to post proportions of ancient genomes using modern blogger calculators.

---

No, if you want to know which modern population is closest to an ancient one, you model the ancient one on modern components. I've already explained the problems with basing it on an entire genome of an ancient sample, especially if you don't have the right ones(for example, before other discoveries, calcs based on ancients, not modern samples, showed Sardinians, Bedouins and Georgians all with the same amount of EEF, even though PCA plots told a completely different story, yet people bought into this bullshit which basically said Sardinians were MENA). Calcs based on modern components definitely match up better with PCAs than calcs based on ancient components, which really, have results that are all over the place and don't make sense. Eurogenes K15 is still the most accurate calc out there, too many other calcs tell British people they're Croatian based on ancients.

Dad:


Population
ANE 15.71
ASE 1.70
WHG-UHG 62.49
East_Eurasian 1.04
West_African -
East_African 1.44
ENF 17.63

Mother:


Population
ANE 16.25
ASE 1.99
WHG-UHG 62.82
East_Eurasian 1.31
West_African 0.47
East_African 0.72
ENF 16.43

Quote:

---

Originally Posted by XenophobicPrussian

In short, ANE had no WHG, because it had EHG instead, but EHG had 25% WHG. Makes sense.

---

ane had no whg, yes.
whg (loschbour so on) had some ane, yes.
ane had ehg, no. -- the opposite, ehg had ane. ehg is the child of ane, being 75 % like ane and the rest 25 % being whg.

Quote:

---

Originally Posted by XenophobicPrussian

Modern Chuvash absolutely has a better case for having more ANE than ANE itself, just like WHGs like La Brana, Loshbour, Villabruna, KO1 are more WHG than paleolithic WHGs such as El Miron, Vestonice, etc(and these are not different populations, Vestonice/etc wasn't replaced, these are populations after genetic drift)

---

you are still saying modern chuvash is more whats called ane (ancient north eurasian) than mal'ta boy ancient genome (ma1) had? while calculator reference for ane is ma1? if there was another reference than ma1 and ag3 for calculators to use, then there would be slight possibility. but given calculators we have today, its technically impossible that chuvash is more ane than mal'ta boy. its like saying lithuanians are more whg than bichon hg, and then saying lithuanians should be reference for whg despite all the mix they have. thats like changing the whole meaning of the given word whg.

Quote:

---

Originally Posted by XenophobicPrussian

Mal'ta should not be the reference for ANE, it is 24k years old, there has been too much drift genetic between then and now, which is why we also use mesolithic WHGs over paleolithic Europeans.

---

maybe it shouldnt. but it is the reference. its not me who use mal'ta and ag3 as reference for ane, its those academicians and calculator bloggers what ever. that doesnt have to be direct admixture without drift, we have ane through some hgs. when a reference is used, the calculators treat it as 100 % being same stuff. whats lumped in it, its lost in a sense. (small andamese isnt the point here, calculators interested in the huge other part of mal'ta which cannot be explained by other ancient.) so for instance when there is a ehg component, all the ane and extra whg will be in it. and when there is a whg component as well, it will show you amont of extra whg you have (in addition to you have in ehg which is hidden to viewer of a spread sheet).

Quote:

---

Originally Posted by XenophobicPrussian

If you want to know which modern population is closest to an ancient one, you model the ancient one on modern components.

---

but unless you base your ane component on modern chuvash which is not the case, its not possible they will be more ane than mal'ta.
and no actually you shouldnt attempt to model ancient using moderns, instead one of the way to do is checking the shared genetic drift using various analysis and then rank the modern pops in ascending of percentage shared with that ancient in question. and ta da.

Quote:

---

Originally Posted by XenophobicPrussian

I've already explained the problems with basing it on an entire genome of an ancient sample, especially if you don't have the right ones(for example, before other discoveries, calcs based on ancients, not modern samples, showed Sardinians, Bedouins and Georgians all with the same amount of EEF, even though PCA plots told a completely different story, yet people bought into this bullshit which basically said Sardinians were MENA).

---

when there is a eef component that doesnt distinguish highly diverged ancestries from each other (which is funny i agree), its possible some irrelevant to each other folks have similar amount -- like 70 and even 80% (that suggests 70 % of their ancestry is common? no!). in reality that these three doesnt have equal amount of that farmer like ancestry, one of them (bed) has more basal and then natufian-like one. the other (sard) has more whg in addition to basal and natufian-like. the other (geo) has considerable ehg (and very small whg) in addition to basal, but not natufian-like. all these stuff are too different than each other to be lumped, i agree. a sardinian at best share only 25 % of its ancestry as common with bedouin. im saying this as hidden ancestry but when you have west med component as sardinian reference then no there will be nothing common.

Quote:

---

Originally Posted by XenophobicPrussian

Calcs based on modern components definitely match up better with PCAs than calcs based on ancient components, which really, have results that are all over the place and don't make sense. Eurogenes K15 is still the most accurate calc out there, too many other calcs tell British people they're Croatian based on ancients.

---

Spoiler!

modern vs ancient based calculator, well not really, they are equally imperfect if any thing. when its done using modern geographical ancestrals, it has disadvantage too. beause there would be some scapegoats when it comes to project the result (using pca). why? some modern pops which are relatively less mixed than other (for instance sardianians) will be used as reference (sardinian reference for a west med component) beause they are better proxy, in this case overall sardinians will be 70-80 % west med allthough this west med is actually mix of very very different stuff in it. and some thing with 80 % same thing will end up very different place in pca when others have it like 30 % at most. so sardinians will be the scapegoat in order to better calculate other ethnichities and they will be placed in very far point of a crazy 2d pca. if there was a very similar modern pop to sardinians, then the same problem for that too.

when its done using ancient ones, it has its advantages too. and actually geographic and ancient based ones overlap more than we think because calculators always having the same problem to group irrelevant stuff together -- either way. because almost every thing sucks, eurogenes k15 is of course one of relatively better ones to calculate but not the pca (actually davidski didnt release a pca of it, some one else did a 2d world scale pca of it which doesnt even correlate with fsts that davidski shared. mds for components tell a very different story. actually you know what? you shouldnt even run a pca using percentages on spread sheet because it ignores the relationship between components, common snps shared with others. you should run using snps from raw data -- that way even when its 2d there will be not that much huge problem of funny clusterings. but not every one can do it and when its done it doesnt really correlate with world political map. you know what i mean? thats the real one. that k15 pca was obviously done using only percentages released (even not exactly correlate with them) and thats why im telling this again, on that pca a half amerindian half tunisian can plot the exact place with an half indian half french despite these two people will be autosomally and in all aspect very different than each other as one can be. while this the case one cannot use that pca for any purpose. i repeat, these two gipotetical people plot in same place -- suggesting to eyes of untraineds that,they are similar. this is where the bullsh*t starts. :shrug: havent you ever though about that? thats why its the worst and only that worst pca thats going around that common (people who share bunch of common snps end up in wrong places where there is not actual share but ancestry from a similar). as to british being too similar to croatian which is some thing different, thats the problem of samples used, its the problem with the oracles mainly. it cas do with very low k too and there is other thing which i will not really expose here)) making a calculator is really tough job, we have several ancient genomes that share some ancestry with each other but we cant randomly choose one as ancestor to two very different populations. we should check for some speficic patterns. like some ancient ancestry (x) show up as donor exclusively only and only with some thing else (z) where as the other ancient (y) which share some similarity with the first but still different show up without what the other have (z). that way we can say one have ancetsry from x and other from y. thats why we shouldnt have a single component with having x and y together because of some thing common. whats not common is what matters more.

My results:

Population
ANE 18.69
ASE 1.51
WHG-UHG 64.30
East_Eurasian -
West_African -
East_African 0.60
ENF 14.89

By comparison, a Czech guy:

Population
ANE 16.99
ASE 1.99
WHG-UHG 63.61
East_Eurasian 0.23
West_African -
East_African 0.99
ENF 16.19

ANE 18.02
ASE 1.79
WHG-UHG 66.34
East_Eurasian -
West_African 0.55
East_African 1.28
ENF 12.02

ANE 12.94
ASE 1.21
WHG-UHG 40.63
East_Eurasian 7.37
West_African 2.36
East_African 9.65
ENF 25.85

ANE 15.39
ASE 1.85
WHG-UHG 62.47
East_Eurasian -
West_African -
East_African 0.77
ENF 19.52