PDA

View Full Version : Please post your 23andme ancestry composition



Censored
02-27-2018, 12:39 AM
I am interested in the results of other South Asians only so I can compare.

poi
02-27-2018, 04:40 PM
I am interested in the results of other South Asians only so I can compare.

At the default speculative rate, 100% South Asian.
At the 90% conservative rate, 95% South Asian, 5% Unassigned.

Mine is v5 chip. Also, unlike others, I do not see the breakdown by countries. Does anyone with v5 chip get the breakdown?

misanthropy
02-27-2018, 07:34 PM
Hyderabadi Muslim. Some Chaush paternally (3-4 gens back), Syed maternally, rest is local.

http://i.imgur.com/zzY8h80.png

http://i.imgur.com/yqbJJp6.png

http://i.imgur.com/932rNhT.png

misanthropy
02-27-2018, 07:38 PM
But in general 23andme is such as waste for South Asians, because most will get 90-100% South Asian and it doesn't break down by region. 23andme considers the region from Afghanistan to Bangladesh as South Asian. Only thing it can show is some ancient trace DNA for many of us.

Btw, I'm interested in seeing if anyone else gets trace NW European in particular besides me. Not sure where it's from. Most here will get Eastern Euro I believe.

Censored
02-27-2018, 08:05 PM
But in general 23andme is such as waste for South Asians, because most will get 90-100% South Asian and it doesn't break down by region. 23andme considers the region from Afghanistan to Bangladesh as South Asian. Only thing it can show is some ancient trace DNA for many of us.

Btw, I'm interested in seeing if anyone else gets trace NW European in particular besides me. Not sure where it's from. Most here will get Eastern Euro I believe.

Where is your Euro coming from? Is it coming from the Yemeni side? Seems too high especially for a South Indian.

I personally have zero Euro only a bit of Middle Eastern and Sub-Saharan African.

misanthropy
02-27-2018, 09:31 PM
Where is your Euro coming from? Is it coming from the Yemeni side? Seems too high especially for a South Indian.

I personally have zero Euro only a bit of Middle Eastern and Sub-Saharan African.

For a South Indian, yeah I guess so. I'm not sure, but my wild speculation would be that it's related to the Yemeni component, perhaps from some common ancient trace med ancestry with Euros near coastal North Africa (my GEDmatch results occasionally show Egypt, Palestine, etc. for the Arab component). On the Relative Finder tool I get a similar amount of Anglo matches as Desis.

Tirunelvi
02-28-2018, 04:07 AM
21814

21815

21816

parasar
02-28-2018, 04:17 AM
Speculative:
South Asian 96.9%
European 2.0%
East Asian & Native American 0.3%
Sub-Saharan African < 0.1%
Western Asian & North African 0.0%
Melanesian 0.0%
Unassigned 0.7%

Conservative:
South Asian 65.3%
European 0.2%
Sub-Saharan African< 0.1%
Unassigned 34.5%

parasar
02-28-2018, 04:21 AM
But in general 23andme is such as waste for South Asians, because most will get 90-100% South Asian and it doesn't break down by region. 23andme considers the region from Afghanistan to Bangladesh as South Asian. Only thing it can show is some ancient trace DNA for many of us.

Btw, I'm interested in seeing if anyone else gets trace NW European in particular besides me. Not sure where it's from. Most here will get Eastern Euro I believe.

Euro 2.0%
Finnish0.1%
Broadly European1.4%
Broadly Northwestern European0.5%
Broadly Southern European< 0.1%

parasar
02-28-2018, 04:21 AM
Euro 2.0%
Finnish0.1%
Broadly European1.4%
Broadly Northwestern European0.5%
Broadly Southern European< 0.1%

No East European

Tirunelvi
02-28-2018, 04:22 AM
Speculative:
South Asian 96.9%
European 2.0%
East Asian & Native American 0.3%
Sub-Saharan African < 0.1%
Western Asian & North African 0.0%
Melanesian 0.0%
Unassigned 0.7%

Conservative:
South Asian 65.3%
European 0.2%
Sub-Saharan African< 0.1%
Unassigned 34.5%

Do you have India or any other South Asian countries as a country match with the new Beta results?.

poi
02-28-2018, 04:32 AM
Do you have India or any other South Asian countries as a country match with the new Beta results?.

I am curious to see what others are seeing in their breakdown. My reports page does not show anything but the generic South Asian.

Tirunelvi
02-28-2018, 04:33 AM
But in general 23andme is such as waste for South Asians, because most will get 90-100% South Asian and it doesn't break down by region. 23andme considers the region from Afghanistan to Bangladesh as South Asian. Only thing it can show is some ancient trace DNA for many of us.

Btw, I'm interested in seeing if anyone else gets trace NW European in particular besides me. Not sure where it's from. Most here will get Eastern Euro I believe.

From what I've seen on my DNA relatives some of the Brahmins I have score trace Northwest European and Scandinavian which varies with there siblings with some other relatives scoring above 1% Southern European. I would also assume that most would get Eastern Euro but I have not personally seen any Indian score Eastern European yet. I guess groups with more Steppe-related ancestry sometimes have it reflected in there 23andme results with minor Euro traces.

pnb123
02-28-2018, 04:43 AM
https://image.ibb.co/b4g3qx/pic1.png

Censored
02-28-2018, 04:52 AM
For a South Indian, yeah I guess so. I'm not sure, but my wild speculation would be that it's related to the Yemeni component, perhaps from some common ancient trace med ancestry with Euros near coastal North Africa (my GEDmatch results occasionally show Egypt, Palestine, etc. for the Arab component). On the Relative Finder tool I get a similar amount of Anglo matches as Desis.

Nevermind. I just looked at other results and it seems like I'm the only Indian on the planet without any Euro in me. lol

poi
02-28-2018, 04:56 AM
https://image.ibb.co/b4g3qx/pic1.png

Damn, dude! Right on!

pnb123
02-28-2018, 05:02 AM
Damn, dude! Right on!

Is this feature not available for v5 ?

Sapporo
02-28-2018, 11:21 AM
@pnb123

My 23andMe ancestry composition is still the same. Is this only for newer chip users (I'm V3)?

ssamlal
02-28-2018, 12:20 PM
My results: v5 chip @ 50% (speculative):

21820

pnb123
02-28-2018, 03:31 PM
@pnb123

My 23andMe ancestry composition is still the same. Is this only for newer chip users (I'm V3)?

Yes, the 23andme ancestry composition is still the same, but they added country/nationality feature in their new beta. Btw mine is V4.

Kaido
02-28-2018, 03:53 PM
https://i.imgur.com/PEo7HG3.png

khanabadoshi
02-28-2018, 03:54 PM
I'm trying to find this Beta feature... but in the meanwhile I found this interesting...




Ancestry Composition Populations
23andMe Customers
Public Data*
Total Individuals


East Asian and Native American

Native American (Colombian, Karitiana, Maya, Pima, Surui), East Asian, Japanese, Korean (South Korean), Yakut, Mongolian (Daur, Hezhen, Mongolian, Oroqen, Tu, Xibo), Chinese (Chinese, Han, Hong Kongese, Taiwanese) Southeast Asian (Burmese, Cambodian, Indonesian, Lao, Malaysian, Filipino, Thai, Vietnamese)
808

I was expecting 5983429582340 people. Seriously though.. at least 1000.


560

1368

How many Native Americans?




European

Southern European, Italian (Italian, Northern Italian, Tuscan), Balkan (Albanian, Bosnian and Herzegovinian, Bulgarian, Croatian, Greek, Macedonian, Maltese, Montenegrin, Romanian, Serbian), Sardinian, Iberian (France Basque, Portuguese, Spanish), Northwestern European, British and Irish (Irish, United Kingdom), French and German (Austrian, French, German, Belgian, Dutch, Swiss), Scandinavian (Danish, Norwegian, Swedish), Finnish, Ashkenazi, Eastern European (Belarusians, Czechs, Hungarians, Polish, Russian, Slovak, Slovene, Ukrainian)
6421

I was expecting more.
But this is probably why Euro results are more differentiated... I mean not just cuz there are so many customers.. but literally they have much more usable data to make the groups.


421




6842

...pretty sure the academic samples are better modeled by 23andme USERS data than vice vera lol... I mean they are such a blip.




Middle Eastern and North African

Middle Eastern (Armenian, Azerbaijani, Cypriot, Georgian, Druze, Iranian, Iraqi, Lebanese, Turkish, Syrian), North African (Algerian, Bahrani, Bedouin, Egyptian, Jordanian, Kuwaiti, Moroccan, Mozabite, Palestinian, Saudi Arabian, Tunisian, Emirati, Yemeni) well that's why Misanthropy scored North African. What a strange NA definition.

550

..damn
quite a lot.


176

^^ this is why NA is so broad.


726



Oceanian

Broadly Oceanian (Non-Austronesian Melanesian, Palauan, Tongan)

3

<-- 1 for each.


36

this is part of the reason why we don't
understand ASI, AAs, tribals, just anything at all really etc.


39



South Asian

Broadly South Asian (Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So Pakistani has got to be based on Punjabi
India is based on literally everything in India?

Afghan is based on both Pashtun and Tajiks in afghanistan.. but Pathan is its own thing for Pakistan?

Uygur... seriously. They had the HDGP samples and flipped coin "South Asia or East Asia? Head or Tails".

They aren't even putting in any effort.

207

behnchod..bas?
..seriously.
That's it.
You know why? Cuz no damn Desi agrees to research.

Pretty sure there are more than 200 people just named Muhammad Khan on 23andme...


615


HGDP
1000G

Lots. of samples of the same pops... 100 PJL 100 Bengali.. etc.. 50 Kalash..

Even 23andme probably doesn't have Xing's Kashmiri samples lol


822



Sub-Saharan African

West African (Bantu, Cameroonian, Ghanian, Ivorian, Liberian, Luhya, Mandenka, Nigerian, Sierra Leonean, Yoruba), East African (Eritrean, Ethiopian, Maasai, Somali), Central and South African (Biaka Pygmies, Mbuti Pygmies, San)

So just to reiterate the countries just to the west of these East african countries are called NORTH African in 23andme's atlas of the world.

228

393

621




So 23andme bases our admixture profiles from a total set of 10418 people.
That's a lot of people.. but like not a lot too.. for such a big company... and probably 100-200-300 blocks are of the same group

Mingle
02-28-2018, 04:12 PM
https://i.imgur.com/PEo7HG3.png

Interesting, quite a bit different from mine.

https://i.imgur.com/lb0zid4.png?1

Kaido
02-28-2018, 04:21 PM
Interesting, quite a bit different from mine.

I had similar results before I added a parent and my results were phased, here's the pre-phased composition

https://i.imgur.com/gHT6jSt.png

Mingle
02-28-2018, 04:39 PM
I had similar results before I added a parent and my results were phased, here's the pre-phased composition

You merged a parent's results with your own results? What does that mean and why would you do that?

Tirunelvi
02-28-2018, 04:39 PM
https://image.ibb.co/b4g3qx/pic1.png
Where is that Western Asian coming from?.

Kaido
02-28-2018, 04:43 PM
You merged a parent's results with your own results? What does that mean and why would you do that?

My mother took the test, and from there 23andme took her results and used them to make mine more accurate.

khanabadoshi
02-28-2018, 05:07 PM
So are the results updated, but some have Countries/Regions? Or... only some results are updated and those show countries? or.... there is a beta tab/button to click?

Tirunelvi
02-28-2018, 05:13 PM
So are the results updated, but some have Countries/Regions? Or... only some results are updated and those show countries? or.... there is a beta tab/button to click?

A small number of users have the new beta results, according to users who are on the beta there is a small chance in % but mostly just the new addition of Countries. In the blog post today by 23andme it should be coming to all users within the coming months.

Kaido
02-28-2018, 05:19 PM
Is there a list of all 150 populations?

Censored
02-28-2018, 07:58 PM
https://i.imgur.com/PEo7HG3.png

Thanks brother, I have been looking for Pashtun results for a long time. Also how the fuck does phasing make such drastic differences? 23andme needs to get their shit together.

Were you tested on V5 chip.

pnb123
02-28-2018, 08:47 PM
Where is that Western Asian coming from?.
23andme says that one of my ancestors who lived around 1760- 1850 is of 100% Western Asian origin. Idk anything beyond that.

pnb123
02-28-2018, 08:50 PM
So are the results updated, but some have Countries/Regions? Or... only some results are updated and those show countries? or.... there is a beta tab/button to click?

I think only some people got beta update. Thereís no button to click or anything like that. It just shows up in the place where your ancestry %ages are found.

Kaido
02-28-2018, 11:31 PM
Thanks brother, I have been looking for Pashtun results for a long time. Also how the fuck does phasing make such drastic differences? 23andme needs to get their shit together.

Were you tested on V5 chip.

Haha yeah I was surprised at the change too, but the drop in ME made more sense to me.

And this was on the V4 chip.

khanabadoshi
03-01-2018, 12:04 AM
Broadly South Asian
(Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So based on what's included in Broadly South Asian... shouldn't we all be 100% this thing? Like... wth.. Hazara + Kalash + Nepali + madar!chordo UYGHUR! all in SA!?
I guess they dumped ALL the HGDP samples and some 23andme peeps but the Afghan sample sets are small and maybe that's why Afghan Pashtun score less in this? PATHAN HGDP from Pakistan is relatively numerous and so people from Peshawar, Mardan, Swat and such are probably getting accounted from in this group.
Or.. I wonder if things are swayed by absolute numbers... like there PJL + BEB + Gujurat samples are like 300 on their own... there are 615 samples total... that maybe actually be what is happening. They may have just thrown all the HGDP samples from these groups in and didn't even bother to even ish out... or maybe accounted for actual population percentages? Like Most people in SA live in North India so accordingly they have x% of the samples from there, small % in Balochistan so small amount of samples from there.

How can a component that has literally every extreme population any of us could be AND every middle population between not account for 100% of all of us? I need to read their white paper on this.

Tirunelvi
03-01-2018, 02:27 AM
Is there a list of all 150 populations?

21850

bmoney
03-01-2018, 02:52 AM
South Asian
99.8%
Broadly South Asian
99.8%
European
0.2%
Northwestern European
0.1%
British & Irish
0.1%
Broadly European
< 0.1%

poi
03-01-2018, 02:52 AM
Broadly South Asian
(Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So based on what's included in Broadly South Asian... shouldn't we all be 100% this thing? Like... wth.. Hazara + Kalash + Nepali + madar!chordo UYGHUR! all in SA!?
I guess they dumped ALL the HGDP samples and some 23andme peeps but the Afghan sample sets are small and maybe that's why Afghan Pashtun score less in this? PATHAN HGDP from Pakistan is relatively numerous and so people from Peshawar, Mardan, Swat and such are probably getting accounted from in this group.
Or.. I wonder if things are swayed by absolute numbers... like there PJL + BEB + Gujurat samples are like 300 on their own... there are 615 samples total... that maybe actually be what is happening. They may have just thrown all the HGDP samples from these groups in and didn't even bother to even ish out... or maybe accounted for actual population percentages? Like Most people in SA live in North India so accordingly they have x% of the samples from there, small % in Balochistan so small amount of samples from there.

How can a component that has literally every extreme population any of us could be AND every middle population between not account for 100% of all of us? I need to read their white paper on this.

Lol, 23andme's ancestry is a joke, but they are geopolitically true(okay may be no Uighurs lol). Having said that, it is what it is -- they do it for business reasons(and/or technical limitations) probably.

I am curious to see what significantly East Asian admixed Nepalis like Newars or Tamangs score. Do they also score 90%-100% South Asian? If they have Hazaras(who are always significantly East Asian admixed) as South Asian, then I am pretty sure biologically South Asian means non-and-partial-WestEurasian. A rather broad category.

poi
03-01-2018, 02:54 AM
South Asian
99.8%
Broadly South Asian
99.8%
European
0.2%
Northwestern European
0.1%
British & Irish
0.1%
Broadly European
< 0.1%

Bro, what do you get in your conservative 90% mode?

bmoney
03-01-2018, 02:55 AM
Broadly South Asian
(Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So based on what's included in Broadly South Asian... shouldn't we all be 100% this thing? Like... wth.. Hazara + Kalash + Nepali + madar!chordo UYGHUR! all in SA!?
I guess they dumped ALL the HGDP samples and some 23andme peeps but the Afghan sample sets are small and maybe that's why Afghan Pashtun score less in this? PATHAN HGDP from Pakistan is relatively numerous and so people from Peshawar, Mardan, Swat and such are probably getting accounted from in this group.
Or.. I wonder if things are swayed by absolute numbers... like there PJL + BEB + Gujurat samples are like 300 on their own... there are 615 samples total... that maybe actually be what is happening. They may have just thrown all the HGDP samples from these groups in and didn't even bother to even ish out... or maybe accounted for actual population percentages? Like Most people in SA live in North India so accordingly they have x% of the samples from there, small % in Balochistan so small amount of samples from there.

How can a component that has literally every extreme population any of us could be AND every middle population between not account for 100% of all of us? I need to read their white paper on this.

Intellectual laziness and lack of an SA market in ancestral genetics

I saw this on my 23andme front page today, hopefully they will provide more than one segment for a sub-continent of 1.75 billion people:

Coming soon!

23andMe's Ancestry Composition update with 120 additional regions will be coming soon to all customers.

poi
03-01-2018, 02:58 AM
You guys have some xplainin' to do why you aint scoring at least 99% on this:



South Asian
Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur

:P

bmoney
03-01-2018, 02:59 AM
Bro, what do you get in your conservative 90% mode?


South Asian
99.5%
Broadly South Asian
99.5%
European
< 0.1%
Broadly European
< 0.1%
Unassigned
0.4%

pnb123
03-01-2018, 03:05 AM
Lol, 23andme's ancestry is a joke, but they are geopolitically true(okay may be no Uighurs lol). Having said that, it is what it is -- they do it for business reasons(and/or technical limitations) probably.

I am curious to see what significantly East Asian admixed Nepalis like Newars or Tamangs score. Do they also score 90%-100% South Asian? If they have Hazaras(who are always significantly East Asian admixed) as South Asian, then I am pretty sure biologically South Asian means non-and-partial-WestEurasian. A rather broad category.

I’ve a Chhetri relative (Rana). He’s scoring 24% E Asian on 23andme. Another Chhetri relative is getting 20% on 23andme. These numbers are pretty close to what Chhetris are getting on HarappaWorld. Another Chhetri relative is getting 64% (I think he’s quarter Chhetri). . He was scoring similarly in HarappaWorld. So I would assume Tamangs and Newars who are significantly more East Asian admixed will score more.

bmoney
03-01-2018, 03:12 AM
You guys have some xplainin' to do why you aint scoring at least 99% on this:



South Asian
Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur

:P

Might as well get a Homo sapiens to Neanderthal breakdown

Oh wait that's right, I actually have more neanderthal ancestry than non-SA ancestry in 23andme

Censored
03-01-2018, 03:15 AM
Broadly South Asian
(Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So based on what's included in Broadly South Asian... shouldn't we all be 100% this thing? Like... wth.. Hazara + Kalash + Nepali + madar!chordo UYGHUR! all in SA!?
I guess they dumped ALL the HGDP samples and some 23andme peeps but the Afghan sample sets are small and maybe that's why Afghan Pashtun score less in this? PATHAN HGDP from Pakistan is relatively numerous and so people from Peshawar, Mardan, Swat and such are probably getting accounted from in this group.
Or.. I wonder if things are swayed by absolute numbers... like there PJL + BEB + Gujurat samples are like 300 on their own... there are 615 samples total... that maybe actually be what is happening. They may have just thrown all the HGDP samples from these groups in and didn't even bother to even ish out... or maybe accounted for actual population percentages? Like Most people in SA live in North India so accordingly they have x% of the samples from there, small % in Balochistan so small amount of samples from there.

How can a component that has literally every extreme population any of us could be AND every middle population between not account for 100% of all of us? I need to read their white paper on this.

23andme seems to literally be the only company Iíve gotten results from that shows me as any % African which gives me huge red flags. Iím worried I made a mistake by testing with them...what DNA service would you have recommended?

poi
03-01-2018, 03:33 AM
I’ve a Chhetri relative (Rana). He’s scoring 24% E Asian on 23andme. Another Chhetri relative is getting 20% on 23andme. These numbers are pretty close to what Chhetris are getting on HarappaWorld. Another Chhetri relative is getting 64% (I think he’s quarter Chhetri). . He was scoring similarly in HarappaWorld. So I would assume Tamangs and Newars who are significantly more East Asian admixed will score more.

Wow, so "Nepali" means the Indic groups like Brahmins and non-tribal Madhesi castes?

bmoney
03-01-2018, 03:38 AM
Wow, so "Nepali" means the Indic groups like Brahmins and non-tribal Madhesi castes?

Thats not fair to Nepal - if anything Indic groups came from the South and West and diffused into Nepal judging by the ancient NE Asian hap O found there

poi
03-01-2018, 03:48 AM
Thats not fair to Nepal - if anything Indic groups came from the South and West and diffused into Nepal judging by the ancient NE Asian hap O found there

In this case, I'm not complainig lol. For once, bahuns are considered natives. Seriously, though, what they should do, is remove Nepal altogether. There are literally 300 distinct ethnicities speaking 300 languages from 4 linguistic families living in the same strip of the Himalayan hills.

pnb123
03-01-2018, 03:51 AM
Wow, so "Nepali" means the Indic groups like Brahmins and non-tribal Madhesi castes?

Idk which groups are exactly included. But I think Pahari Brahmins were major focus. Western Nepal was included as Central Asia in FtDNA (along with North India, Pakistan, and Kashmir), so maybe focus was mostly on Western Nepali Paharis on this one too. We would have to look if Non Tribal Madheshis are getting Nepal as their match to find out if they were included too.

Censored
03-01-2018, 03:52 AM
South Asian
99.8%
Broadly South Asian
99.8%
European
0.2%
Northwestern European
0.1%
British & Irish
0.1%
Broadly European
< 0.1%

Another South Indian scoring European... Wow, I officially AM the only Indian in the world who doesn't!

poi
03-01-2018, 03:53 AM
Edit- oops, didn't read properly. My bad.

poi
03-01-2018, 04:03 AM
Idk which groups are exactly included. But I think Pahari Brahmins were major focus. Western Nepal was included as Central Asia in FtDNA, so maybe focus was mostly on Western Nepali Paharis. We would have to look if Non Tribal Madheshis are getting Nepal as their match to find out if they were included too.

I have absolutely no hits on non-Bahun nepalis. So you have some Chettris. Any Madhesis?

pnb123
03-01-2018, 04:09 AM
I have absolutely no hits on non-Bahun nepalis. So you have some Chettris. Any Madhesis?

No Madheshis, but some Chhetris (maybe 5-6). Majority are Bahuns. I think I’m getting Chhetris because at some point they also descended from Bahuns. As Bahun marrying Chhetri, would result in Chhetri offspring. Or Bahun marrying any non-Dalit castes would also result in Chhetri. There is one Chhetri relative with G haplogroup, which is nonexistent in any Bahun 23andme relative.

poi
03-01-2018, 04:19 AM
No Madheshis, but some Chhetris (maybe 5-6). Majority are Bahuns. I think Iím getting Chhetris because at some point they also descended from Bahuns. As Bahun marrying Chhetri, would result in Chhetri offspring. Or Bahun marrying any non-Dalit castes would also result in Chhetri. There is one Chhetri relative with G haplogroup, which is nonexistent in any Bahun 23andme relative.

That makes sense, but I don't have any. Is it due to much less people usinf v5 versus v4? Heck, there are Europeans but no nonBahun Nepalis.may be v5 relatives will build up as time passes.

bmoney
03-01-2018, 04:45 AM
In this case, I'm not complainig lol. For once, bahuns are considered natives. Seriously, though, what they should do, is remove Nepal altogether. There are literally 300 distinct ethnicities speaking 300 languages from 4 linguistic families living in the same strip of the Himalayan hills.

Agree

Mingle
03-01-2018, 06:45 AM
In this case, I'm not complainig lol. For once, bahuns are considered natives. Seriously, though, what they should do, is remove Nepal altogether. There are literally 300 distinct ethnicities speaking 300 languages from 4 linguistic families living in the same strip of the Himalayan hills.

Brahmins are usually not considered natives? When did Khas people first become a majority in Western Nepal?

In Nepal, I assume that Khas people would be genetically somewhat different from their mainstream generic Indian sample. So Nepal probably shouldn't be removed. If Nepal is removed, then I don't think it makes much sense to have Pakistan, Sri Lanka, or Bangladesh either.

For a region such as South Asia, it makes no sense to divide it based on very genetically diverse countries that are not even loosely based on ethnolinguistics in the first place (unlike Europe). They should have divided South Asia into categories such as Tajik, Pashtun, Baloch, Northwest Indian, North Indian, East Indian, South Indian, etc. As flawed as that approach may be, it would be an improvement over the current setup.

It causes a lot of confusion the way they did it. But I understand that this product is mostly targeted towards Americans of mixed European background and they don't care about the genetic components we're familiar with (CHG, ASI, etc) but are instead interested in which country their ancestry comes from. It's a good thing that we have GEDmatch.

Mingle
03-01-2018, 06:46 AM
Broadly South Asian
(Afghan, Balochi, Bangladeshi, Brahui, Burusho, Hazara, Indian, Kalash, Makrani, Nepalese, Pakistani, Pathan, Sindhi, Sri Lankan, Uygur)

So based on what's included in Broadly South Asian... shouldn't we all be 100% this thing? Like... wth.. Hazara + Kalash + Nepali + madar!chordo UYGHUR! all in SA!?
I guess they dumped ALL the HGDP samples and some 23andme peeps but the Afghan sample sets are small and maybe that's why Afghan Pashtun score less in this? PATHAN HGDP from Pakistan is relatively numerous and so people from Peshawar, Mardan, Swat and such are probably getting accounted from in this group.
Or.. I wonder if things are swayed by absolute numbers... like there PJL + BEB + Gujurat samples are like 300 on their own... there are 615 samples total... that maybe actually be what is happening. They may have just thrown all the HGDP samples from these groups in and didn't even bother to even ish out... or maybe accounted for actual population percentages? Like Most people in SA live in North India so accordingly they have x% of the samples from there, small % in Balochistan so small amount of samples from there.

How can a component that has literally every extreme population any of us could be AND every middle population between not account for 100% of all of us? I need to read their white paper on this.

Do you know what their Indian sample might be based off of? Do you think it's the same as the generic 'Indian' component in GEDmatch? I think the generic Indian component is an Uttar Pradeshi Brahmin. Correct me if I'm mistaken.

poi
03-01-2018, 06:57 AM
Brahmins are usually not considered natives? When did Khas people first become a majority in Western Nepal?

In Nepal, I assume that Khas people would be genetically somewhat different from their mainstream generic Indian sample. So Nepal probably shouldn't be removed. If Nepal is removed, then I don't think it makes much sense to have Pakistan, Sri Lanka, or Bangladesh either.

For a region such as South Asia, it makes no sense to divide it based on very genetically diverse countries that are not even loosely based on ethnolinguistics in the first place (unlike Europe). They should have divided South Asia into categories such as Tajik, Pashtun, Baloch, Northwest Indian, North Indian, East Indian, South Indian, etc. As flawed as that approach may be, it would be an improvement over the current setup.

It causes a lot of confusion the way they did it. But I understand that this product is mostly targeted towards Americans of mixed European background and they don't care about the genetic components we're familiar with (CHG, ASI, etc) but are instead interested in which country their ancestry comes from. It's a good thing that we have GEDmatch.

Typically brahmins were imported by kings, so they had been more mobile than other groups. Plus, the political narrative in South Asia(India/Nepal at least) is around tribals=natives, lowercaste=oppressed, uppercastes=invaders. There is some truth to that, so that is what usually carries over to the western journals/studies etc.

And the history of Khas is shrouded in mystery. People do not even want to associate with that relic group... but I suspect the Khas has been in Western Nepal for a long time. With future full Y and mtDNA analysis(or whole genome sequencing for that matter), perhaps we can figure out subclades that might point to the actual migration patterns.

And I agree with you regarding 23andme's decisions. I actually think they are doing what is for the best. Plus, they provide the raw genotyping, so we can use whatever calculator to figure out CHG and EHG and ASI and ANE and WTF. :P

khanabadoshi
03-01-2018, 07:40 AM
Do you know what their Indian sample might be based off of? Do you think it's the same as the generic 'Indian' component in GEDmatch? I think the generic Indian component is an Uttar Pradeshi Brahmin. Correct me if I'm mistaken.

I skimmed thru the white paper.. their process is pretty damn involved.. they use the entire HGDP set and 1000 Genomes. They don't define the differences between anything that fits in broadly South Asian, because they cannot confidently say that this is Nepali or this is Bengali or whatever with at least 50% certainty across the board. Basically If the any of the groups show in the South Asian list combined together cross 50% - 100% probablilty of being the ancestral population for that section of the chromosome... its gets defined as "South Asian" -- they probably can see which populations in the South Asian group are actually the ones there... but they aren't confident enough to say "yes...this part is definitely from Baloch and that from Pashtun...


Check out their white paper, skip the equations.. that's not for me or you to look at haha.. but read the method.. you'll get the idea of why they don't define the South Asians. Also, they make 25 clusters on a PCA... so Uyghur is falling into a closer cluster with the South Asians then the East Asians I guess.

I'm a little too sleepy to type all that askofjogvj out.

MonkeyDLuffy
03-01-2018, 08:17 AM
23andme ancestry composition gives me cancer every time I open it.

https://i.imgur.com/1NK89hs.png

Censored
03-01-2018, 09:56 AM
23andme ancestry composition gives me cancer every time I open it.

https://i.imgur.com/1NK89hs.png

Thanks, I was waiting for yours lol. Why does it give you cancer?

I can relate to the feeling btw.

Sapporo
03-01-2018, 10:40 AM
Jatt Sikhs I share with on 23andMe range anywhere from 88-100% South Asian (most are probably around 96-97%). The non South Asian parts are mostly European or Middle Eastern with sometimes a few % unassigned. Some individuals score a lot more European and others more Middle Eastern.

https://i.imgur.com/TixFKn9.png

I presume that after my mother and father's V5 results come in, my data will be phased and I'll get near 100% South Asian.

itsmrhoax
03-01-2018, 11:02 AM
Lol, 23andme's ancestry is a joke, but they are geopolitically true(okay may be no Uighurs lol). Having said that, it is what it is -- they do it for business reasons(and/or technical limitations) probably.

I am curious to see what significantly East Asian admixed Nepalis like Newars or Tamangs score. Do they also score 90%-100% South Asian? If they have Hazaras(who are always significantly East Asian admixed) as South Asian, then I am pretty sure biologically South Asian means non-and-partial-WestEurasian. A rather broad category.


I’ve a Chhetri relative (Rana). He’s scoring 24% E Asian on 23andme. Another Chhetri relative is getting 20% on 23andme. These numbers are pretty close to what Chhetris are getting on HarappaWorld. Another Chhetri relative is getting 64% (I think he’s quarter Chhetri). . He was scoring similarly in HarappaWorld. So I would assume Tamangs and Newars who are significantly more East Asian admixed will score more.


Thats not fair to Nepal - if anything Indic groups came from the South and West and diffused into Nepal judging by the ancient NE Asian hap O found there


Brahmins are usually not considered natives? When did Khas people first become a majority in Western Nepal?

In Nepal, I assume that Khas people would be genetically somewhat different from their mainstream generic Indian sample. So Nepal probably shouldn't be removed. If Nepal is removed, then I don't think it makes much sense to have Pakistan, Sri Lanka, or Bangladesh either.

For a region such as South Asia, it makes no sense to divide it based on very genetically diverse countries that are not even loosely based on ethnolinguistics in the first place (unlike Europe). They should have divided South Asia into categories such as Tajik, Pashtun, Baloch, Northwest Indian, North Indian, East Indian, South Indian, etc. As flawed as that approach may be, it would be an improvement over the current setup.

It causes a lot of confusion the way they did it. But I understand that this product is mostly targeted towards Americans of mixed European background and they don't care about the genetic components we're familiar with (CHG, ASI, etc) but are instead interested in which country their ancestry comes from. It's a good thing that we have GEDmatch.

I mustíve joined this forum at the right time. So Iím Nepali (father chhetri and mother limbu)

Whatíd likely be my expected ancestry percentages?

bmoney
03-01-2018, 11:15 AM
I must’ve joined this forum at the right time. So I’m Nepali (father chhetri and mother limbu)

What’d likely be my expected ancestry percentages?

Welcome to the forum! good to see some non Khas Brahmin Nepalis :P

Have you got an ancestry kit yet?

itsmrhoax
03-01-2018, 12:49 PM
Welcome to the forum! good to see some non Khas Brahmin Nepalis :P

Have you got an ancestry kit yet?

Thanks! No not yet. Iím not sure which one to go for

I mean Iím doing some reasearch but my parents donít seem to know or care too much about their origins lol. Anyway, how much truth is there to people of Khas origins being from Iran/central asia and India? I donít even know if my father is the ďIranicĒ or ďIndicĒ type of chhetri though he does look more Iranian than he does Indian if that helps. And Iím assuming my mongloid side would get me some East Asian? So where does that leave South Asia and how far does it go back, will I have as much SA as I previously thought? (At least 90%+)

So many questions. I which I was better aware of our genealogy!

poi
03-01-2018, 01:59 PM
I must’ve joined this forum at the right time. So I’m Nepali (father chhetri and mother limbu)

What’d likely be my expected ancestry percentages?

Welcome to the forum. If you take the test from Ancestry or ftdna, upload your raw file to Gedmatch. We will learn a lot! Especially curious about Limbu.

poi
03-01-2018, 02:16 PM
Thanks! No not yet. Iím not sure which one to go for

I mean Iím doing some reasearch but my parents donít seem to know or care too much about their origins lol. Anyway, how much truth is there to people of Khas origins being from Iran/central asia and India? I donít even know if my father is the ďIranicĒ or ďIndicĒ type of chhetri though he does look more Iranian than he does Indian if that helps. And Iím assuming my mongloid side would get me some East Asian? So where does that leave South Asia and how far does it go back, will I have as much SA as I previously thought? (At least 90%+)

So many questions. I which I was better aware of our genealogy!

Definetely take the test from Ancestry or FTDNA. Or both. We can combine results when we run different calculators. Do it as soon as possible, because it is atleast a month's grueling wait to get the results.

In your specific case, despite the phenotype, I think Chettris will be similar to Bahuns except for higher East Asian of varying degree. Your Limbu will increase your East Asian, so curious to see by how much.

Is there a chance you could also test someone from the Limbu side of your family?

MonkeyDLuffy
03-01-2018, 02:24 PM
Thanks, I was waiting for yours lol. Why does it give you cancer?

I can relate to the feeling btw.

Because it doesn't give you any details lol. As others and khana showed their frustration with piece of joke calculator they use, its a cancer.

bmoney
03-01-2018, 02:25 PM
Thanks! No not yet. I’m not sure which one to go for

I mean I’m doing some reasearch but my parents don’t seem to know or care too much about their origins lol. Anyway, how much truth is there to people of Khas origins being from Iran/central asia and India? I don’t even know if my father is the “Iranic” or “Indic” type of chhetri though he does look more Iranian than he does Indian if that helps. And I’m assuming my mongloid side would get me some East Asian? So where does that leave South Asia and how far does it go back, will I have as much SA as I previously thought? (At least 90%+)

So many questions. I which I was better aware of our genealogy!

Khas Brahmins genetically cluster with their Indic neighbours - Pahari Brahmins from states like Uttarakhand, so no doubt they initially were from this pop or a cousin pop to them

They have surprisingly maintained a genetic moat from other Nepalis so to speak, whereas Brahmins in southern India and UP are less divergent from people within their state - maybe it suggests the Khas are relative newcomers to the area as @Poi has mentioned before

I gotta say though the concept of genetic countries in South Asia is meaningless - a Sinhalese Sri Lankan loosely clusters with a Bangladeshi, and a Pakistani Punjabi would cluster more with Indian Punjabis than an Indian Punjabi would to a UP-ite. Better to use self-reported ethnicity

When you say look Indian do you mean look Madhesi? there is no Indian 'look' as such

Also you probably would score way higher E Asian than Nepali Brahmins based on the Chettri samples we've seen so far and your Limbu side would increase this even further - Im guessing your SA (steppe+ASI+Baloch) could be as low as 50-60%

pnb123
03-01-2018, 03:45 PM
I must’ve joined this forum at the right time. So I’m Nepali (father chhetri and mother limbu)

What’d likely be my expected ancestry percentages?

Welcome to the forum.

Mingle
03-01-2018, 03:54 PM
a Sinhalese Sri Lankan loosely clusters with a Bangladeshi

How is this possible? From the Bengali results I have seen, they score quite a bit more East Eurasian than even Anatolian Turks. I've also seen a Bengali hit the 20% mark in his East Eurasian score on GEDmatch. Sinhalas I assume would have very little East Asian in comparison.

pnb123
03-01-2018, 03:56 PM
Khas Brahmins genetically cluster with their Indic neighbours - Pahari Brahmins from states like Uttarakhand, so no doubt they initially were from this pop or a cousin pop to them

They have surprisingly maintained a genetic moat from other Nepalis so to speak, whereas Brahmins in southern India and UP are less divergent from people within their state - maybe it suggests the Khas are relative newcomers to the area as @Poi has mentioned before

I gotta say though the concept of genetic countries in South Asia is meaningless - a Sinhalese Sri Lankan loosely clusters with a Bangladeshi, and a Pakistani Punjabi would cluster more with Indian Punjabis than an Indian Punjabi would to a UP-ite. Better to use self-reported ethnicity

When you say look Indian do you mean look Madhesi? there is no Indian 'look' as such

Also you probably would score way higher E Asian than Nepali Brahmins based on the Chettri samples we've seen so far and your Limbu side would increase this even further - Im guessing your SA (steppe+ASI+Baloch) could be as low as 50-60%

Actually there’s only one sample from Uttarakhand and he’s showing way more East Asian than any Khas Brahmin from nepal. If this is a norm for other Uttarakhand Brahmins, then they would cluster with Chhetris. So right now we don’t know if they cluster with us.

poi
03-01-2018, 06:15 PM
Actually thereís only one sample from Uttarakhand and heís showing way more East Asian than any Khas Brahmin from nepal. If this is a norm for other Uttarakhand Brahmins, then they would cluster with Chhetris. So right now we donít know if they cluster with us.

Yep, that Uttarkhand Brahmin clusters with Nepali Chettri. The harappa pca might be of interest for our new member.

https://anthrogenica.com/showthread.php?13310-HarappaWorld-PCAs

itsmrhoax
03-01-2018, 06:44 PM
Khas Brahmins genetically cluster with their Indic neighbours - Pahari Brahmins from states like Uttarakhand, so no doubt they initially were from this pop or a cousin pop to them

They have surprisingly maintained a genetic moat from other Nepalis so to speak, whereas Brahmins in southern India and UP are less divergent from people within their state - maybe it suggests the Khas are relative newcomers to the area as Poi has mentioned before

I gotta say though the concept of genetic countries in South Asia is meaningless - a Sinhalese Sri Lankan loosely clusters with a Bangladeshi, and a Pakistani Punjabi would cluster more with Indian Punjabis than an Indian Punjabi would to a UP-ite. Better to use self-reported ethnicity

When you say look Indian do you mean look Madhesi? there is no Indian 'look' as such

Also you probably would score way higher E Asian than Nepali Brahmins based on the Chettri samples we've seen so far and your Limbu side would increase this even further - Im guessing your SA (steppe+ASI+Baloch) could be as low as 50-60%

That probably wasnít the smartest thing to say on this forum haha. Anyway, not madehsi, you just wouldnít think heíd be from anywhere in India as he has more of a persian look but then you have men who look irani/afghan in India so its impossible to really tell.

And yeah definitely. I know a chhetri who did the test and got like 40% central Asian. And the Limbu side could also be really varied so yeah, I plan to do it soon.

poi
03-01-2018, 07:19 PM
That probably wasn’t the smartest thing to say on this forum haha. Anyway, not madehsi, you just wouldn’t think he’d be from anywhere in India as he has more of a persian look but then you have men who look irani/afghan in India so its impossible to really tell.

And yeah definitely. I know a chhetri who did the test and got like 40% central Asian. And the Limbu side could also be really varied so yeah, I plan to do it soon.

Interesting about the phenotype. Chettris can have varying looks, so I would not doubt the phenotype part. Genetically, however, Chettris are not different component-wise from Bahuns, except for higher East Asian. Is it possible to have your dad tested? Also, "Central Asian" could mean a lot of things within the genetic context. FTDNA has Nepalis as part of the Central Asian cluster (I score 82% Central Asian in FTDNA).

MonkeyDLuffy
03-01-2018, 11:00 PM
Yea Central Asian is simply including the east Asian in it. Chhettris score very high east Asian. Phenotype again does not go along with genotype. Even someone with 65-66 SI can have Rosey white skin tone. That's why phenotype is considered pseudo science.

Censored
03-01-2018, 11:09 PM
Yea Central Asian is simply including the east Asian in it. Chhettris score very high east Asian. Phenotype again does not go along with genotype. Even someone with 65-66 SI can have Rosey white skin tone. That's why phenotype is considered pseudo science.

Yes, my family on my dad’s side is mixed and some people look completely Indian but their brother or sister look straight out of Afghanistan or Turkey. Genetic expression can be wild.

itsmrhoax
03-01-2018, 11:17 PM
Interesting about the phenotype. Chettris can have varying looks, so I would not doubt the phenotype part. Genetically, however, Chettris are not different component-wise from Bahuns, except for higher East Asian. Is it possible to have your dad tested? Also, "Central Asian" could mean a lot of things within the genetic context. FTDNA has Nepalis as part of the Central Asian cluster (I score 82% Central Asian in FTDNA).

Is there a reason why FTDNA does this and Ancestry not? What would be the more "accurate" result. And I'm guessing a lot of you aren't fond of 23andme as nobody is suggesting it? lol

Sapporo
03-01-2018, 11:22 PM
Is there a reason why FTDNA does this and Ancestry not? What would be the more "accurate" result. And I'm guessing a lot of you aren't fond of 23andme as nobody is suggesting it? lol

23andMe's V5 chip is kind of crap (compared to V3 and V4). It's only worthwhile combined with FTDNA or another chip. I'd recommend FTDNA, Ancestry.com and then 23andMe. After that, LivingDNA, MyHeritage or Geno 2.0 (National Geographic).

bmoney
03-01-2018, 11:31 PM
How is this possible? From the Bengali results I have seen, they score quite a bit more East Eurasian than even Anatolian Turks. I've also seen a Bengali hit the 20% mark in his East Eurasian score on GEDmatch. Sinhalas I assume would have very little East Asian in comparison.

Refer to pois Harappa pca

And youre right about Sinhalas not scoring E Asian like Bengalis - Im guessings its the similarity of the other segments that pull them towards Bengali.

As i've mentioned in the other thread they were meant to have migrated from the Vanga kingdom

bmoney
03-01-2018, 11:32 PM
Actually thereís only one sample from Uttarakhand and heís showing way more East Asian than any Khas Brahmin from nepal. If this is a norm for other Uttarakhand Brahmins, then they would cluster with Chhetris. So right now we donít know if they cluster with us.

Thats Uttaranchal Brahmin not Uttarakhand - the sample looks admixed ive seen it

No I was referring to other Pahari Brahmins who wont look like the Harappa Uttaranchal Brahmin

poi
03-02-2018, 12:07 AM
Is there a reason why FTDNA does this and Ancestry not? What would be the more "accurate" result. And I'm guessing a lot of you aren't fond of 23andme as nobody is suggesting it? lol

Different companies do different things, probably for business reasons and to attract/keep their customers happy, but all of them provide "raw" genotyping on a select SNPs. Using the raw genotype files, so we can use them on many other calculators. Those companies use different genotyping chips(to cut costs, eventually they will do full genome sequencing).

Starting August 2017, 23andme started using their latest v5 chip, which does not work properly with existing calculators. FTDNA and Ancestry still use older chips, so existing calculators(built by amateurs and pros alike) work much better.

If you're still deciding, you should pick either FTDNA or Ancestry because you want to use your data in those calculators. If you are content with whatever those companies provide, go with 23andme, since it is the latest chip.

poi
03-02-2018, 12:15 AM
Thats Uttaranchal Brahmin not Uttarakhand - the sample looks admixed ive seen it

No I was referring to other Pahari Brahmins who wont look like the Harappa Uttaranchal Brahmin

Lol, I did now know that difference. Is the cricketer Rishabh Pant Uttaranchal or Uttarkhand?

MonkeyDLuffy
03-02-2018, 01:27 AM
Uttrakhand is a state which includes rishikesh/ haridawar and dehradun regions as well. The brahmins from these regions are more similar to UP and Haryanvi Brahmins. Uttranchal is used more for Garhwal.

khanabadoshi
03-02-2018, 02:05 AM
Idk which groups are exactly included. But I think Pahari Brahmins were major focus. Western Nepal was included as Central Asia in FtDNA (along with North India, Pakistan, and Kashmir), so maybe focus was mostly on Western Nepali Paharis on this one too. We would have to look if Non Tribal Madheshis are getting Nepal as their match to find out if they were included too.


Based on their method, the 23andme users who all identified their 4 grandparents as Nepali are all examined, outliers removed, people who they deem probably don't actually have 4 grandparents from Nepal are removed, close relatives are removed, and what's left gets used as the Nepali base. On average they removed no less than 10% of the samples that claim 4 grandparents from any given region.

So Nepali likely consists of what the majority of Nepalis who have taken a 23andme test are.

pnb123
03-02-2018, 02:11 AM
Based on their method, the 23andme users who all identified their 4 grandparents as Nepali are all examined, outliers removed, people who they deem probably don't actually have 4 grandparents from Nepal are removed, close relatives are removed, and what's left gets used as the Nepali base. On average they removed no less than 10% of the samples that claim 4 grandparents from any given region.

So Nepali likely consists of what the majority of Nepalis who have taken a 23andme test are.

Most of my relatives are Nepali Bahuns, so yes it makes sense for me to score Nepal, that too with 100% confidence. Maybe other groups are included as well. Do they not specify any ethnicities ?

Varun R
03-02-2018, 02:12 AM
21881

The Tamil/ Kannadiga Brahmins all show 99-100% S Asian, and the vast majority fall on the higher end of the spectrum. Many show 100% S Asian.

pnb123
03-02-2018, 02:26 AM
Thats Uttaranchal Brahmin not Uttarakhand - the sample looks admixed ive seen it

No I was referring to other Pahari Brahmins who wont look like the Harappa Uttaranchal Brahmin

Uttarakhand used to be called Uttaranchal before. They’re the same thing. But yes, I agree with you there probably are Brahmins who are less East Asian admixed than that sample in Harappa. But we won’t know this until we’ve more samples from Uttarakhand Brahmins.

pnb123
03-02-2018, 02:28 AM
Lol, I did now know that difference. Is the cricketer Rishabh Pant Uttaranchal or Uttarkhand?

AFAIK, they’re the same thing.

khanabadoshi
03-02-2018, 02:30 AM
23andMe's V5 chip is kind of crap (compared to V3 and V4). It's only worthwhile combined with FTDNA or another chip. I'd recommend FTDNA, Ancestry.com and then 23andMe. After that, LivingDNA, MyHeritage or Geno 2.0 (National Geographic).

Just to clarify. V5's raw data is currently crap for analyzing ancestry with the SNPs we've been used to using. Essentially no one has caught up and the calculators and ancient raw data is all on older standards. The V5 chip however, is probably the best thing on the market. It tests for more SNPs than V4, just not the same ones. So it's not "backwards compatible". In that sense everyone should understand it's massive upgrade... bigger than the difference of the previous chips, because you could still compare V3 and V4 to a reasonable level. Illumina and 23andme are basically saying after looking at all the data we have.... all those SNPs we used to test for are pretty useless or highly predictable -- these SNPs that we didn't test for previously are way more important and informative for assigning ancestry.

So, the wisest thing for anyone to do who hasn't tested is get an FTDNA test (which is comparable to V4) so that you can test on the current slew of calculators and compare to ancient samples that are on those standards, and then get a V5 chip later. Things will take time to adjust to V5, and if you are thinking they won't... that's a mistake. Illumina makes the chips, 23andme just customizes them to its own proprietary research. LivingDNA went straight to the newer chip, didn't even bother with any other. FTDNA and Ancestry will likely switch as well to keep up. Illumina has already demonstrated that this chip is considerably more accurate than their last, so it will become the standard at some point and everyone with older chips will be left in a position that people on V5 currently feel now. However, the difference is the people with V5 will get tests made for them, and everyone else will slowly start to notice less and less calcs are working for them, until finally nothing.

In terms of a cost-benefit look at things. Maybe Paul Gill was right... Whole Genomes is the way to go.

Kurd
03-02-2018, 02:46 AM
I just noticed that S Asians are not getting W Asian for the most part. They are getting percentages from far away places like Europe, N Africa, and W Africa before W Asian.

Does this make any sense? Of course not. Considering that W Asian includes neighboring Iranian and Caucasian this should by far be the largest category after S Asian for S Asians, yet many are getting 0 here, yet a little African and European.

S Asians should look at their unassigned category (90% confidence mode). I believe almost all of this is W Asian. Instead they are throwing it into S Asian in speculative mode (50% confidence)

So with 23 minor ancestry gets wacked twice, once when they take up to 49/100 markers in their haplotype windows and convert this minor ancestry to major, and the 2nd time when they take the unassigned and assign it to major.

The reverse forIranians and other W Asians. The majority of the 25-35% unassigned is probably S Asian

khanabadoshi
03-02-2018, 02:53 AM
Most of my relatives are Nepali Bahuns, so yes it makes sense for me to score Nepal, that too with 100% confidence. Maybe other groups are included as well. Do they not specify any ethnicities ?


This is how they start out. They specify ethnicity if they can say that only these alleles can come from that ethnicity, not that one is
like that ethnicity. The standard for classifying is higher. If you are coming out as 100% Nepali, they are saying that on every chromosome, the order of your alleles on phased segments is specific to a person from Nepal. ie. if I had the same order of alleles as you in that segment I would get assigned Nepali for that segment. That is a high bar they have established. They are able to say (at 90%, 80%, 70%, 50% levels of confidence) that this segments of alleles only occurs among Nepalis across the entire set of chromosomes, and not in anyone else in South Asia.


When 23andme models you, they actually make 4 fake grandparents based on you, and their grandparents and so on, back 4 generations... and then let it simulate all the way down to see if their fake ancestors for you in any combination can produce your alleles. That's pretty damn involved.

Part of their process is this:




2.1.2 Training data
We trained the local classifier on more than 9700 individuals divided into K = 25reference populations. We denote N the number of training individuals. The reference populations are a combination of countries and broader geographic regions. We required that a population contain at least 25 individuals. When a country did not contain enough individuals, it was grouped with countries from the same geographic region. The process was repeated until each population contained enough individuals. Next, we describe the steps involved in building the training set. We started by building a dataset of N∗ = 10699unrelated individuals with known/self-reported ancestry. Out of the 10699 individuals, 1793 came from publicly available datasets (1000 Genomes: 765, CEPHHGDP: 941, HapMap3: 87). The remaining 8906 were research-consented 23andMe members who reported via a survey on the 23andMe website that their four grandparents were born in the same country. The text of the questions asked was “In which country was your (mother’s mother) born?”, for each of the four grandparents. The answer was chosen from a list of countries, which included the option “I don’t know”. This number includes 23andMe members who indicated that all four of their grandparents had Ashkenazi Jewish ancestry. This was ascertained by a separate question on the same survey. The text of that question reads: “If your (mother’s mother’s) birthplace does not fully describe (her) ancestry, please provide additional ancestry-relevant information here. Your answer may be any mix of national, ethnic, religious, or other labels, separated by semicolons. To illustrate: transnational or ethnic identities, such as Ashkenazi, Basque, and Cherokee, are often more relevant to ancestry than birthplace. Or, birthplace may not accurately reflect ancestry for a child of immigrants. For example, if this person was born in the US to immigrants from India, you might write Hindu; Indian.” The text supplied by members in response to this question was searched for occurrences of “Ashkenazi”. We ensured that all the reference individuals were distantly related by removing individuals from the sample until no two individuals shared more than an estimated 100 cM identical-by-descent (IBD). IBD estimates were obtained according to Henn et al. (2012). We then performed genome-wide PCA on the remaining individuals. We computed the distance of each individual to his/her 20 nearest-neighbors in the space defined by the first ten principal components (PCs), which enabled us to infer the empirical distribution of distance to nearest neighbors for each of the K populations. Finally, we removed individuals outside of the 5th-quantile interval of the distribution. Table S1 and Table S2 summarize the composition of our training set.

Mingle
03-02-2018, 03:04 AM
Refer to pois Harappa pca

And youre right about Sinhalas not scoring E Asian like Bengalis - Im guessings its the similarity of the other segments that pull them towards Bengali.

As i've mentioned in the other thread they were meant to have migrated from the Vanga kingdom

I saw them cluster together on poi's PCA. I was just confused on how they ended up clustering together when one has almost no EE whereas the other has it in very large quantities. I guess it makes sense that they are more similar in other components, but their EE difference is HUGE so I was just wondering if anyone had a more detailed explanation on the reason why that happened.

bmoney
03-02-2018, 03:05 AM
I saw them cluster together on poi's PCA. I was just confused on how they ended up clustering together when one has almost no EE whereas the other has it in very large quantities. I guess it makes sense that they are more similar in other components, but their EE difference is HUGE so I was just wondering if anyone had a more detailed explanation on the reason why that happened.

Also check this out

https://en.wikipedia.org/wiki/Genetic_studies_on_Sinhalese#Predominantly_Bengali _origin

Censored
03-02-2018, 03:11 AM
I just noticed that S Asians are not getting W Asian for the most part. They are getting percentages from far away places like Europe, N Africa, and W Africa before W Asian.

Does this make any sense? Of course not. Considering that W Asian includes neighboring Iranian and Caucasian this should by far be the largest category after S Asian for S Asians, yet many are getting 0 here, yet a little African and European.

S Asians should look at their unassigned category (90% confidence mode). I believe almost all of this is W Asian. Instead they are throwing it into S Asian in speculative mode (50% confidence)

So with 23 minor ancestry gets wacked twice, once when they take up to 49/100 markers in their haplotype windows and convert this minor ancestry to major, and the 2nd time when they take the unassigned and assign it to major.

The reverse forIranians and other W Asians. The majority of the 25-35% unassigned is probably S Asian

Ideally most Indians should not be getting any Euro or African if they are testing only 500 years back so I canít complain about them not getting West Asian either.

khanabadoshi
03-02-2018, 03:16 AM
I just noticed that S Asians are not getting W Asian for the most part. They are getting percentages from far away places like Europe, N Africa, and W Africa before W Asian.

Does this make any sense? Of course not. Considering that W Asian includes neighboring Iranian and Caucasian this should by far be the largest category after S Asian for S Asians, yet many are getting 0 here, yet a little African and European.

S Asians should look at their unassigned category (90% confidence mode). I believe almost all of this is W Asian. Instead they are throwing it into S Asian in speculative mode (50% confidence)

So with 23 minor ancestry gets wacked twice, once when they take up to 49/100 markers in their haplotype windows and convert this minor ancestry to major, and the 2nd time when they take the unassigned and assign it to major.

The reverse forIranians and other W Asians. The majority of the 25-35% unassigned is probably S Asian


They have such a thought-out and logical process and where they fail really is just in the most simple matter -- understanding that you don't have to group everything by a country's borders. They are taking the method that is useful in Europe (and works because they have such a large set of data from Europe) and applying to regions that have far more diverse populations (with far less people tested) and it causes issues.

Though, when reading their paper, one can understand the how and why of all this. They are definitely trying to be more accurate, but they value confidence > specificity; thus, in order to be as accurate and confident as possible for regions like ours, they inherently had to be less specific.

Kurd
03-02-2018, 03:17 AM
I just noticed that S Asians are not getting W Asian for the most part. They are getting percentages from far away places like Europe, N Africa, and W Africa before W Asian.

Does this make any sense? Of course not. Considering that W Asian includes neighboring Iranian and Caucasian this should by far be the largest category after S Asian for S Asians, yet many are getting 0 here, yet a little African and European.

S Asians should look at their unassigned category (90% confidence mode). I believe almost all of this is W Asian. Instead they are throwing it into S Asian in speculative mode (50% confidence)

So with 23 minor ancestry gets wacked twice, once when they take up to 49/100 markers in their haplotype windows and convert this minor ancestry to major, and the 2nd time when they take the unassigned and assign it to major.

The reverse forIranians and other W Asians. The majority of the 25-35% unassigned is probably S Asian


Whereas S Asians are assigned 90%+ S Asian, W Asians are assigned 90%+ Middle Eastern. Here is a N Iraqi Kurd to illustrate. The S And E Asian percentages are severely underestimated.

https://i.imgur.com/DrDjdM8.jpg


Although in their ancestry time line it seems that they are sort of cognizant of the fact that they underestimate minor ancestry, because the ancestry time line does not correlate well with their ancestry percentages

https://i.imgur.com/7ky2maP.jpg


Thus, users should look at the unassigned in conservative mode (90%) in this case it is 24%. This is where most of the S Asian is hiding for W Asians, and W Asian for S Asians. Instead they toss it in with the major ancestry. The reason it is unassigned in the 1st place could be that it is very similar (for ex similarity in allele frequencies for Iranian/Caucasian and western S Asian)

https://i.imgur.com/Job6wvN.jpg

Kurd
03-02-2018, 03:27 AM
They have such a thought-out and logical process and where they fail really is just in the most simple matter -- understanding that you <em>d</em><em>on't have to</em> group everything by a country's borders. They are taking the method that is useful in Europe (and works because they have such a large set of data from Europe) and applying to regions that have far more diverse populations (with far less people tested) and it causes issues.<br>
<br>
Though, when reading their paper, one can understand the how and why of all this. They are definitely trying to be more accurate, but they value confidence &gt; specificity; thus, in order to be as accurate and confident as possible for regions like ours, they inherently had to be less specific.

The problem is that your average consumer is led to believe that these are very homogenous populations who have lived in their respective bubbles for eons, and that for example Iranians and Pakistanis (however you want to define this) are&nbsp; genetically very distinct from each other, which of course they are not.

misanthropy
03-02-2018, 03:35 AM
The problem is that your average consumer is led to believe that these are very homogenous populations who have lived in their respective bubbles for eons, and that for example Iranians and Pakistanis (however you want to define this) are genetically very distinct from each other, which of course they are not.

Exactly. These are modern borders. Just playing around with this interactive world history atlas shows how those divisions (e.g. Iranian vs Pakistani vs Afghan) are pretty moot: http://geacron.com/home-en/

poi
03-02-2018, 03:36 AM
Whereas S Asians are assigned 90%+ S Asian, W Asians are assigned 90%+ Middle Eastern. Here is a N Iraqi Kurd to illustrate. The S And E Asian percentages are severely underestimated.

https://i.imgur.com/DrDjdM8.jpg


Although in their ancestry time line it seems that they are sort of cognizant of the fact that they underestimate minor ancestry, because the ancestry time line does not correlate well with their ancestry percentages

https://i.imgur.com/7ky2maP.jpg


Thus, users should look at the unassigned in conservative mode (90%) in this case it is 24%. This is where most of the S Asian is hiding for W Asians, and W Asian for S Asians. Instead they toss it in with the major ancestry. The reason it is unassigned in the 1st place could be that it is very similar (for ex similarity in allele frequencies for Iranian/Caucasian and western S Asian)

https://i.imgur.com/Job6wvN.jpg

I am 95% South asian in 90% conservative mode. Doesn't that 95% already mask plenty of West/East Asian? So, beyond geopolitical borders, 23andme is useless. Atleast FTDNA has " Ancient Origins". The components there aren't relevant for South Asians, though.

bmoney
03-02-2018, 03:41 AM
Uttarakhand used to be called Uttaranchal before. Theyíre the same thing. But yes, I agree with you there probably are Brahmins who are less East Asian admixed than that sample in Harappa. But we wonít know this until weíve more samples from Uttarakhand Brahmins.

Youre correct sorry what I meant was what Monkey referred to

Uttranchal is used more for Garhwal.- Some Garhwalis do have significant E Asian mix and its phenotypically visible

To clarify I mean Kumaoni, Dogra, Himachali Brahmins etc would be closer to Nepali Brahmins than other pops

Kurd
03-02-2018, 03:43 AM
I am 95% South asian in 90% conservative mode. Doesn't that 95% already mask plenty of West/East Asian? So, beyond geopolitical borders, 23andme is useless. Atleast FTDNA has " Ancient Origins". The components there aren't relevant for South Asians, though.

of course it does, because a S Asian Bengali or Nepali reference when broken down is actually Indian tribal + W Asian + steppe + E Asian + ..... So the E Asian is included in S Asian. Don’t forget if the 100 SNP haplotype windows have 40 E Asian SNPs, those E Asian SNPs are assigned S Asian by their smoothing algorithm (majority rule)

khanabadoshi
03-02-2018, 03:50 AM
The problem is that your average consumer is led to believe that these are very homogenous populations who have lived in their respective bubbles for eons, and that for example Iranians and Pakistanis (however you want to define this) are&nbsp; genetically very distinct from each other, which of course they are not.


Yes. And thus it's time for:

PUBLIC SERVICE ANNOUNCEMENT FOR THE HINDI SECTION OF ANTHROGENICA

Based on the published methods of 23andme we must accept some fault.

We did not specify who we were to 23andme, and since they only have ~200 samples to use from us with the HGDP and 1000Genomes samples ... we are definitely affecting how things are analyzed. We are simultaneously the CONTROL and the VARIABLE of these tests.

Accordingly, I went through all my 14 kits on 23andme and changed their groupings. I checked MIDDLE EASTERN, ASIAN, and OTHER -- this will cause 23andme to at least compare to an Iranian at some point. Before I had just checked Asian.

In OTHER, you can write down whatever you want. Separate each ethnic/clan/tribe/caste group with a ";" -- they specify that in their paper that their algorithm searches for this. I wrote:

PAKISTAN: 43.75% Saraiki-Baloch; 25% Kashmiri; ...; 6.25% UNKNOWN SOUTH ASIAN.

I think it is important to specify what you don't know as well.

Then, in all sections where one can specify Location, Region, Tribes, Castes ... whatever.. I wrote everything as detailed as possible. I ALSO SPECIFIED WHAT IS MATERNAL AND PATERNAL... because STEP 1 of anything 23andme does is PHASE YOU, and at some point they do or will factor in what you say is Maternal/Paternal on questioners into that phased equation.

I guarantee the vast majority of Europeans were very specific in their answers which is how so many subgroups came to be in Europe on their tests. We are NOT specific beyond countries/language. We should be more specific.

Kurd
03-02-2018, 04:01 AM
Yes. And thus it's time for:

PUBLIC SERVICE ANNOUNCEMENT FOR THE HINDI SECTION OF ANTHROGENICA

Based on the published methods of 23andme we must accept some fault.
We did not specify who we were to 23andme, and since they only have ~200 samples of to use form us with the HGDP and 1000Genomes samples ... we are definitely affecting how things are analyzed. We are simultaneously the CONTROL and the VARIABLE of these tests.

Accordingly, I went through all my 14 kits on 23andme and changed their groupings.
I checked MIDDLE EASTERN, ASIAN, and OTHER -- this will cause 23andme to at least compare to an Iranian at some point. Before I had just checked Asian.

In OTHER -- you can write down whatever you want. Separate each ethnic/clan/tribe/caste group with a ";" -- they specify that in their paper that their algorithm searches for this.

I wrote: PAKISTAN: 43.75% Saraiki-Baloch; 25% Kashmiri; etc etc etc 6.25% UNKNOWN SOUTH ASIAN. I think it is important to specify what you don't know as well.

Then in all sections where one can specify Location, Region, Tribes, Castes ... whatever.. I wrote everything as detailed as possible. I ALSO SPECIFIED WHAT IS MATERNAL AND PATERNAL... because STEP 1 of anything 23andme does is PHASE YOU, and at some point they do or will factor in what you say is Maternal/Paternal on questioners into that phased equation.

I guarantee the vast majority of Europeans were very specific in their answers which is how so many subgroups came to be in Europe on their tests. We are NOT specific beyond countries/language. We should be more specific.


Doesn’t hurt to try, but I honestly don’t think that will change things much, because of the way their algorithm works. For them It’s almost like trying to guess where you are from with the results forced to confirm with the guess via mathematical manipulation. Having such a huge bullseye such as S Asia or ME makes it easyto guess correctly. Correct guesses = less consumer complaints. It’ All business :)

Edit: I can see them tossing people out who they feel are too mixed between their groupings, or labeling them outliers if they don’t serve their purpose

poi
03-02-2018, 04:02 AM
of course it does, because a S Asian Bengali or Nepali reference when broken down is actually Indian tribal + W Asian + steppe + E Asian + ..... So the E Asian is included in S Asian. Don’t forget if the 100 SNP haplotype windows have 40 E Asian SNPs, those E Asian SNPs are assigned S Asian by their smoothing algorithm (majority rule)

Pnb mentioned earlier in this thread that even Chettri Nepali are showing 20% East Asian in 23andme. So, atleast for Nepali samples, the smoothing isn't affecting too much. Yet.

Kurd
03-02-2018, 04:10 AM
Pnb mentioned earlier in this thread that even Chettri Nepali are showing 20% East Asian in 23andme. So, atleast for Nepali samples, the smoothing isn't affecting too much. Yet.

I’m sure it’s affecting them also, however the little extra E Asian they have is putting them over the edge. For ex, take one of those 100 SNP windows they use Lets say you have 45 E Asian SNPs and the Chettri has 55 E Asian SNPs. That window will get assigned 100% S Asian for you, whereas for the Chettri it will get assigned 100% E Asian. A couple more SNPs are just enough to push it from 100% S Asian to 100% E Asian.

Add all those windows up. You get the idea...

poi
03-02-2018, 04:16 AM
Iím sure itís affecting them also, however the little extra E Asian they have is putting them over the edge. For ex, take one of those 100 SNP windows they use Lets say you have 45 E Asian SNPs and the Chettri has 55 E Asian SNPs. That window will get assigned 100% S Asian for you, whereas for the Chettri it will get assigned 100% E Asian. A couple more SNPs are just enough to push it from 100% S Asian to 100% E Asian.

Add all those windows up. You get the idea...

Statistics based genetic horoscope. Where did you see this industry headed (prediction wise)?

khanabadoshi
03-02-2018, 04:22 AM
Doesnít hurt to try, but I honestly donít think that will change things much, because of the way their algorithm works. For them Itís almost like trying to guess where you are from with the results forced to confirm with the guess via mathematical manipulation. Having such a huge bullseye such as S Asia or ME makes it easyto guess correctly. Correct guesses = less consumer complaints. Ití All business :)

Edit: I can see them tossing people out who they feel are too mixed between their groupings, or labeling them outliers if they donít serve their purpose


Totally agree. I just think we are not an insignificant contribution to the fact there is a huge Bull's Eye. So, might as well resolve my side of this affair before I launch my consumer complaints haha.

Kaido
03-02-2018, 02:24 PM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

Censored
03-02-2018, 08:41 PM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

16% at 90% confidence.

khanabadoshi
03-02-2018, 10:27 PM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

90% Confidence on Left and 50% Confidence on Right. Some of the unassigned is huge.....





90% CONFIDENCE

South Asian
East Asian
ME/NA
European
Oceanian
Unassigned
50% CONFIDENCE

South Asian
ME/NA
East Asian
European
SSA
Oceanian
Unassigned


Khanabadoshi
69.00%
0.40%



30.60%

Khanabadoshi
94.90%
1.90%
0.90%
1.60%


0.70%


Brother
56.90%
0.10%

0.10%

42.90%

Brother
93.10%
2.90%
0.80%
0.70%


2.40%


Sister
46.80%
0.30%



52.90%

Sister
87.00%
8.40%
1.40%
1.30%
0.10%

1.70%




















"Father"
70.70%
0.30%

0.70%

28.30%

"Father"
95.90%
0.70%
0.80%
1.70%
0.20%

0.70%


"Mother"
71.90%
0.90%
0.10%


27.00%

"Mother"
88.00%
3.60%
3.80%
2.70%


1.90%




















"Maternal Grandfather"
83.20%




16.80%

"Maternal Grandfather"
97.40%
1.20%
0.20%
0.20%

0.20%
0.80%


Maternal Grandmother
13.40%
0.30%

0.70%

85.50%

Maternal Grandmother
57.40%
26.50%
6.60%
5.90%


3.60%


mGM Brother
11.80%
0.60%

1.50%

86.10%

mGM Brother
57.00%
21.30%
10.30%
6.70%


4.70%




















2nd Cousin
64.50%


0.10%

35.40%

2nd Cousin
96.40%
2.20%
0.10%
0.30%
0.10%

0.90%


Mohmand Pashtun
17.00%
0.20%
3.30%
0.40%
0.10%
79.10%

Mohmand Pashtun
70.60%
23.00%
1.80%
1.90%

0.10%
2.50%


Multani
60.30%


0.50%

39.20%

Multani
96.50%

0.80%
2.00%
0.20%

0.60%


Lahori
71.70%
0.10%

0.20%

28.00%

Lahori
95.60%
1.90%
0.50%
1.30%


0.70%

poi
03-02-2018, 10:44 PM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

So far, it looks like people tested on their v5 chip seem to have much lower unassigned at 90% conservative confidence. Mine 90% confidence is only 5% unassigned. bmoney's 90% is just 0.1% unassigned.

Kurd
03-02-2018, 11:24 PM
90% Confidence on Left and 50% Confidence on Right. Some of the unassigned is huge.....





90% CONFIDENCE
South Asian
East Asian
ME/NA
European
Oceanian
Unassigned
50% CONFIDENCE
South Asian
ME/NA
East Asian
European
SSA
Oceanian
Unassigned


Khanabadoshi
69.00%
0.40%



30.60%

Khanabadoshi
94.90%
1.90%
0.90%
1.60%


0.70%


Brother
56.90%
0.10%

0.10%

42.90%

Brother
93.10%
2.90%
0.80%
0.70%


2.40%


Sister
46.80%
0.30%



52.90%

Sister
87.00%
8.40%
1.40%
1.30%
0.10%

1.70%




















"Father"
70.70%
0.30%

0.70%

28.30%

"Father"
95.90%
0.70%
0.80%
1.70%
0.20%

0.70%


"Mother"
71.90%
0.90%
0.10%


27.00%

"Mother"
88.00%
3.60%
3.80%
2.70%


1.90%




















"Maternal Grandfather"
83.20%




16.80%

"Maternal Grandfather"
97.40%
1.20%
0.20%
0.20%

0.20%
0.80%


Maternal Grandmother
13.40%
0.30%

0.70%

85.50%

Maternal Grandmother
57.40%
26.50%
6.60%
5.90%


3.60%


mGM Brother
11.80%
0.60%

1.50%

86.10%

mGM Brother
57.00%
21.30%
10.30%
6.70%


4.70%




















2nd Cousin
64.50%


0.10%

35.40%

2nd Cousin
96.40%
2.20%
0.10%
0.30%
0.10%

0.90%


Mohmand Pashtun
17.00%
0.20%
3.30%
0.40%
0.10%
79.10%

Mohmand Pashtun
70.60%
23.00%
1.80%
1.90%

0.10%
2.50%


Multani
60.30%


0.50%

39.20%

Multani
96.50%

0.80%
2.00%
0.20%

0.60%


Lahori
71.70%
0.10%

0.20%

28.00%

Lahori
95.60%
1.90%
0.50%
1.30%


0.70%







Wow theyíre really struggling with your M GM, her brother, and the Momand Pashtun with 12, 13 and 17% S Asian in 90% confidence mode.

This means that even with their liberal definition of S Asian and their algorithm to maximize it that most of the S Asian for these individuals they are only 50% sure of !

Hopefully their new added regions will help...

jb24
03-02-2018, 11:32 PM
New member here (Jatt Sikh). I have heard of 23andme, but I have no knowledge about such tests or what it could all cost?

MonkeyDLuffy
03-03-2018, 01:39 AM
New member here (Jatt Sikh). I have heard of 23andme, but I have no knowledge about such tests or what it could all cost?

Are you still in Punjab? It's hard to get that in India unless you've a relative in West who can ship it to you. It costs about $129 CAD for ancestry only.

Sapporo
03-03-2018, 01:49 AM
New member here (Jatt Sikh). I have heard of 23andme, but I have no knowledge about such tests or what it could all cost?

Depending on where you're located, FTDNA, 23andMe V5 or Ancestry DNA are your best options. Not sure how much you are willing to spend but imo, it's best to invest into a combination of FTDNA for older Gedmatch calculators and 23andMe V5 for the newer Illumina chip. If you're in the US, FTDNA is $79 for autosomal analysis. You might find slight discounts on Groupon or other sites. 23andMe V5 is $99. Sometimes it can be had for around $79 or less during sales or holiday specials. If you have Amazon prime, I believe 23andMe and Ancestry DNA can be purchased through their site.

https://www.familytreedna.com/products/family-finder

https://www.23andme.com/dna-ancestry/

Tirunelvi
03-03-2018, 05:32 AM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

20% at 90% Confidence, but I'm phased with a child.

itsmrhoax
03-27-2018, 08:40 AM
Since FTDNA is a US site showing only US prices, can I order it from the UK?

misanthropy
03-28-2018, 04:21 AM
What kinda of unassigned percentages are you guys getting? I noticed at 90% conservative confidence I had 28% unassigned, whereas my mother at the same confidence has 65% unassigned, wonder what's going on there.

44.6% unassigned @ 90% confidence (and 55.1% South Asian)

I don't know if it's true, but I asked about those who have higher unassigned percentages (regardless of confidence levels) in a thread on here and someone said that it's probably due to ancestry at the border zones of major 23andme regions. Again, not sure if it's true.

Edit: Off the 23andme site


It is also possible to see a percentage of your DNA listed as “Unassigned.” There are two reasons why a piece of DNA might have unassigned ancestry:

The piece of DNA matches many different populations from around the world.
The piece of DNA does not match any of the reference populations very well.

Tirunelvi
04-03-2018, 08:04 PM
Results from the new 23andme update. I wonder how diverse the backgrounds of the Indians on the reference panel are.
22475
22476

misanthropy
04-03-2018, 09:31 PM
About time!

I got no matches at all :(

https://i.imgur.com/0Hp9t2q.png

Censored
04-04-2018, 03:39 AM
About time!

I got no matches at all :(

https://i.imgur.com/0Hp9t2q.png

It's absurd. What are these people doing?

22495
22496

khanabadoshi
04-04-2018, 04:03 AM
About time!

I got no matches at all :(



Just checked mine, after seeing your post -- same thing for me:

https://i.gyazo.com/07800b31bc600659041415c9a04f2cef.png

misanthropy
04-04-2018, 05:02 AM
Just checked mine, after seeing your post -- same thing for me:


Yeah, I'm kinda disappointed because I wanted to narrow down my 'Western Asian' to the closest country, so I can get an idea if it's from my paternal or maternal side. My current theory is that it's mostly from my maternal side with the older Persian Sufi descent (as opposed to my paternal Hadhrami component, which is geographically much further away).

Yours was Uzbekistan though, right? Edit: I see your Uzbek and the Iranian flag.

Zaid
04-04-2018, 08:18 PM
Same here. Frustating, but oh well...

https://i.imgur.com/SPDQ7GM.png

Hanna
04-05-2018, 05:20 PM
Posted in the wrong thread. Just noticed this thread is for South Asians but I'll share my updated results anyway. Nothing new.

Western Asian
85.2% Turkey
Broadly Western Asian & North African 0.2%

European 12.8%
Italian 2.3%
Balkan 1.4%
Broadly Southern European 5.2%
Broadly European 3.9%

East Asian & Native American 1.3%
Siberian 0.6%
Manchurian & Mongolian 0.2%
Broadly East Asian 0.5%

Unassigned

22546
22545

poi
04-06-2018, 01:52 AM
They updated my breakdown from 100% South Asian to 100% Nepal. Lol

22551

Censored
04-06-2018, 06:14 AM
Just checked mine, after seeing your post -- same thing for me:

https://i.gyazo.com/07800b31bc600659041415c9a04f2cef.png

You got a surprisingly high amount of South Asian given your family history.

Myth
04-08-2018, 05:36 AM
Half Ashkenazi dads side and half German moms side on new results..kind of annoyed no new information with update.

22553

Sapporo
04-08-2018, 08:20 AM
Got the email from 23andMe regarding new updates but no changes to my ancestry composition or my father's. My mother's account has no changes either.

Kaido
04-08-2018, 12:46 PM
No changes to the composition.
https://i.imgur.com/qXRHyif.png

https://i.imgur.com/BoIsP71.png

khanabadoshi
04-09-2018, 12:59 AM
No changes to the composition.


https://i.imgur.com/BoIsP71.png

The Mohmand got some assignments:

https://i.gyazo.com/2031df90d429b9bd0346f58adeda8f00.jpg

My paternal Uncle - .jam , representative of my Father also got assignments:

https://i.gyazo.com/7f819449a3e73c2c4fa5ae7197c9c3ff.jpg

My maternal grandfather's sister - .sadia, representative of my maternal Grandfather:

https://i.gyazo.com/465d105bd164b87e179012f18a3c9a95.jpg

The Multani - Syed/Durrani:

https://i.gyazo.com/d87387f1f3e3bf9f82f68fb8301f10ba.jpg

Other than that, no one got assignments. Of the ones that have some matches, only my Uncle got assignments in any group outside of South Asia (the European section).
Interesting that everyone matched more strongly with Afghanistan than Pakistan, and no one has a match with India except the Mohmand (weakly). Very strange.

Saad2016
04-09-2018, 02:02 AM
No changes to the composition.

what does these four and five dots indicated?

khanabadoshi
04-09-2018, 03:57 AM
what does these four and five dots indicated?

Which population you are matching closest to in their reference.

Saad2016
04-09-2018, 10:29 AM
Which population you are matching closest to in their reference.

reference population base don their own data base? say like 5 people declared themselves as pakistanies and undergone 23andme test, 23andme added them to their database of references now when a 6th person takes the same test 2 months down the road then depending on how much he resembles the first 5 he will be given dots?

Magnetic
04-09-2018, 12:12 PM
edit: sorry I have just seen that this thread is meant for south asian members by OP

anyway too late :p my results :

http://up.picr.de/32332993hx.jpg

http://up.picr.de/32333024vy.jpg

Mingle
04-09-2018, 08:40 PM
https://i.imgur.com/NQcX8cI.png?1

https://i.imgur.com/67MoOhN.png?1

khanabadoshi
04-10-2018, 01:46 AM
My Brother got a weak match.

https://i.gyazo.com/8a2f9a6c11ca39cb95f3850fc885ae8f.png

Zaid
04-10-2018, 05:01 AM
Will the bubbles update with time? All of mine are still empty.

khanabadoshi
04-10-2018, 05:26 AM
Will the bubbles update with time? All of mine are still empty.

It seems so. Just 3 days ago, none of my kits had any bubbles filled. So they are updating slowly.

poi
04-10-2018, 05:43 AM
It seems so. Just 3 days ago, none of my kits had any bubbles filled. So they are updating slowly.

Mine don't even have empty bubbles. Not that I expect things to change.

https://i.imgur.com/givepFA.png

vettor
04-10-2018, 05:46 AM
5 dots only for italian for 27% ............I wonder if they have italian and austrian tyrol numbers ?????

https://s20.postimg.org/a0qs5x3a5/andy_5_dots.jpg (https://postimages.org/)

Censored
04-10-2018, 06:37 AM
Mine don't even have empty bubbles. Not that I expect things to change.

https://i.imgur.com/givepFA.png

How does a Nepali Brahmin get 100% South Asian when even I can't? Incredible. Lol

Tirunelvi
04-10-2018, 04:03 PM
No East European


The Mohmand got some assignments:

https://i.gyazo.com/2031df90d429b9bd0346f58adeda8f00.jpg

My paternal Uncle - .jam , representative of my Father also got assignments:

https://i.gyazo.com/7f819449a3e73c2c4fa5ae7197c9c3ff.jpg

My maternal grandfather's sister - .sadia, representative of my maternal Grandfather:

https://i.gyazo.com/465d105bd164b87e179012f18a3c9a95.jpg

The Multani - Syed/Durrani:

https://i.gyazo.com/d87387f1f3e3bf9f82f68fb8301f10ba.jpg

Other than that, no one got assignments. Of the ones that have some matches, only my Uncle got assignments in any group outside of South Asia (the European section).
Interesting that everyone matched more strongly with Afghanistan than Pakistan, and no one has a match with India except the Mohmand (weakly). Very strange.
Really your uncle got assignments for the European section? Which country?.

Sapporo
04-10-2018, 04:22 PM
I get 4 dots with India. No other dots whatsoever. My father gets 2 dots with India. Nothing else.

What are the country assignments even based upon? The countries that you have the option to list your grandparents are from? That's the only thing I could think of since I sometimes see Jatt Sikh matches on 23andMe show list India for each of their 4 grandparents. Very rarely, some list Pakistan for one of their grandparents.

Edward J
04-10-2018, 06:42 PM
22588

poi
04-10-2018, 07:36 PM
How does a Nepali Brahmin get 100% South Asian when even I can't? Incredible. Lol

Lol! To be fair, I'm on v5 chip. PNB is same ethnic group as me and he doesn't get 100% either. He is on v4.

Saad2016
04-10-2018, 08:09 PM
22588

Interesting so it in agreement with your real ancestry? are u from Slovenia?

bmoney
04-11-2018, 01:31 AM
Lol! To be fair, I'm on v5 chip. PNB is same ethnic group as me and he doesn't get 100% either. He is on v4.

Although all three of my 23andMe chip versions' ancestry reports are similar to each other, my new 23andMe v5 test is clearly the most consistent with my Doug McDonald BGA analysis (the Standard) and is more specific than my v3 and v4 Ancestry Composition reports. My smaller ancestry components like Native American, Southeast Asian, Ashkenazi Jewish and Oceania holds good at 90% confidence so are likely real affinities. This bodes well for my family genealogical research. My 23andMe v5 results did struggle with my European admixture, but if we consider 23andMe v5 and Illumina's promises of ancestry improvements with this GSA chip I still feel like I hit the jackpot. In this regard 23andMe v5 retains its title as the world's best ethnicity admixture test with the most cutting-edge chromosome painting tool on the market.

http://www.rootsandrecombinantdna.com/2018/01/23andme-chip-versions-comparison.html

I'm happy using v5 as it better detects chromosome level minor admixture, and as shown by the above blog post, seems to be better for mixed people (like South Asians)

23andme has claimed that the v5 chip only looks at markers that reliably predict ancestry whereas other chips/companies might be using non-ancestry specific markes to assume ancestry

The only drawback is the low overlap with current calc SNPs, but calcs designed for v5 would be more accurate than current calcs for South Asians IMO from purely an ancestral perspective

Edward J
04-11-2018, 12:18 PM
Interesting so it in agreement with your real ancestry? are u from Slovenia?

I have a grandfather with full Slovenian ancestry and a grandmother with full Italian ancestry. My next closest recent ancestry is a German/Bohemian great-grandmother and a gg grandmother from England. The rest of my ancestry is mostly Colonial American. (mix of German, UK, and Irish).

So I would say it matches up quite well with my paper ancestry. Also in coordinating with some of the Slovenian groups that I belong to on FB, it seems that it does a good job at picking up the Slovenian element for others with partial ancestry from there.

jb24
04-15-2018, 12:20 AM
deleted

jb24
04-15-2018, 12:27 AM
Are you still in Punjab? It's hard to get that in India unless you've a relative in West who can ship it to you. It costs about $129 CAD for ancestry only.

I am in UK. Thank you for the info.

jb24
04-15-2018, 12:32 AM
Depending on where you're located, FTDNA, 23andMe V5 or Ancestry DNA are your best options. Not sure how much you are willing to spend but imo, it's best to invest into a combination of FTDNA for older Gedmatch calculators and 23andMe V5 for the newer Illumina chip. If you're in the US, FTDNA is $79 for autosomal analysis. You might find slight discounts on Groupon or other sites. 23andMe V5 is $99. Sometimes it can be had for around $79 or less during sales or holiday specials. If you have Amazon prime, I believe 23andMe and Ancestry DNA can be purchased through their site.



Thank you for the info. I will decide which company to test with or it might be worth testing with both FTDNA and 23andMe?

parasar
04-15-2018, 04:13 AM
Broadly South Asian 96.9%
But no circles filled in at present.

agent_lime
04-26-2018, 11:04 AM
So I might be the only South Asian here that got Ancestry.com data onto 23&me free offer.

This shows me the highest south asian out of any calculator I've tried.

https://i.imgur.com/1Z0YchZ.png

https://i.imgur.com/ht6Ig2d.png

https://i.imgur.com/zWcjGSt.png

Ancestry results for comparison-

https://i.imgur.com/jPMdLRZ.png

jortita
04-27-2018, 03:45 AM
22820

Ancestry DNA uploaded to 23andMe results

South Asian 85.2%

Broadly South Asian 85.2% (all India)

East Asian and Native American 11%

Southeast Asian 6.1%

Manchurian and Mongolian 1.2%

Broadly East Asian 1.7%

Broadly East Asian and Native American 2%

European 1.2%

Finnish 0.1%

Broadly Northwestern European 0.2%

Broadly European 0.9%

Western Asian and North African 0.2%

Western Asian 0.2%

Unassigned 2.5%

I personally think 23andMe underestimates the East Asian of South Asians, an example being JF Winstone's mother and I score practically similar East Asian ancestry on gedmatch calculators, however she scored 22.2% East Asian on 23andMe. I have a feeling my East Asian ancestry % might be higher when I test with 23andMe, which I am not able to as I am based in Thailand

FrostAssassin0701
05-05-2018, 07:16 PM
Just 100% South Asian. I had <0.1% East Asian but they got rid of it because it was below 0.05%.