PDA

View Full Version : Breaking down Polynesian DNA components using G25 modelling



BalkanKiwi
04-06-2020, 11:42 PM
I’ve been playing around a bit with Poi’s Runner and modelling various Southeast Asian and Oceania populations to see if I can create a migration map of sorts, tracing my minor Polynesian through Southeast Asia to New Zealand. I must admit my understanding of the statistics and analysis associated with G25 coordinates and the interpretation is not as good as many of you here, so please correct me if I’m wrong on anything.

To begin with on paper my Māori ancestry is very minor (around 1%), however I want to see if it’s possible to model components of Polynesian DNA to see if I have trace amounts within that of the areas Polynesian’s migrated through over several thousands of years. For those of you not overly familiar with Polynesian and Māori migration, here is a quote from this article (https://www.culturalsurvival.org/news/genetic-link-brings-indigenous-taiwanese-and-maori-together) which sums it up well:


Thousands of years later (8,000 to 10,000 years ago) members of the Austronesian language group traveled south from Taiwan and passed through the Philippines and Indonesia. Along the way the Papuans and Austronesians intermarried, giving birth to contemporary Polynesians and ultimately settling in what is now New Zealand. "The Maori are the ultimate end product of these voyage and colonization processes," said Chambers, adding that "a ‘genetic trail’ links New Zealand Maori back all the way to the indigenous people of Taiwan, as do Austronesian languages.

Here is a map using data from this article (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4143916/) that show’s the migration path’s in Asia and Oceania, and the ancestral components of each population. Polynesian’s are mostly Austronesian and Papuan.

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

Besides Māori, I went and swapped it out for every Austronesian and Melanesian population in the G25 calculator to see how interchangeable Māori is. I don’t match every population, but I match a few from Philippines, Indonesia and Papua New Guinea. My highest percentage is Nasoi (1.67%) from Bougainville Island, one of the islands of Papua New Guinea. I get this using a different model which I haven't listed here, as its fit is slightly higher (1.1635). This is shown in the map from ph2ter.

https://i.imgur.com/8usOYuV.png

Below I’ve listed each population I match, with the closet fit from top to bottom, and the distance to each sample in that population. Interestingly (but maybe not surprising), the 0.83% I get for Māori is the same for every Austronesian and Papuan population I match (besides Nasoi).

Nasoi ("fit": 1.374) (Papua New Guinea)


"fit": 1.374,
"Irish": 53.33,
"English_Cornwall": 30,
"Croatian": 15,
"Nasoi": 1.67
• Nasoi:HGDP01027: 60.326195
• Nasoi:HGDP00661: 60.598324
• Nasoi:HGDP00655: 60.694556


Kosipe ("fit": 1.0082) (Papua New Guinea)

"fit": 1.0082,
"Irish": 66.67,
"Serbian": 18.33,
"English": 14.17,
"Kosipe": 0.83,
• Kosipe:GS000035365: 78.294494


Koinanbe ("fit": 1.0122) (Papua New Guinea)

"fit": 1.0122,
"Irish": 67.5,
"Serbian": 19.17,
"English": 12.5,
"Koinanbe": 0.83,
• Koinanbe:GS000035255: 80.201681


Papuan ("fit": 1.0134) (Papua New Guinea)

"fit": 1.0134,
"Irish": 67.5,
"Serbian": 19.17,
"English": 12.5,
"Papuan": 0.83,
• Papuan:HGDP00541: 78.968809
• Papuan:HGDP00542: 79.233931
• Papuan:HGDP00545: 79.931080
• Papuan:HGDP00540: 81.205769
• Papuan:HGDP00552: 81.580418



Australian ("fit": 1.3593)

"Irish": 52.5,
"English_Cornwall": 30,
"Croatian": 15.83,
"Australian": 1.67,
• Australian:B_Australian-4: 63.590219
• Australian:B_Australian-3: 64.015839


Maori ("fit": 1.5418)

"fit": 1.5418,
"Irish": 52.5,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Maori": 2.5,
• "Maori:S_Maori-1: 28.244222"

Agta ("fit": 1.5637) (Philippines)

"fit": 1.5467,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Agta": 1.67,
• Agta:GS000035435: 51.639106


Aeta ("fit": 1.1015) (Philippines)

"fit": 1.1015,
"Irish": 55.83,
"Scottish": 24.17,
"Serbian": 19.17,
"Aeta": 0.83,
• Aeta:GS000035250: 51.618677
• Aeta:GS000035257: 53.133173


Batak ("fit": 1.5637) (Indonesia)

"fit": 1.5637,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Batak": 1.67
• Batak:GS000035252: 54.016611
• Batak:GS000035253: 55.276758


Bajo ("fit": 1.1848) (Philippines)

"fit": 1.1848,
"Irish": 69.17,
"Serbian": 19.17,
"English": 10.83,
"Bajo": 0.83,
• Bajo:GS000017004: 56.817300
• Bajo:GS000017005: 57.990166
• Bajo:GS000017006: 58.470105


Indonesian Bali ("fit": 1.2154)

"fit": 1.2154,
"Irish": 70,
"Serbian": 19.17,
"English": 10,
"Indonesian_Bali": 0.83,
• "Indonesian_Bali:GRC10041153: 57.677006",
• "Indonesian_Bali:GRC10041154: 57.722013",
• "Indonesian_Bali:GRC10041174: 57.863440",
• "Indonesian_Bali:GRC10041162: 57.899742",
• "Indonesian_Bali:GRC10041168: 57.982084",
• "Indonesian_Bali:GRC10041159: 58.022552",
• "Indonesian_Bali:GRC10041166: 58.056610",
• "Indonesian_Bali:GRC10041170: 58.192115",
• "Indonesian_Bali:GRC10041161: 58.352414",
• "Indonesian_Bali:GRC10041167: 58.486995",
• "Indonesian_Bali:GRC10041171: 58.534598",
• "Indonesian_Bali:GRC10041155: 58.582123"


Malayan ("fit": 1.5)

"fit": 1.5,
"Irish": 53.33,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Malayan": 1.67,


I’ve done a quick diagram to map out the migration path using these results.

https://i.imgur.com/nHo8ANXh.jpg

Its interesting to see that G25 coordinates can be useful in this way, even for very minor components. It would be great to see what this would look like for someone with more substantial Polynesian DNA. I can only surmise that the 0.83% is the same for all of them because they are interchangeable in a sense. You can’t have Polynesian without Papuan, Indonesian or indigenous Filipino. I’m not sure why I’m closer to Papuan populations. I assume I inherited more of this component. I hope whoever reads this finds it interesting. If you have any comments or if I've have made errors, please correct me.

Osiris
04-07-2020, 07:08 PM
Very cool. Austronesian languages and their descendants of the earliest known speakers always seemed so ripe for studying.

BalkanKiwi
04-07-2020, 07:55 PM
Very cool. Austronesian languages and their descendants of the earliest known speakers always seemed so ripe for studying.

Agreed. I see you're part Hawaiian. Unfortunately there aren't many members here with Polynesian ancestry, so a welcome to the forum!

Ideally we would have more samples from places like New Caledonia, Cook Islands, Tahiti etc to fill in the gap between Papua New Guinea and the final stop (New Zealand, Hawaii etc).

Osiris
04-08-2020, 12:25 AM
I would love for someone to do a full on study of the Austronesians and their descendants. I'm always hoping the University of Hawaii will do something. Maybe partner with other universities throughout the region. Seems like such a slam dunk to take that and combine it with archaeology and have something profound and important to say.

BalkanKiwi
04-08-2020, 01:18 AM
I would love for someone to do a full on study of the Austronesians and their descendants. I'm always hoping the University of Hawaii will do something. Maybe partner with other universities throughout the region. Seems like such a slam dunk to take that and combine it with archaeology and have something profound and important to say.

I'm surprised they haven't already thought of this. If the University of Hawaii were to team up with the University of Auckland and the University of Fiji for example, a lot of good work can be done. I've briefly played around with the new Chinese samples David has added to the G25, and I pick up trace amounts for various populations (0.4-0.6%). There is a lot of potential to do a detailed study which could be linked partially back to ancient China.

Bealfire
04-08-2020, 01:22 AM
I think it would be a great thing to have. Amazing also how people from Lapita culture absorbed Melanesian ancestry before sailing eastwards and Melanesians absorbed Lapita ancestry and adopted the Malayo-Polynesian language .

BalkanKiwi
04-08-2020, 02:37 AM
I think it would be a great thing to have. Amazing also how people from Lapita culture absorbed Melanesian ancestry before sailing eastwards and Melanesians absorbed Lapita ancestry and adopted the Malayo-Polynesian language .

I was taught both Maori and Indonesian in primary school (Maori in New Zealand, Indonesian in Australia). I have a very basic understanding of both languages, but there's clear similarities between various words, i.e. 2 in Maori is Rua, and in Indonesian its Dua, 5 in Maori is Rima, and in Indonesian its Lima.

BalkanKiwi
05-20-2020, 10:18 AM
This is some quick modelling I did with the ancient Vanuatu and Tongan samples:

"fit": 1.653,
"Irish": 49.17,
"English": 29.17,
"Croatian": 20,
"TON_2500BP": 1.67,
"closestDistances":
"English:HG01790: 2.173339",
"Irish:Irish12: 2.683066",
"Croatian:Croatia_Cro140: 4.483113",
"TON_2500BP:Sk10: 24.904264"

"fit": 1.6581,
"Irish": 48.33,
"English": 30.83,
"Croatian": 19.17,
"VUT_2900BP_all": 1.67,
"closestDistances":
"English:HG01790: 2.173339",
"Irish:Irish12: 2.683066",
"Croatian:Croatia_Cro140: 4.483113",
"VUT_2900BP_all:I1370_all: 25.022574"

Hurricane
05-27-2020, 10:29 PM
Just playing around


"fit": 1.9363,
"Berber_Algeria": 65.83,
"Kikuyu": 26.67,
"Yoruba": 5.83,
"Australian": 1.67,

"fit": 1.9668,
"Berber_Algeria": 65,
"Kikuyu": 27.5,
"Yoruba": 5.83,
"VUT_500BP_all": 1.67,

"fit": 1.9871,
"Berber_Algeria": 65.83,
"Kikuyu": 25.83,
"Yoruba": 6.67,
"Nasoi": 1.67,

"fit": 1.9875,
"Berber_Algeria": 66.67,
"Kikuyu": 25.83,
"Yoruba": 5.83,
"VUT_1200BP": 1.67

"fit": 1.99,
"Berber_Algeria": 64.17,
"Kikuyu": 26.67,
"Yoruba": 7.5,
"VUT_1100BP": 1.67,

"fit": 2.0027,
"Berber_Algeria": 63.33,
"Kikuyu": 26.67,
"Yoruba": 8.33,
"Kosipe": 1.67

"fit": 2.0529,
"Berber_Algeria": 65.83,
"Kikuyu": 26.67,
"Yoruba": 5.83,
"Indonesian_Java": 1.67,
"fit": 2.0671,
"Berber_Algeria": 62.5,
"Kikuyu": 26.67,
"Yoruba": 9.17,
"Koinanbe": 1.67

"fit": 2.0853,
"Berber_Algeria": 62.5,
"Kikuyu": 27.5,
"Yoruba": 8.33,
"Indonesian_Bali": 1.67,

"fit": 2.1135,
"Berber_Algeria": 65,
"Kikuyu": 25.83,
"Yoruba": 7.5,
"Aeta": 1.67,

"fit": 2.1244,
"Berber_Algeria": 62.5,
"Kikuyu": 27.5,
"Yoruba": 8.33,
"Maori": 1.67,

"fit": 2.1248,
"Berber_Algeria": 67.5,
"Kikuyu": 26.67,
"Yoruba": 5,
"Hawaiian": 0.83,

"fit": 2.1643,
"Berber_Algeria": 62.5,
"Kikuyu": 28.33,
"Yoruba": 7.5,
"TON_2500BP": 1.67,]

I would have loved to simulate with the Alorese,Pantarese and Timorese samples available on the K36 runner

BalkanKiwi
05-28-2020, 09:41 AM
Just playing around



I would have loved to simulate with the Alorese,Pantarese and Timorese samples available on the K36 runner

That's interesting. As I've mentioned elsewhere, I'm not greatly knowledgeable in African genetics. Is it normal for North Africans to pick up small amounts of Oceanian? It's as shame those samples aren't in the G25. Does the G25 have any populations that are similar to those?

Since my initial post, I've been playing around with the FreeForm nMonte (all of those models were done with Poi's Runner). I've found doing FreeForm works better, and leads to a slightly higher percentage if I use English Cornwall instead of English. Here are a few (I believe the Maori sample isn't full Maori). I've played around with modern Chinese populations, and Dongxiang scores the highest with 1.67%. Its also the darkest colored area in China on ph2ter's map. Given its proximity to Taiwan, I'm not sure if its suggestive of migration into the area.



"fit": 1.5418,
"Irish": 52.5,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Maori": 2.5,

"fit": 1.3593,
"Irish": 52.5,
"English_Cornwall": 30,
"Croatian": 15.83,
"Australian": 1.67,

"fit": 1.6359,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 13.33,
"VUT_2900BP_all": 2.5,

"fit": 1.6224,
"Irish": 53.33,
"English_Cornwall": 26.67,
"Croatian": 17.5,
"TON_2500BP": 2.5,

"fit": 1.374,
"Irish": 53.33,
"English_Cornwall": 30,
"Croatian": 15,
"Nasoi": 1.67,

"fit": 1.5467,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Agta": 1.67,

"fit": 1.5637,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Batak": 1.67,

"fit": 1.5,
"Irish": 53.33,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Malayan": 1.67,

"fit": 1.6562,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Dongxiang": 1.67,

Ebizur
05-28-2020, 11:47 AM
That's interesting. As I've mentioned elsewhere, I'm not greatly knowledgeable in African genetics. Is it normal for North Africans to pick up small amounts of Oceanian? It's as shame those samples aren't in the G25. Does the G25 have any populations that are similar to those?

Since my initial post, I've been playing around with the FreeForm nMonte (all of those models were done with Poi's Runner). I've found doing FreeForm works better, and leads to a slightly higher percentage if I use English Cornwall instead of English. Here are a few (I believe the Maori sample isn't full Maori). I've played around with modern Chinese populations, and Dongxiang scores the highest with 1.67%. Its also the darkest colored area in China on ph2ter's map. Given its proximity to Taiwan, I'm not sure if its suggestive of migration into the area.



"fit": 1.5418,
"Irish": 52.5,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Maori": 2.5,

"fit": 1.3593,
"Irish": 52.5,
"English_Cornwall": 30,
"Croatian": 15.83,
"Australian": 1.67,

"fit": 1.6359,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 13.33,
"VUT_2900BP_all": 2.5,

"fit": 1.6224,
"Irish": 53.33,
"English_Cornwall": 26.67,
"Croatian": 17.5,
"TON_2500BP": 2.5,

"fit": 1.374,
"Irish": 53.33,
"English_Cornwall": 30,
"Croatian": 15,
"Nasoi": 1.67,

"fit": 1.5467,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Agta": 1.67,

"fit": 1.5637,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Batak": 1.67,

"fit": 1.5,
"Irish": 53.33,
"English_Cornwall": 27.5,
"Croatian": 17.5,
"Malayan": 1.67,

"fit": 1.6562,
"Irish": 52.5,
"English_Cornwall": 31.67,
"Croatian": 14.17,
"Dongxiang": 1.67,
The Dongxiangs are Muslims who inhabit some parts of Gansu Province in northwestern China and who speak a Mongolic language of a peculiar sort, possibly (though not necessarily) altered through the influence of some substrate language. They are notable for a high degree of male-mediated admixture from Western Eurasia. A sample of Dongxiangs studied by Dan Xu and Shaoqing Wen in "The Silk Road: language and population admixture and replacement" included only the following Eastern Eurasian Y-DNA haplogroups: 6.54% O2a1c-JST002611, 6.54% C2-M217(xM93, P39, M48, M407, P53.1), 5.61% O2a2b1-M134(xM117), 4.67% O2a2b1a1-M117, 2.80% O1b1a1a-M95, 2.80% N1a2b-P43, 2.80% D1a2a-P47, 1.87% D1a1-M15, 1.87% NO-M214(xO-M175, LLY22g), 1.87% O2a2-P201(xM7, P164), 1.87% O1a1-P203, 1.87% Q1a1-M120/N14, 0.93% Q1a-MEH2(xM120, M25, M346), 0.93% N1a1a-M178, and 0.93% O1b-M268/P31(xPK4, M176) for a total of 43.91% Y-DNA of Eastern Eurasian origin or affinity. (Of course, the Q subclades may also be considered to be ultimately of Western Eurasian origin.)

The remainder of this sample of Dongxiang males belonged to J-M304 (16.82%), R1a-M17 (14.02%), R2a-M124 (9.35%), G-M201 (5.61%), E-M96 (3.74%), H-M69 (1.87%), R1b1b1-M73 (1.87%), R1b1b2-M269 (0.93%), I (0.93%), and L (0.93%).

BalkanKiwi
05-28-2020, 08:10 PM
The Dongxiangs are Muslims who inhabit some parts of Gansu Province in northwestern China and who speak a Mongolic language of a peculiar sort, possibly (though not necessarily) altered through the influence of some substrate language. They are notable for a high degree of male-mediated admixture from Western Eurasia. A sample of Dongxiangs studied by Dan Xu and Shaoqing Wen in "The Silk Road: language and population admixture and replacement" included only the following Eastern Eurasian Y-DNA haplogroups: 6.54% O2a1c-JST002611, 6.54% C2-M217(xM93, P39, M48, M407, P53.1), 5.61% O2a2b1-M134(xM117), 4.67% O2a2b1a1-M117, 2.80% O1b1a1a-M95, 2.80% N1a2b-P43, 2.80% D1a2a-P47, 1.87% D1a1-M15, 1.87% NO-M214(xO-M175, LLY22g), 1.87% O2a2-P201(xM7, P164), 1.87% O1a1-P203, 1.87% Q1a1-M120/N14, 0.93% Q1a-MEH2(xM120, M25, M346), 0.93% N1a1a-M178, and 0.93% O1b-M268/P31(xPK4, M176) for a total of 43.91% Y-DNA of Eastern Eurasian origin or affinity. (Of course, the Q subclades may also be considered to be ultimately of Western Eurasian origin.)

The remainder of this sample of Dongxiang males belonged to J-M304 (16.82%), R1a-M17 (14.02%), R2a-M124 (9.35%), G-M201 (5.61%), E-M96 (3.74%), H-M69 (1.87%), R1b1b1-M73 (1.87%), R1b1b2-M269 (0.93%), I (0.93%), and L (0.93%).

Yes, a shared Western Eurasia ancestor makes more sense. The other Chinese populations I only get noise levels for (0.83%), so your answer seems more logical than through Polynesian migration.

Hurricane
06-14-2020, 09:42 PM
That's interesting. As I've mentioned elsewhere, I'm not greatly knowledgeable in African genetics. Is it normal for North Africans to pick up small amounts of Oceanian? It's as shame those samples aren't in the G25. Does the G25 have any populations that are similar to those?


No there are no other similar populations.

I only saw 2 NA who scored about 1% oceanian on AncestryDNA. That's not groundbreaking.Must be noise.

BalkanKiwi
06-15-2020, 09:33 AM
No there are no other similar populations.

I only saw 2 NA who scored about 1% oceanian on AncestryDNA. That's not groundbreaking.Must be noise.

Indeed. The only was to be sure otherwise would be to see if there are any Oceanian like segments on a chromosome painting, and/or see if you get any Polynesian matches, but its unlikely with around 1% or less (as in my case).

uflakmoon
06-20-2020, 12:03 AM
Does anyone know how Malagasy fit into this? Are they Austronesian or other Asian?

BalkanKiwi
06-20-2020, 05:56 AM
Does anyone know how Malagasy fit into this? Are they Austronesian or other Asian?

The Malagasy are mostly a mix of Austronesian and East African.