Thursday, April 6, 2017

Finally! A Gedmatch Admixture Guide!

For those unaware, Gedmatch.com is a website where you can upload your raw DNA data for further analysis and matching with people from other companies who have also upload their data.

Parts 3 and 4 on Admixture Proportions by Chromosome and Chromosome Painting now available.

Part 1 - Admixture Proportions

Introduction
Despite all the help articles available on Gedmatch.com, none of them really offer a comprehensive guide to understanding the admixture calculators for newbies. Most of them are guides on understanding DNA in general, or how to upload your data, or using the one-to-many or one-to-one tools. In fact, there is a very good beginners guide to the matching side of things found here. But the most common questions I see about Gedmatch are “which admixture calculator do I use?” and “what do the results mean?” There is a Gedmatch wiki page on admixture, and there is Kitty Cooper's slide presentation, but I don’t think they really answer all the questions most people are looking for, especially regarding Oracle. Even Googling the topic only turns up spotty results from forums and blogs, nothing that really lays it all out. Since no one else has done it, here is my attempt. Please keep in mind I am no expert and have no formal education in genetics, this is just the knowledge I’ve gathered over the years from various sources as a result of trying to understand my own DNA results.

Admixture is a scientific term for the ethnicity percentages you received from a DNA company like Ancestry.com, FamilyTreeDNA, 23andMe, or MyHeritage. It’s important to understand that each admixture project on Gedmatch is created by a different person, mostly academics. Note that most of the admixture results will include some basic info on the calculator, either on the results page, or through a link from the creator. However, the info provided may still be technical and difficult to understand for the average person, because they were primarily created for academic purposes. This is an attempt to translate some of that info into something more understandable to the average user. I apologize that this guide favors info on European backgrounds, but that is simply what I’m most familiar with, being a European descendant myself.

Be aware that it’s common practice in DNA admixtures to refer to populations from prehistoric times as “ancient”, even though this is a bit of a misnomer. In historical terms, ancient history marks the beginning of recorded history, but here, “ancient” generally refers to the time before written history, prehistory. Some time periods might be specified as “neolithic”, or “paleo/paleolithic” etc.

Select a project from the drop down menu, leaving the other
options as they are, then click "continue"
Step 1: Pick a project.
There are 7 projects to choose from in the Admixture (Heritage) tool (found under "Analyze your data" and "DNA raw data"), but what are they? What do they mean? Which one should you pick? Here’s a basic breakdown:

(Note: below the projects drop down menu there are options like "Admixture Proportions (with link to Oracle)" and "Chromosome Painting", etc. Don't mess with those for now, just stick with the top default option, Admixture Proportions (with link to Oracle), as that is what this guide will cover.)

  1. MDLP
This is a global calculator and attempts to break your results down into different parts of the world. It’s good as an overview, but if, for example, you already know you’re European, it’s probably unnecessary. It’s also heavy on ancient groups. The blog for this project is found here: http://magnusducatus.blogspot.com/

  1. Eurogenes
As the name suggests, this is primarily for people with European backgrounds. While it does have populations outside Europe, there are usually more sub-continental regions for Europe than any other continent. I highly recommend this as the go-to project for people with sole European ancestry. The blog for this project is found here: http://bga101.blogspot.com.au/

  1. Dodecad
This project says it focuses primarily on Eurasians, but most of the calculators are geared more towards Asian and African ancestry than European. It’s not ideal for Europeans, but may be useful for people with mixed ancestry. The blog for this project can be found here: http://dodecad.blogspot.com/

  1. HarappaWorld
This calculator is primarily for people with South Asian ancestry. The blog for this project can be found here: http://www.harappadna.org/

  1. Ethiohelix
This is an African based project, though it does have options for people with mixed backgrounds (but always including African). There is no Native American in this project at all. The blog for this project is found here: http://ethiohelix.blogspot.com/

  1. puntDNAL
This is primarily a project on ancient DNA. There is no website, but questions and comments about should be directed to Abdullahi Warsame at puntdnalking@gmail.com

  1. GedrosiaDNA
This project focuses primarily Eurasian (especially Indian and Asian) and ancient DNA. There is no website, but for further questions, please contact the creator at Dilawerkh4@gmail.com


Once you've selected a project, you need to enter your kit
number and then select a specific calculator.
Step 2: Pick a calculator.
You’ll find that for each project, there are often several calculators to choose from. How to choose? What do they mean? What are the differences? Well, for starters, the numbers following a ‘K’ indicate how many populations (or regions/categories) that calculator includes. So for example, Eurogenes EUtest V2 K15 has 15 populations. So choose one depending on how many regions you want to break your results down into. Keep in mind the more populations and therefore the more specific the regions are, the more speculative the results will be.

Don't forget to put in your kit number - if you've forgotten it, go back to the home page and copy it.

Certain other tests may be specific to deeper, more ancient (prehistoric) ancestry, like Hunter-Gatherer vs Farmer. Any abbreviation that starts with ‘A’ probably stands for ‘ancient’, but I will post a comprehensive terminology list at the end of this guide. These calculators for ancient DNA aren’t very useful if you’re just looking for an opinion on your more recent ethnicity results.

Other calculators might be specific to certain types of ancestry. For example, Eurogenes’ Jtest is specific to Ashkenazi Jewish ancestry. There’s no need to run this test if you don’t have any Jewish ancestry. In fact, you might get false results in Ashkenazi if you run this calculator and have no Jewish ancestry.

(Note: ignore the option below the calculator drop down menu, this is for data collection purposes. If all 4 of your grandparents are from the same ethnic group and you want your DNA to be a part of the sample groups they use to create these calculators and determine populations, then go ahead and fill it out. Otherwise, you can ignore it.)

Here’s a more detailed breakdown of each calculator.

MDLP
  • MDLP K11 Modern - 11 global populations including ancient
  • MDLP K16 Modern - 16 global populations including ancient and modern - results page includes full population descriptions
  • MDLP K23b - 23 global populations including ancient
  • MDLP World22 - 22 global populations including ancient, full details including maps of what areas each category covers are found here - there are several Native American categories so this may be ideal for Native American ancestry
  • MDLP World - 12 global populations, probably the original MDLP calculator

Eurogenes
  • Eurogenes K13 - 13 global populations, mostly European. Creator made this the default as it “seems to hit the spot for most people” with European background. Details here
  • Eurogenes EUtest V2 K15 - 15 global populations, mostly European, also a popular option. Details including regional maps for each category found here
  • Eurogenes ANE K7 - 7 populations, Ancient North Eurasian, meaning this looks at ancient DNA mostly in Europe, Western Asia, and Africa. Details found here and some maps available here
  • Eurogenes K9b - 9 global populations, approximates Geno 2.0 analysis
  • Eurogenes K9 - 9 global populations, map available here (population descriptions no longer available)
  • Eurogenes K10 - 10 global populations, map available here (population descriptions no longer available)
  • Eurogenes K11 - 11 global populations, map available here (population descriptions no longer available)
  • Eurogenes K12 - 12 global populations. North European ancestry is said to do well with this calculator. Map available here (population descriptions no longer available)
  • Eurogenes K12b - 12 global populations, excluding Native American (Amerindian), map available here (population descriptions no longer available)
  • Eurogenes K36 - 36 global populations, mostly European. This is the most detailed breakdown for Europeans, but that also makes it highly speculative. Details found here
  • Eurogenes Hunter-Gatherer vs Farmer - 12 ancient Hunter-Gatherer vs Farmer populations. Map available here
  • Jtest - Jewish Ashkenazi, 14 global populations but mostly European, this is essentially the EUtest with an Ashkenazi category. Details including maps are here
  • EUtest - 13 global populations, mostly European minus Jewish Ashkenazi. Details including maps are here

Dodecad
  • Dodecad V3 - 12 populations, mostly Asian and African, 2 European, no Native American. More info
  • Africa9 - 9 populations, all African except one European (no Asian, no Native American). More info
  • World9 - 9 global populations, not specific to any continent so good as an overview regardless of your ancestry. More info
  • Dodecad K7b - 7 populations, mostly Asian, 2 European, 1 African, no Native American. More info
  • Dodecad K12b - 12 populations, mostly Asian, 3 African, 2 Middle East, 2 European, no Native American. More info and population maps

HarappaWorld
  • HarappaWorld only has one calculator and as explained above, it’s primarily for South Asian ancestry. It does include some European, African, and Native American populations, but its focus is on South Asian: Indians, Pakistanis, Bangladeshis and Sri Lankans.

Ethiohelix
  • EthioHelix K10 + French - 10 populations, 9 African, one “French” which acts as a European population. This is really only useful/accurate for people with mixed African and European ancestry. Maps available here
  • EthioHelix K10 + Japanese - 10 populations, 9 African, one “Japanese” which acts as an Asian population. Only useful for people with a mix of African and Asian ancestry. Maps
  • EthioHelix K10 + Palestinian - 10 populations, 9 African, one “Palestinian” which acts as a Middle Eastern population. Only useful for people with a mix of African and Middle Eastern ancestry. Maps
  • EthioHelix K10 Africa Only - 10 strictly African populations, nothing else. Do not use if you have no African ancestry as results won’t be accurate. Maps

puntDNAL
  • puntDNAL K10 Ancient - 10 ancient populations, incorporates Caucasus HG as well as Early Neolithic Farmers and Western European HG.
  • puntDNAL K12 Ancient - 12 populations, utilizing ancient oracle, more info provided on results page
  • puntDNAL K12 Modern - 12 populations utilizing modern oracle, more info provided on results page
  • puntDNAL K13 - 13 modern populations, focuses primarily on Asia (6 Asian populations, 3 African, 2 European, 1 Oceania, 1 Native American)
  • puntDNAL K15 - 15 populations, focuses primarily on Africa (particularly East Africa), but also includes some West Asia, and Europe. More info
  • puntDNAL K8 African only - 8 populations, as the name suggest, it’s strictly an African calculator

GedrosiaDNA
  • Eurasia K9 ASI - 9 populations, modeled around the ancient Ancestral South Indian component, no Native American. More info on population descriptions
  • (Removed) Eurasia K10 CHG - 10 ancient populations, modeled on Caucuses Hunter Gatherers, more info on population descriptions
  • (Removed) Eurasia K11 CHG-NAF - 11 ancient populations, modeled on Caucuses Hunter Gatherers and Neolithic Anatolian Farmers, more info on population descriptions
  • Gedrosia K3 - 3 populations, Eastern Eurasian, Western Eurasian, and Sub-Saharan African. More details
  • (Removed) Gedrosia K15 - 15 populations with a focus on the Indian subcontinent. Population descriptions
  • (Removed) Eurasia K14 - 14 populations, using the same Neolithic and Bronze Age source data as the K14 Neolithic calculator, plus some modern populations
  • Eurasia K14 Neolithic - 14 global populations, focus is on ancient Neolithic and Bronze Age genomes from across Eurasia. Population descriptions
  • Gedrosia K12 - 12 populations, designed for individuals of predominantly South Asian and West Asian ancestry for inferring gedrosian Balochi admixture. No Native American. More info
  • (Removed) Gedrosia K11 - 11 populations with a focus on Kalash Indo European peoples of Pakistan. Population descriptions
  • Ancient Eurasia K6 - 6 ancient populations, primarily Europe, Asia, and in between, 1 African, no Native American. Further descriptions are available on results page.
  • Near East Neolithic K13 - 13 ancient populations, with a focus on the Near East. Details provided on results page.


Step 3: Understanding the results: A Terminology Guide
A list of populations you might see and a brief description. I did not include some of the most self-explanatory ones. Some that I have listed might still be obvious to some people, but I’ve seen others ask about them on occasion. If there isn’t one listed here, you might learn a lot by just googling it. There is also a good abbreviation guide here: https://isogg.org/wiki/Abbreviations
Keep in mind different calculators may use different terms to refer to the same region or population.

  • Amerindian or Amerind - Native American (ie, American Indian meshed into one word)
  • Anatolian - mostly Turkey
  • Ancestral Altaic - Asia (excluding South), and Eastern Europe
  • ANE - Ancient North Eurasian
  • Archaic African - broad category for prehistoric Africans
  • Archaic Human - broad category for prehistoric humans around 500,000 years ago
  • ASE - Ancient/Ancestral South Eurasian
  • Ashkenazi - Ashkenazi Jewish of central/eastern Europe (not the same as Sephardic Jewish)
  • ASI - Ancient/Ancestral South Indian
  • Australian - aboriginals of Australia
  • Australoid - “people indigenous to Southeast Asia, South Asia, Australia, Melanesia, Polynesia, Micronesia, and historically parts of East Asia.” (Wikipedia)
  • Austronesian - “relating to or denoting a family of languages spoken in an area extending from Madagascar in the west to the Pacific islands in the east.” (Google)
  • Baloch - people of Iranian Plateau and Arabian Peninsula (primarily the Middle East)
  • Baltic - regions surrounding the Baltic sea
  • Bantu - Central and south Africa
  • Basal - Basal Eurasian?
  • Beringian - areas surround the Bering Strait (Eastern Russia and Alaska)
  • Biaka - aka Aka, “nomadic Mbenga pygmy people who live in southwestern Central African Republic and the Brazzaville region of the Republic of the Congo” (Wikipedia)
  • Caucasian/Caucasus - people of the Caucasus region, the border between Europe and Asia in between the Black sea and the Caspian Sea
  • CHG - Caucuses Hunter Gatherers
  • EHG - Eastern Hunter-Gatherer
  • ENF - Early Neolithic Farmer
  • Fennoscandian - Scandinavia and Finland
  • Gedrosia - Modern day Makran (semi-desert coastal strip in Balochistan, in Pakistan and Iran, along the coast of the Persian Gulf and the Gulf of Oman)
  • Khoisan - Southern Africa
  • Mbuti - “one of several indigenous pygmy groups in the Congo region of Africa” (Wikipedia)
  • Melanesian - “a subregion of Oceania (and occasionally Australasia) extending from the western end of the Pacific Ocean to the Arafura Sea, and eastward to Fiji.” (Wikipedia)
  • Mesoamerican - Native American in Mexico, Central and South America
  • NAF - Neolithic Anatolian Farmer
  • Oceanian - Aboriginals of the Pacific Ocean islands (may include Australia depending on calculator)
  • Omotic - Southwest Ethiopia
  • Papuan - New Guinea and surrounding islands
  • Pastoralist - Sheep or cattle farmer
  • Pygmy - “certain peoples of very short stature in equatorial Africa and parts of Southeast Asia.” (Google)
  • San - Bushmen of southern Africa
  • SEA - South East Asian
  • SSA - Sub-Saharan African
  • Steppe - “ancient North Eurasian hunter-gatherers' heritage, which was subsequently shown to have an influence in later eastern hunter-gatherers and to have spread into Europe via an incursion of Steppe herders” (MDLP K16)
  • Tungus-Altaic - Northeast China and Siberia
  • WHG - Western Hunter-Gatherer
  • WHG-UHG - Western Hunter-Gatherer/Unknown Hunter-Gatherer
  • Volga-Ural - Part of Russia (central)


Conclusion
Which project and calculator you go with greatly depends on your known ancestry. I know all this info is probably still a little overwhelming even with (or perhaps because of!) this guide. If you’re of European descent, and a newcomer to Gedmatch, and you just want a second opinion on your ethnicity results from any of the Big 3 companies (Big 4 now maybe, with MyHeritage joining the bandwagon), I’d recommend Eurogenes K13 or K15. Personally, I tend to prefer K15, because there are maps available showing specifically what regions are covered by which populations. Certainly, you can play around with any of the other Eurogenes calculators too (except Jtest if you’re not Jewish). Most of the other projects and calculators are either geared more towards ancient DNA, other continents, or a mixed ancestry. You may find a non-bias global calculator in some of the other projects, but it’s probably not going to provide the breakdown of Europe you’re looking for.

If you’re looking for an ancient calculator, I again tend to stick to one of Eurogenes’ (HG vs F, or ANE), but MDLP have some good options too. There’s also a couple in puntDNAL which I don’t think have a bias towards any one type of ancestry.

If you’re African, Asian, or of mixed heritage, there are a number of options to choose from, but I unfortunately can’t recommend any over any others. Most global calculators will include Amerindian (I have noted when one doesn’t), but MDLP World22 seems to have the most categories for Native Americans and may be ideal for that.

I was surprised to realize Eurogene's Jtest is the only one that offers an Ashkenazi (or other Jewish) category, so if you're Jewish, it looks like this is your only option. However, it should be noted that there are many Jewish populations typically included in Oracle/Oracle 4 (see below for more details), not just for Jtest but any other given calculator too, in any project. They may include (not necessarily limited to) the following: Ashkenazi Jewish, Sephardic Jewish, French Jewish, Algerian Jewish, Tunisian Jewish, Libyan Jewish, Italian Jewish, Iranian Jewish, Iraqi Jewish, Kurdish Jewish, Syrian Jewish, Georgian Jewish, Turkish Jewish, Romanian Jewish, Uzbekistan Jewish, Azerbaijan Jewish, Moroccan Jewish, Yemenite Jewish, Ethiopian Jewish, and Tat Jewish.

It is frustrating that maps, or at least population descriptions, aren’t available for every calculator, but this is a free service, after all. It’s actually pretty amazing all the work the project creators do to provide this for free.



Part 2 - Oracle
Say what now?

Introduction
The second most common questions I see about Gedmatch are about Oracle. What is it? What do the results means? Oracle is an attempt to pinpoint your origins to a more specific population or region. You'll find many are narrowed down as specifically as regions within countries, or specific religious groups. There are two options: Oracle and Oracle 4. You will find buttons for them listed under your admixture results. Note that not all admixture calculators have Oracle available so pick a calculator that both suits your background and offers Oracle. There is a third button which just says "Spreadsheet" but there is a good explanation for this from Roots & Recombinant DNA so there's no need for me to go over it.

Oracle
Oracle will list your admixture results, then something called Single Population sharing, and finally Mixed Mode Population Sharing.

  • Single Population Sharing attempts to pinpoint a specific, single population that your DNA most closely matches, with a list of the top 20. The distance will tell you how closely you match each group, so the smaller the distance number is, the more closely you match. It is assuming your ancestors all came from the same area/population.
  • Mixed Mode Population Sharing will show you your top 20 of two specific, combined populations in order of how closely you match those populations. It is assuming your ancestors came from more than one location/population. Again, the distance will tell you how closely you match this combo of populations, while the percentage will tell you how much of your DNA matched which population.

Oracle 4
Oracle 4 is essentially the same as Oracle, except it expands on it by providing combinations of 3 and 4 specific populations. The single and double combinations can be different from original Oracle though, so don’t bypass Oracle thinking you’ll get that and more with Oracle 4, it’s best to examine both.

  • Using 1 population approximation works the same as Single Population Sharing in Oracle, but I’ve noticed the results are sometimes different, so they’re obviously using a slightly different calculation. Reading the results work the same though: they are showing you a list of specific populations you most closely match, with the distant showing you just how closely you match.
  • Using 2 population approximation also works the same as Mixed Mode Population Sharing but again, results may vary, and for some reason only lists your top 1 result instead of the top 20.
  • Using 3 population approximation works the same as 2, but with a combination of 3 populations instead. It only lists one result. You may sometimes find two of the three populations are the same.
  • Using 4 population approximation obviously uses a combination of 4 specific populations you most closely match and lists your top 20 combos. This was designed especially for people who have 4 grandparents from 4 different places but it can also work well if most of your ancestry is mainly from 4 different places/populations.

Conclusion
Be aware that the results from Oracle and Oracle 4 will vary depending on what admixture calculator you used, which is why they are found on the admixture results page, and not as a separate calculator. Also keep in mind the results are speculative, but I have found they do often make some sense, and in some cases, can be remarkably accurate. Check out my blog post on a deeper analysis of my Oracle results here.

If you feel like you've got a good handle on this, continue onto Parts 3 and 4, Admixture Proportions by Chromsome and Chromosome Painting.

62 comments:

  1. Excellent information. Thank you!

    ReplyDelete
  2. This is really useful and needed. Thank you! Great work.

    ReplyDelete
  3. I have always found Gedmatch's admixture very confusing. The information you have shared has cleared it up. Thank you for your excellent article.

    ReplyDelete
  4. Thank you. Lots of information to filter. However, I'm so gratefully you shared this awesome knowledge.

    ReplyDelete
  5. Great. Any info on chromosome painting?

    ReplyDelete
    Replies
    1. I might do another guide on the other options by Admixture by chromosome and Chromosome painting if it seems like there's a demand for that.

      Delete
    2. I would love that, I cannot figure out how to read the chromosome paintings at all!!!

      Delete
  6. Very useful! Thank you very much.

    ReplyDelete
  7. Hello, your information is quite helpful. I was wondering, if you have a key for some of the place listed as 2 letter abbreviations when you look at the Oracles? I can guess on some and others leave wondering where they are referencing. Here are some examples that I just do not known: Hu, PT, ES, RO, and IE. Thanks for your information in the blog post.

    ReplyDelete
    Replies
    1. That's interesting, I don't think I've ever come across two letter abbreviations in Oracle - you must match some populations I don't. Which calculator are you using when you get these results?

      Delete
    2. I think I figured it out... It is European country abbreviations. Silly, I finally put the right key words into google https://mfi-assets.ecb.int/resultMfiD/abbreviations Thanks again for your information.

      Delete
    3. Hungary, Portugal, Spain, Romania, Ireland

      Delete
  8. This comment has been removed by a blog administrator.

    ReplyDelete
  9. Excellent article. Do you know what to do on the Oracle if the single population is totally different from the four population estimate? Do you pick the one with the lower number or the one that is a better fit to your known history? Also, I had several ZERO diatances. Is that good? Thank you for any help you can provide. I've been scouring the Internet.

    ReplyDelete
    Replies
    1. Zero distance from a population on Oracle would mean you matched that population exactly, so that's very good. I don't think there is ever a definitive answer though, since this is all very speculative data. I would take it all into consideration.

      Delete
  10. Excellent article. Do you know what to do on the Oracle if the single population is totally different from the four population estimate? Do you pick the one with the lower number or the one that is a better fit to your known history? Also, I had several ZERO diatances. Is that good? Thank you for any help you can provide. I've been scouring the Internet.

    ReplyDelete
  11. I think I'm probably just really stupid. I still don't exactly know how to read the results. I have an idea, but I don't like to assume anything. For example, what exactly does your first photo, the pie chart, say?

    What do the spreadsheets mean? How do you read them?

    ReplyDelete
    Replies
    1. The pie chart is just a visual of the percentages to the left. So it's saying I got 30.63% in the region labelled North Sea (areas that border the North Sea), 20.73% in Atlantic (European areas that border the Atlantic Ocean), 13.57% in Western Mediterranean and 10.87% in Eastern Mediterranean (those are my Italian ancestry), etc.

      As for spreadsheets, there is a good explanation here: http://www.rootsandrecombinantdna.com/2015/12/gedmatchs-new-spreadsheet-feature.html

      Delete
    2. Thanks for the explanation and link! It makes much more sense now.

      Delete
  12. I'm just trying to find out who I am. I know I'm American Indian but I don't know how much.

    ReplyDelete
  13. Thank you so much! So helpful! I was looking at all this and not understanding. This helped so much!

    ReplyDelete
  14. Thank you - that is so helpful. I'm new to this and have been utterly bewildered to date. That explains a lot.

    ReplyDelete
  15. this is an amazing blog, thank you so much for making it clear and simple to understand!

    ReplyDelete
  16. Your blog clarifies it much more than Gedmatch. I know I'm mixed...very much but thr closest I have come to 0 is 2.5 with the 4 population mode.

    ReplyDelete
  17. Thank you, this helps a lot.

    ReplyDelete
  18. Thank you! This is so helpful.

    ReplyDelete
  19. Maybe I missed it, but is there a map for the Eurogenes K13 model?

    ReplyDelete
    Replies
    1. Unfortunately not - there are some maps for K15 though, which has some similar populations you could probably apply to K13.

      Delete
  20. OMG, thank you, believe it or not you have actually made me understand!

    ReplyDelete
  21. Excellent! 👏🏻👏🏻👏🏻 Thank you for putting all this into context. I'd figured out much of this over time, however this is a perfect place to come to as a reference guide.

    ReplyDelete
  22. Thank you so much for taking the time to write and publish this. It has been very helpful and appreciated!

    ReplyDelete
  23. These calculators are so wishy-washy. Why would you believe any of them when they tell you different things on different calculators. I also find it hilarious how one person told me that the "SSA at 1%+ was only real in Southern Europeans and that 0.x% SSA for all other Europeans was just noise and to not count it". LOL. Agenda much. Looking at the K3 averages page SSA, even at tiny percentage shows up in Norwegians as an average in the averages section of the calculator..

    Something like 0.08. It's tiny but it's still there yet people insist that it's only legitimate when it comes to Southern Europeans.

    ReplyDelete
    Replies
    1. All the calculators are certainly estimates and no ethnicity/admixture report should be taken literally. But it is still interesting to see the different reports and how they work.

      I imagine the person talking about African results for Southern Europeans versus Northern Europeans was referring to the fact that many Southern European do get consistent trace amounts from Africa because that Mediterranean area is so mixed. I'm not so sure about SSA, but I have South Italian/Sicilian ancestry and I do frequently get trace amounts in North Africa and I know it's not uncommon among Italians/Mediterraneans (it's also common for Italians to get results in the Middle East/Caucasus). This person sounds like this is maybe what they were referring to, but perhaps misunderstood it slightly.

      As for Gedrosia K3, it's a very limited calculator with only 3 populations, defined as follows:

      1- E Eurasian - This component peaks in E & SE populations such as Ami, Nivkh, Dai, Han, and Ulchi, at about 100%, followed by Siberian & other Asian populations such as Nganasans, Tibetans, Subba, and Mongola.

      2- W Eurasian - This component peaks in Neolithic European farmers such as Stuttgart, and LBK culture, as well as in most modern European populations at over 95%.

      3- SSA (Sub-Saharan African) - This component peaks in Sub-Saharan African populations such as Yoruban, Esan, and Luhiya at over 97%.

      It is probably not really an ideal calculator if you are of recent European descent and looking for a calculator that represents that.

      Delete
  24. This is so much fun!! I've gt some literature to read now. I'm not up on Human Population Genetics.

    ReplyDelete
  25. you're a life saver. time saver. great job on tackling this subject, it is very helpful!

    ReplyDelete
  26. Thank you! An excellent comprehesnive guide which the website lacks!

    ReplyDelete
  27. Thank you for creating this page. What a huge help in deciphering Gedmatch. Excellent!

    ReplyDelete
  28. Thank you, this is very useful. I have a question though. When running Oracle you get, let's say, a distance of 4 what does that number represent: 4k years?

    ReplyDelete
    Replies
    1. I don't believe the distance number represents any time period, I think it's just a scale - such as from 0 to 20 with 0 being the closest, and 20 being the furthest away. That's sort of an example because I'm not sure how high the distance number goes - highest I've seen is 15.99 for mine. The lower the number, the closer you matched the samples for that population.

      Delete
  29. Great Blog, my father was adopted so I only know my mums ancestry which is English and Welsh. When I run the K13 model I get single South Dutch 5.26 and mixed 91.9 south west English & 8.1 Abhkasian total distance 1.76. With K15 model I am getting single South West English 5.33 and mixed South West English 92.40 & 8.3 Tabassaran total distance 3.65. For reference my kit number is M691240. I have no idea what this information is telling me, help!!!

    ReplyDelete
    Replies
    1. Well, all admixture/ethnicity is fairly speculative and it's particularly difficult to determine maternal side from paternal just based on ethnicity alone. Abhkasian and Tabassaran are from the Caucasus area (above the Middle East: https://en.wikipedia.org/wiki/Caucasus ) If you're consistently seeing results in that area, there could be something to it.

      Delete
  30. You are amazingly generous to spell the GEDmatch admixture out for the new, and even seasoned, DNA genealogists! Thank you. I just uploaded my sister's DNA from 23andme new V.5 to the GEDmatch's beta Genesis Admixture calculator and your guide was helpful.

    ReplyDelete
  31. Extremely helpful information History Chick. I have a couple of questions. You said you preferred Eurogenes K13 or K15. K15 because of the maps. Did you mean Eurogenes EUTEST V2 K15? Also, how does one access the maps. Thank you very much.

    ReplyDelete
    Replies
    1. Yep, that's the one - the population maps are available on the creator's blog, there should be a link to it above. Here it is: http://bga101.blogspot.com/2013/10/eurogenes-k15-now-at-gedmatch.html

      Delete
    2. Got it! Thank you so very much.

      Delete
  32. Don't know how to know if my earlier question was seen. Should perhaps have 'ticked', 'notify me.

    ReplyDelete
    Replies
    1. I responded to your question above, you may need to click on "replies" below it - or try here: http://genealogical-musings.blogspot.com/2017/04/finally-gedmatch-admixture-guide.html?showComment=1504574025547#c8899410746532421922

      Delete
  33. This was very helpful. They should link to your post on their site. :)

    ReplyDelete
    Replies
    1. Thanks! That would be nice, but it seems they might only add guides from professional geneticists.

      Delete
  34. I don't understand why K13 should be more accurate for people of European ancestry compared to the regular EUtest (not the new k15). Can you explain this ? Because it makes no sense to me. It would only make sense to me if , perhaps, the people using K13 were mostly Eastern European.

    ReplyDelete
    Replies
    1. That is what the creator of the project and calculators has said. Do you feel K13 is heavy weighted towards Eastern Europe? I haven't found that to be the case in my experience. It doesn't really have an Eastern Europe category to begin with - I think the regions are broader than that (I'm guessing Eastern Europe is within "Baltic" but that this covers more than just Eastern Europe - probably anything around the Baltic Sea). I don't know of any reason why K13 shouldn't be more accurate for most people - of course it depend on the individual, but I guess most people report that the particular algorithm for K13 works the best for them. Your mileage may vary, of course, but if you'd like more info on how the calculators work behind the scenes and why one might generally be more accurate than another, you would have to contact the project's creator.

      Delete
  35. First, how accurate are the Admixture Calculators? What are their Specificities, Sensitivities, Positive Predictive Values, Negative Predictive Values, and Confidence Intervals? I can't find these in GEDMatch.com. I can't take any of these calculators seriously without these values.

    Second, what does the "GED" stand for in GEDMatch?

    ReplyDelete
    Replies
    1. You would probably have to contact each of the projects creators to get that kind of technical information. Or check their blogs - they might share that kind of thing on their blogs. If not, you would have to ask them.

      However, all ethnicity or admixture reports are only ever going to be estimates, since people of neighboring regions share some DNA, making it not always possible to tell them apart with any real accuracy. I wouldn't take any ethnicity/admixture report very literally, it's just fun and interesting to explore. How accurate they will be depends on the individual - you will likely find some are more accurate for you than others and they probably won't be the same ones for everyone. Like I say, some are designed for people of certain ancestry so if you use one that's not ideal for your ancestry, it's probably not going to be very accurate.

      I assume the GED in GEDmatch stands for the same thing that GED in "gedcom" stands for: Genealogical Data.

      Delete
  36. All of the tests mentioned above are interesting. But in each of them, they only mention Amerindian as an afterthought, if mentioned at all. I would hope someday someone would create a genetic test that was centered upon finding small amounts of American Indian genetic information. The few tests available that do look at it do so as an afterthought, and minimize its role.

    ReplyDelete
    Replies
    1. Unfortunately that's probably because many Native Americans in the US are reluctant to participate in supplying samples. So working with small sample groups doesn't allow them to break it down further. The majority of the samples come from Latin America and Canada. That said, MyHeritageDNA attempts to break it down into 3 regions, and LivingDNA attempts 4. DNA Tribes has the most with 7 regions, but I honestly couldn't say how reliable they are: http://dnatribes.com/populations.php - also note how there's no category that covers the eastern US. That's because tribes in the east were largely pushed west and forced to integrate with western tribes, which creates another problem in finding full blooded tribal members that aren't a mixture of different tribes.

      Delete
  37. I have two Italian grandparents, one Portuguese grandparent and one Ashkenazi Jewish grandparent. My results have a lot of Italian and Jewish groups both Ashkenazi and Sephardi, but all of my results seem to have either Palestinian, Syrian, Druze, Lebanese or Samaritan...Do you know if this is common among any of the three populations I come from? I know Arabs were in Italy and that Jews are also related to Middle Eastern populations...Just trying to find out why these groups come up so much in my results.

    ReplyDelete
    Replies
    1. It's not unusual for Mediterraneans to get results in the Middle East too. I don't know much about Jewish DNA but Samaritans are Jewish so I assume what you're seeing is at least not unheard of.

      Delete
  38. I am Filipino and I found the Dodecad V3 the most useful calculator.

    ReplyDelete