Wednesday, November 21, 2012

Genographic Project 2.0 - First Look!

I am very fortunate to have been given the opportunity to get a glimpse into one of the first test results returned for the National Geographic's new Geno 2.0 (now called the GenoChip) and I am happy to be able to share some screen shots with my readers. I haven't had time to reach any conclusions yet and probably won't write a review until I receive my own results, which currently sit at 60% complete. 


However, the screenshots can speak for themselves.

YOUR STORY
The results open with:


You can click through to the different sections for MUCH more information. These results are for a female, thus the question mark under paternal line.

YOUR MAP
There are heat maps for the haplogroups. For mtDNA Haplogroup U, they walk you through the migration from the root of Hg L3 to Hg N to Hg R and, finally, to Hg U.



WHO AM I?
NatGeo has discovered that nine ancestral regions make up each of our genomes. They are Northeast Asian, Mediterranean, Southern African, Southwest Asian, Oceanian, Southeast Asian, Northern European and Sub-Saharan African. According to NatGeo each of us is a blend of these nine ancestral regions, except Native Americans, Oceanians and the Khoisan people.






They compare our DNA to 43 reference populations, each made up of distinct blends of these nine regions:

































Each of us will receive an estimate of which of these 9 regional affiliations and 43 populations we most closely resemble genetically. Here is an example:



There is also an estimate of hominid ancestry:




OUR STORY
The community aspect of the results is called "Our Story". You will be the center of your universe with those with the most similar genetics clustering around you:


So far there aren't enough participants sharing stories,


but that will soon change!


RAW DATA
In case you were wondering what the raw data looks like, here is the beginning of the file, which contains the mtDNA SNPs (with the ID removed from Column A and allele values removed from columns D and E):


and some of the Y-SNPs (no-calls here):






































In the download of the raw data, the Y-SNPs are listed as no-calls for women, as above. This file can be accessed by going to "Expert Options" under Profile and downloaded as a CSV file. (Download example file here.)


That's all for now... Enjoy!

My deepest appreciation to Sharon Schmidt for her generosity and willingness to share with me and our entire community.

[*For additional screen shots, clarifications to some of the questions raised in the comments below and information on third party resources, please see this post.]

35 comments:

  1. How did they choose German as her first Reference Population when she is obviously much, much, more similar to Greek which was her second Reference Population? I really hope they don't make such a blunder with my results. I really think a child could have compared them and chosen Greek over German.

    ReplyDelete
    Replies
    1. I don't know Eochaidh, but you appear to be right.
      Thanks for your comment,
      CeCe

      Delete
    2. It's a bit unsettling, CeCe. Her Mediterranean score and her Northern European score are nearly an exact reverse of the German Reference.
      Making the mistake is one thing, but posting such a glaring error doesn't speak well for quality control.
      Hopefully, they'll fix it and explain what happened.
      Thanks for posting it, Miles

      Delete
  2. Hi Cece,
    Can you please publish the section mentioning: "Helping understand your results: Learn more about this analysis and the science behind how it was conducted", it may contain some clarification.
    Thanks!
    Joss.

    ReplyDelete
    Replies
    1. Hi Joss, Here it is:
      In this section, we display your affiliations with a set of nine world regions. This information is determined from your entire genome so we’re able to see both parents’ information. Your percentages reflect both ancient and recent genetic influences.

      What do you mean—ancient and recent?

      Human genetic patterns have been created over tens of thousands of years as our ancestors migrated around the globe. People living in the same geographic region are more likely to share similar patterns of genetic markers than people living on opposite sides of the world, because their ancestors were more likely to have encountered each other and had children together. Over time, this has made people from, say, Senegal more similar to each other genetically than they are to people from China.

      Migration has also served to disperse these regional patterns over time. For instance, the spread of agriculture from the Middle East into Europe also dispersed Middle Eastern genetic patterns as these early agriculturists moved into Europe. This is why someone who is say, Irish and Scottish on both sides of their family going back many generations would show Southwest Asian and Mediterranean components in their regional affiliations—not because their grandparents were from those parts of the world, but because over thousands of years, all Europeans have mixed with people from these regions and have retained traces of this in their DNA.

      Delete
    2. Cont...
      But your results also reveal details about your more recent ancestry.

      If your parents were from very different parts of the world—say Denmark and Japan—this would be reflected in your regional percentages. In this case, since you get half of your genome from your mother and half from you father, you would be half Danish and half Japanese. At the genetic level, this would show up as half of the regional percentages that each of your parents had—Northern European, Mediterranean, Northeast Asian, and so on.

      The percentage of your DNA that comes from each of your ancestors drops by half as we go back through the generations—you are carrying half of your genome from your mother and father, but only one-quarter from each of your grandparents. Because of this, our ability to see your recent ancestry decreases with each generation in the past. If, say, your great-grandmother (three generations in the past) was Native American, you should see that roughly 13% of your DNA is Native American. Our limit is six generations, which would represent around 2% of your DNA. Beyond that, we can’t be certain that the percentages are significant, and this is why you won’t see regional affiliations of less than 2%.

      People with recent ancestry from very different populations can have a mix of these regions that is not typically seen in indigenous populations. Hispanics, for instance, will have some of the typical European components from their European ancestors, but also Native American and even African components as a result of the recent mix of world cultures and populations that has occurred in the Americas over the past 500 years.

      Read more analysis of different ancestral stories in our participant case studies.

      How the Closest Populations Are Determined

      Modern day indigenous populations around the world carry particular blends of these regional affiliations. We compared your DNA results to the averages from each of 43 reference populations we currently have in our database and estimated which of these populations were most similar to you in terms of the genetic markers you carry.

      Notice some unusual populations listed for you?

      This doesn’t mean that you belong to these groups, only that these were the groups in our 43 reference populations that were closest to you. As we expand our set of reference populations, you may find that you are closer to another group. This simply reflects the ongoing scientific refinement of the Genographic reference dataset, as well as improvements in our methodology for assessing your closest populations.

      Remember, the regional blends were determined over thousands of years, so you may see surprising regional percentages reflected in these populations.

      Delete
  3. I'm not seeing the German as first Reference Population - was it corrected since you first put the images up?

    At any rate, thanks so much to S. Schmidt for sharing, and CeCe for posting this for all to see!

    ReplyDelete
    Replies
    1. I now see the pages showing the German as first Reference Population. For some reason, they were among other missing pages when viewed using Firefox on my computer. I tried another browser (Safari) and can now see all pages in this example.

      I agree that it makes no sense, based on auntsha's results.

      Delete
  4. I am not a genius with this stuff but I did question this in my posts at ftdna and 23. I put a question mark after the German & Greek listing... wondering a mistake? We will see, as the the others get their results. The feedback will be overwhelming if results are not in line. Thanks for the critique, it is always hardest being the first! :-)

    ReplyDelete
  5. I stumbled across your blog from Google. I was trying to find others who have received their results from Geno 2.0. I am a participant as well, and was curious on how many people received their results. So far, there have been at least 18 people who shared their stories on genographic.com.

    ReplyDelete
    Replies
    1. Hi Anonymous,
      From the number on the raw data file, it looks like there were about 800 people who received their results. Most are probably busy with family right now.
      How did you feel about your results?
      Thanks for commenting!
      CeCe

      Delete
    2. Where did they share their stories? I took a look at the Geno page and I did not see anything.

      Luis

      Delete
  6. Thank you Cece, Auntsha and others for sharing their opininons on these results. Does NG have a place for people to comment on their results?

    Luis

    ReplyDelete
  7. I must be crazy! :) The majority of people I have encountered find these results to be perfectly logical. Either that, or they believe that Geno 2.0 is not giving a #1 and #2 match and that Geno 2.0 is leaving it up to the tester to interpret the closest match.

    Let me put this out one more time and see if I'm nuts!

    Auntsha: Med 47%, No Euro 31%, SW Asian 20%
    German Pop: Med 36%, No Euro 46%, SW Asian 17%
    Greek Pop: Med 54%, No Euro 28%, SW Asian 17%

    Doesn't it seem that the German Reference Population score is almost the exact opposite of Auntsha's score on Med and No Euro? Doesn't it also seem that Auntsha's Med and No Euro percentages are only 7 and 3 percentage points different from the Greek Reference Population?

    Perhaps I'm loosing my mind. I see a stark difference, and, therefore, a glaring error.

    ReplyDelete
    Replies
    1. I don't see an error at all. The German and Greece are the TWO CLOSEST Reference Populations given to her in order to judge for herself what is most likely based on her own knowledge. The data for those populations may change slightly so they offer the two closest one for you to go over. Why did they put the German first and not the Greece - who knows - maybe the computer ordered it according to spelling. Listing it first does not mean that it is THE closest to her it is just one of the two closest offered by NG's results. That is the way I am reading it - of course I am new to this so I could be off my rocker :)

      Delete
    2. I agree with Eochaidh, because not only is the Greek closer to Auntsha's profile by differences in percentage, but the Tuscan (Italy) ref. pop. is exactly the same as the Greek ref. pop., and the Iberian ref. pop. seems even closer. Shown with differences from Auntsha in parentheses:
      Auntsha: Med. 47% , N. Eur 31%, SWA 20%
      Grk/Tusc.: Med. 54% (+7%) N Eur 28% (-3%), SWA 17%(-3%)
      Iberian: Med. 48% (+1%), N Eur 37% (+6%), SWA 13% (-7%)
      German: Med. 36% (-9%), N Eur 46% (+15%), SWA 17%(-3%)

      There is more No. European in the German, by a larger percentage than any of the differences in the Greek, Tuscan, or Iberian references.
      My 2 cents, but maybe a statistician could make more out of it all.

      Delete
    3. The newest twist on FTDNA's forum is that NatGeno doesn't have a clue about matches and they don't care what testers results are. They're just building a database.

      This comes from a Geno 2.0 supporter who thinks our personal results don't matter.

      I believe that personal results and matches do matter, especially to those who are missing information on a parent or other ancestor.

      Delete
  8. Hi All, You may know me as the author of the mtDNA haplogroup analysis tool mthap. http://dna.jameslick.com/mthap/

    Thanks to Sharon and another contributor who shared their Genographic 2.0 raw data with me, I'm hard at work adding support for Geno2.0 results to mthap. It is already well on the way, but the raw data files are not reporting results with a consistent orientation, so I had to derive this information myself. The problem is that I can't be sure this is completely correct with just two examples. If you have already received your Geno2.0 results and would like to help, please send your data files to me at: james.lick@jameslick.com

    Thank you!

    ReplyDelete
  9. Maybe it has to do with her mtDNA (U4a) which is mostly found in E. and N. Europe particularly in Scandinavia and the Baltic area - from what I read - so they chose Greece and Germany as the most likely. ???

    ReplyDelete
  10. There are some typos in these reference populations descriptions--like calling the Lebanese sample "Iranian" in the text and mislabeling parts of the Malagasy admixture. Maybe there needs to be some editing done on these descriptions.

    ReplyDelete
  11. I'm curious how someone with Ashkenazi Jewish heritage would look according to these reference populations? Maybe they would be closer to the Iranian population than non-Jewish Europeans (i.e. more Southwest Asian)? But they might also have a higher Mediterranean percentage, as seen in many of the Middle-Eastern groups.

    Am I the only one who finds it strange that Kuwaiti, Egyptian, Lebanese, etc. have more Mediterranean than the Greeks? I suppose this just reflects the fact that it comes from the agriculturalists from the Fertile Crescent. The "Mediterranean" group seems that it could be more appropriately labelled something like "paleolithic Middle-Eastern".

    ReplyDelete
  12. yo me hice un test con familytree dna pero no he obtenido resultados que podamos llamar importantes, que puedo obtener con este sistema? cual es el costo?
    gracias
    I did a test with dna FamilyTree but I obtained important results that we can call, I can get with this system? what is the cost?
    thanks

    ReplyDelete
  13. Does anyone have an explanation for why there is no Northeast Asian percentage in the Highland Peruvian reference group? If the 4% portion in the Amerindian (Mexican) reference group is explained by the origin of Native Americans in Asia, I guess we could expect this to be the case for the Peruvians as well, right? Although I guess for some reason the percentage could be lower than 2% and not significant.

    /Dan

    ReplyDelete
  14. Cece, I was looking at your raw file snapshots and wondering how the allele results are going to be reported in two columns. Are they using +/- notation, or are they noting ancestral and derived values? I mentioned it on the U106 list, and Charles Moore sent me back here to ask you about it.

    David

    ReplyDelete
  15. @David I've only seen two samples so far and both are female, so I don't think anyone knows yet how the Y-DNA SNPs will be reported.

    ReplyDelete
  16. I did inquire at The Geno Project about the order my results were in with the lesser %German pop being put first over the Larger Mediterranean. I also had a concern about the German Pop as that was not on the radar for me from using other tools. Here is their reply.
    "Thank you for contacting the Genographic Project. It does not matter which order your population groups are in only which percentages are the higher etc. Please note that most Germans have both Mediterrean and Northern European descent as reflected in your results. Our Algorithm is different than 23&me so I would not expect your results to be identical.
    thank you for participating in the Project."
    I never mentioned 23 only that I used other tools, meaning Gedmatch. Actually 23 is not a good example as the are not very elaborate in their prediction only European, Asian, and African. Just thought you would like to know they were responsive and maybe this will help when you get your results.

    ReplyDelete
  17. I have no problems with the designations. Both Germany and Greece are modern nations, speakers of languages of Asian provenance that entered Europe after the Neolithic probably with farmers and dairy/beef cattle herders. Ancient Germania, and ancient Greece no longer exist genetically except as bits and pieces in modern Europeans whether German or Greek or Russian. France of the days of the Gauls is dead and gone. Clinging to bygone concepts of ethnic origins and "race" is passe. Europe is one big melting pot.

    ReplyDelete
  18. @jlick I'm starting with the assumption that the Y-results will be identical in columns D & E, which seemed to be true for the mt results. For what it's worth, for one kit I was able to look at, I determined that out of 142,173 non-Y SNPs there were only 336 no-calls in column D. At that same rate, there would be about 30 no-calls for the 12,438 Y-SNPs.

    ReplyDelete
  19. I called the help desk for Nat. Geo 2 (which is at FTdna) and had an informative conversation about the meaning of the reference populations. The short answer is that the populations are chosen by similarity of "markers", not by their resemblance to your percentages of the 9 population areas. Unfortunately, the markers are not explained on the site at this point and showing the comparison of the percents in each population area is misleading. They have nothing to do with the choice of the population.
    Here is the example using my results. My ancestry is from Austro-Hungarian Galicia as far back as I can trace. My regional percentages are 41% Northern European, 38% Mediterranean, 18% Southwest Asian and 2% Northeast Asian. My two reference populations are Finnish (57% Nothern European, 17% Southwest Asian, 17% Mediterranean, 7% Northeast Asian) and Greek (54% Mediterranean, 28% Northern European, 17% Southwest Asian). When I looked at the 43 specific populations, I found 2, Bulgarian and Romanian, which matched my regional population percentages much more closely, and were even physically closer to the location of my ancestors' origin. However, the hint to the correct interpretation is in the paragraph "What your results mean" which says "We compared your DNA results to the reference populations we currently have in our database and estimated which of these were MOST SIMILAR TO YOU IN TERMS OF THE GENETIC MARKERS YOU CARRY." Genetic markers in the reference populations, not the similarity of the regional percentages. As I said above, showing those regional comparisons is misleading. They are not the basis of the reference population choice, the markers are. It is interesting to see how you differ in those percentages from the populations that you most closely match in terms of markers, but unfortunately, there is no discussion of the markers. I recommended that such information be included in this section.

    ReplyDelete
  20. It has been three weeks since I received my results and after reading Cece's results decided to go back and review my own again. Well they actually changed my German ref population to Bulgarian and kept the Greek. Bulgarian happens to match my Med-N Eur-SW Asia in %'s almost exactly so don't know if they decided to change it for that reason or what! One of the above posts says they are not going by those %'s so don't know what is going on. I do wonder if some other early results were changed too?

    ReplyDelete
  21. Very Interesting! I was wondering how long it typically takes NatGeo to return results? The labs received my test the last week of November, and my results have not posted as of today. The test is still at 60% complete – it has been at 60% for four weeks! I wonder if this is unusual. I am interested in comparing these results with the results from my Ancestry.com test. I'm hoping they don't differ drastically, otherwise I would lose confidence in this whole genetic testing thing.

    ReplyDelete
    Replies
    1. Hi Vienna1791,
      There hasn't been a "typical" time frame for Geno 2.0 test yet. Your experience is not unusual.

      Your Geno 2.0 results most definitely will differ from AncestryDNA's, but that alone shouldn't make you doubt genetic testing as a whole. First of all, Geno 2.0 is testing very deep ancestry - 1000's of years ago and they try to fit your genetics into a single reference population label. If you are highly mixed like many of us Americans, this doesn't always work perfectly.

      On the other hand, AncestryDNA is not looking as deeply into your ancestry as NatGeo and give you percentages from different reference populations, not just one. This usually works better for people with mixed ancestry. However, I haven't been satisfied with AncestryDNA's Genetic Ethnicity tool. YOu can read my review and comparison of the 4 major companies offering this type of analysis here: http://www.yourgeneticgenealogist.com/2012/12/comparing-admixture-test-results-across.html

      None of these analyses are perfect, but the science is improving quickly.

      CeCe

      Delete
  22. Could you please shed more lights about the meaning of the "percentage" reported in the results? Does it mean "probability" since it is an estimate of how similar the genetic markers are compared to the reference population?

    In the posted reference population for Chinese, the numbers seem to be incorrect. They are far short of 100%. Please check again, perhaps a typo error there.

    ReplyDelete
    Replies
    1. Hi Jasmine,
      You are right about the percentages in the Chinese reference population. It looks like NatGeo must have made an error there. I just screen grabbed it from their site. The percentages should add up to 100%.
      They are indeed just an estimate.
      Thank you for your comment,
      CeCe

      Delete
  23. Hi,
    I just received my results. I was wondering if there were anyways we could see if I have the Roman haplotype with the Genographic project. I had 44% Mediterranean and my best populations were Tuscan and British, but is there anyway I can find if I have some Roman or Ancient Greek ancestors? Or does the fact that I have 44% of Mediterranean genes confirm that I have them?
    Thank You very much for your time!
    Xavier

    ReplyDelete