Hola Armando y NuestrosRancheros,
I just got some FamilyFinder results and analyzed them using DIY Dodecad Wrapper / Globe13 as you suggested. Then I figured out the ethnic percentiles assuming Mexican and Spanish ancestry, but I wonder if the numbers I got would match your calculations. Could you please give me an ethnic percentile estimate based on the information below?
I have not traced back this line yet to the NR area, but other lines from my family tree do go back to Zacatecas (Mazapil and Gruñidora).
Thanks in advance.
VN
Siberian 00.02
Amerindian 20.22
West_African 1.33
Palaeo_African 0.21
Southwest_Asian 8.10
East_Asian 0.00
Mediterranean 35.42
Australasian 0.00
Arctic 1.13
West_Asian 6.01
North_European 25.70
South_Asian 1.51
East_African 0.34
Navarro, FamilyFinder
Hola Armando,
Thanks for the feedback. Indeed the calculations I made are almost identical since the Mediterranean : North European proportions in my samples are very similar to the ones form the modern Spanish_D tables. I find this very interesting, almost as if the second set of results was confirming the first one. I agree that in the example I gave the Southwest Asian, West Asian and South Asian percentiles need to be added to my Spanish total (either all or just part of them).
I wish I could compare my future Geno2.0 results to the FamilyFinder ones, but it won't be so easy since I bought Geno2.0 for myself (given my curiosity on my own ethnic percentiles and my idea of a large database of local populations from the Genographic project) and the FamilyFinder tests for samples from my parents (in order to take advantage of the distant family matching of FamilytreeDNA). Unfortunately I can not afford buying every test for everyone I would like to test in my family.
Thanks again for your time and have a nice week.
Sincerely,
Victoriano Navarro
Navarro, FamilyFinder
Hola Victoriano,
Your Mediterranean and North_European are almost exactly 70% of the Spanish_D. (The D stands for Dodecad). So I took 70% of everything else that Spanish_D has and put it in your Spanish percentage and subtracted that amount from the different Asian and African percentages. Then I placed the remaining amounts into your Amerindian and African percentages. You come out 71.24% Spanish, 27.81 Amerindian, and 0.64 African. You have a lot more Spanish than most Mexicans but about the same as most Alteños that show all or most of their ancestors to be españoles.
The following doesn't align perfectly so you have to count the headers and the columns to be sure you are on the correct group when looking at the percentages.
Victoriano Spanish_D Spanish Amerindian African
Siberian 0.02 0.00 0.00 0.02 0.00
Amerindian 20.22 0.00 0.00 20.22 0.00
West_African 1.33 0.90 0.63 0.00 0.43
Palaeo_African 0.21 0.10 0.07 0.00 0.11
Southwest_Asian 8.10 5.60 3.92 4.18 0.00
East_Asian 0.00 0.00 0.00 0.00 0.00
Mediterranean 35.42 50.40 35.42 0.00 0.00
Australasian 0.00 0.00 0.00 0.00 0.00
Arctic 1.13 0.00 0.00 1.13 0.00
West_Asian 6.01 5.10 4.21 1.80 0.00
North_European 25.70 36.80 25.70 0.00 0.00
South_Asian 1.51 0.40 1.06 0.45 0.00
East_African 0.34 0.60 0.24 0.00 0.10
Total 71.24 27.81 0.64
Saludos,
Armando
Thoughts on recalculating FamilyFinder DNA Geographical %
Hi Armando,
Thanks for the analysis. I also used the Spanish_D data for the calculations and came up with the following results:
When doing the math based on North European percentile: 69.77% Spanish, 21.37% Amerindian, 7.87% Asian and 0.73% African.
When doing the math based on Mediterranean percentile: 70.21% Spanish, 21.37% Amerindian, 7.82% Asian and 0.76% African.
I see you add all the Asian populations when calculating Amerindian %, but I’m not sure if that’s quite right. I agree if you add the East Asian, Siberian and Arctic (at least the Asian and American parts), but not so much when adding the Southwest Asian, East Asian and South Asian percentiles. I don’t know how are the Dodecad people mapping those regions, but at least according to Wikipedia Southwest Asia and Western Asia are the same as the Middle East (Arabic peninsula, Israel and Palestine, Turkey, Iran, Iraq, etc.) while South Asia includes mainly India and Pakistan (other neighbouring countries may or not be including depending on the source). That makes me wonder if those percentiles could also be attributed to Spanish origin, from a little more Moorish or Jewish roots. Do you know where I can find the maps used by Dodecad for their mapping?
In case someone is wondering, these are the spreadsheets we are using for the calculations:
https://docs.google.com/spreadsheet/ccc?key=0ArJDEoCgzRKedGR2ZWRoQ0VaWT…
These are showing data from different populations, and we are using the ones from modern Spanish people (Spanish_D). However, these numbers are averages and may have variations, some bigger than others (especially if we are considering individuals). Among the populations listed there we can find many different ones from Spain: Andalucía, Aragón, Canarias, Cantabria, Castilla-La Mancha, Castilla-León, Cataluña, Extremadura, Galicia, Murcia, País Vasco, Spaniards, Spanish and Valencia. They list other ones which may be closely related to Spain: Sephardic Jews, Portuguese and Morocco Jews. Most of these Spanish populations show similar (but not identical) percentiles, although the data from País Vasco has higher Mediterranean and North European percentiles than the others.
I think we have to consider that most of us do not descend from modern Spaniards. We descend from different kinds of Spaniards (and some Portuguese), who came to New Spain and México from different regions of Spain, at different times between the 1520’s and the 1940’s. In those 400+ years there were many changes in the Spanish genetic pool due to immigration, emigration, deaths and local population shifts, all resulting from wars, diseases, colonization of the territories overseas, politics, economics, the expulsion of unconverted Jews and Moors in 1492, etc. (of course, similar changes occurred elsewhere in the Planet). My point is that the percentiles shown for Spaniards in the spreadsheet may be a good point to start to calculate our own, but we should allow for some flexibility in the results.
I can’t wait to get my own Geno2.0 results and see how does the geographic data I get from that test compares to the Dodecad calculations.
By the way, in the particular case of the data I show, the higher Spanish percentile is in part explained since a grandfather immigrated directly from Spain in the late 1890’s, so for the three remaining grandparents the Spanish (Western Asian)-Native Mexican proportion may be closer to 2.75:1 according to my calculations, 1.7:1 according to yours.
Saludos cordiales desde Torreón,
Victoriano Navarro
Thoughts on recalculating FamilyFinder DNA Geographical %
Hola Victoriano,
Your two calculations aren’t very different from each other. They are almost identical.
I hope the following gives some insight on why I was wrong on the Asian components.
If you look at the spreadsheet all of the Iberian groups, except Basques, have Southwest Asian. That includes the Portuguese. Canarias and Murcia have a tiny amount of East Asian, all other Iberians have none. All Iberians including Basques have West Asian. All Iberians except for those from Andalucía (from the 1,000 Genomes project) have South Asian. Therefore, at least a portion of Southwest Asian, West Asian and South Asian need to be added to your Spanish total if not all of it. You don’t have East Asian and neither do most Iberians. The rest was hard for me to tell. We know that our Amerindian ancestors came from Siberia and they must have had some Asian components and probably some of the other components such as Arctic and Siberian, but once I look at the Maya population from Li et al. (HGDP) I see that they don’t have any of the Asian components. I had overlooked that. Hopefully a new study comes out that includes more Amerindian groups of Mexico and a higher number of subjects. Until then it looks as though all of the Asian components need to be included with the Spanish/Iberian ancestry or attributed to other ancestors.
Of course, the above means I incorrectly added the Asian to the Amerindian. I hadn’t looked closely enough at the spreadsheet and had assumed that a lot of Asian components would have went with the Siberians into North Central and South America. These components are from the past 50,000 years in which multiple waves of people migrated to different parts of Asia, Middle East, and Europe and the Asians went to the Americas 20,000 years ago which I why I had assumed some of those components went with the Siberians to the Americas.
Now let’s look at what Spencer Wells of Geno 2.0 has to say about the Amerindian, Mongolian, and the Iberian populations at https://genographic.nationalgeographic.com/overview-of-regions-and-clos…
-------------------------------------------------------------------------------------------------------------------
AMERINDIAN (MEXICO)
This reference population is based on samples collected from regions in central Mexico. These are the descendants of the original settlers of the Americas over 15,000 years ago, accounting for the 83% Native American percentage. The 5% Northern European, 4% Mediterranean, 3% Southwest Asian come from recent migrations to the Americas by Europeans over the past 500 years. The 4% Northeast Asian component is reflective of the origins of the Native Americans in northeastern Asia over 20,000 years ago
MONGOLIAN
This reference population is based on samples collected from the native populations of Mongolia. In addition to the 67% Northeast Asian component typical of Siberian and Mongolian populations, there is also a western Eurasian contribution of 12% Southwest Asian, and 6% Northern European, which shows the connection to steppe nomads of western Eurasian origin, as well as migrations from the Middle East with the spread of agriculture over the past 10,000 years, along with some recent migration from Russian populations. The 9% Southeast Asian component is likely a result of admixture with groups further to the south, such as the Chinese. The 4% Native American component reflects the fact that the ancestors of today’s Native Americans probably originated in this region.
IBERIAN (SPAIN & PORTUGAL)
This reference population is based on samples collected from people native to Spain and Portugal. The 48% Mediterranean and 13% Southwest Asian percentages reflect the strong influence of agriculturalists from the Fertile Crescent in the Middle East, who arrived here more than 8,000 years ago. The 37% Northern European component likely comes from the pre-agricultural population of Europe—the earliest settlers, who arrived more than 35,000 years ago during the Upper Paleolithic period. Today, this component predominates in northern European populations, while the Mediterranean component is more common in southern Europe.
------------------------------------------------------------------------------------------------------------------
So as you can see he also shows the Amerindian have the Northeast Asian component. Mongolians have Southwest Asian, apart from other Asian components, Native American, and even the Northern European component. The Iberians to have a significant amount of Asian ancestry. You also have to take into consideration that Spencer divides Asians into 3 and not 4 ancestral components like Dienekes Pontikos did with Dodecad Globe13.
I agree that our Spanish ancestors aren’t all going to have been like the population group that I used as the comparison group. However since it should be an average of what they are today then I think is the best option especially considering that our ancestors were from all over Spain including many that were from the Basque Country. Yes, Spanish and Portuguese have some ancestors that were Jews and Moors. Some have more than others. They would have admixed over the centuries. All of those ethnic groups were admixed already with many of the same components as the Spanish_D and none consisted of a single component. Our Iberian ancestors would have been a mix of the Portuguese and Spanish that had already had the admixture of the Jews and Moors as well as any other ethnic group. Yes, the Spanish and Portuguese might have changed some over the past 500 years but it shouldn’t have been large since the autosomal components based on SNPs are from 50,000 years ago to 8,000 years ago for all of the groups. It isn’t as if the markers that are from that far back would have changed so drastically over just 500 years. Scientists are even using the autosomal markers of ancient remains to provide clues of where which groups originally had which autosomal markers. They point to Otzi the Iceman for clues to our past - http://en.wikipedia.org/wiki/%C3%96tzi#Genetic_analysis and autosomal results from such specimens are used in the DNA databases of the commercial companies. I’ll use the Y-DNA haplogroups as an analogy for the markers not changing. The R1b marker still exists from more than 10,000 years ago in Mexicans that descend from Spaniards that populated Mexico 400 or more years ago. Some of those families might have created new subclades but the ancestral marker is not lost. They are still R1b regardless. This is also the most common among Iberians and it is also the most common among Mexicans that have a European paternal line. The same goes with the few Mexicans that have a female European line. They have the most common Spanish mtDNA. Therefore, our Spanish ancestors shouldn’t be that different from the modern Spaniard. Only time will tell for sure. We might not be around when they figure that out. Even so from now, as far as the components, I will focus on commenting that the North European and Mediterranean is high in Iberians and they also have Asian components, the Amerindian is self-named, the African is partially from our Iberian ancestors and directly from Africa for some portion for some people.
Of course there should be flexibility. I am not saying the components should be assigned to the ethnic or geographic groups exactly as I analyzed. I’ve acknowledged my error on the assignment of the Asian components. The assignment of the ancestral components to ethnic groups was my attempt to correct assignment of ethnic labels from FTDNA Population Finder, 23andme, or even Geno 2.0. Dienekes made multiple calculators until he came up with Globe13 which showed to be the best fit. FTDNA admits that Population Finder is in Beta but they assign too much of our components to Middle Eastern population groups. It’s one thing to say they have the Middle Eastern component but yet another to say a person belonged to a specific ethnic group from that region when there is no history of it in the person’s ancestry. They provide labels as to which population group the person likely has ancestry from without taking into consideration the person’s real ancestry or that it looks like they belong to those population groups only because the other ethnic groups also have some of the same ancestral components from 8,000 to 50,000 years ago that have been passed down to both the person’s autosomal DNA and to the ethnic group. Geno 2.0 will probably end up saying you match some other population best and Mexican second if at all. They said one subject is Bulgarian even though they show that she has 35% Native American and Bulgarians have none and all of her other components are within 2% of their Mexican population group. Regardless, Geno 2.0 and Globe13 with the FTDNA raw data file are still the best because they provide their reference population groups and which components those reference population groups contain. FTDNA Population Finder does not do that and 23andme does only provides how many subjects match a certain component. 23andme won’t provide all of the components an ethnic group contains. Geno 2.0 also provides the age and origin of each ancestral component which none of the other do.
My main objective was getting knowledge about what Geno 2.0 and Globe13 have to offer to our group and the failings of FTDNA Population Finder. Everyone can use Globe13, read the source population group data, and decide for themselves what they think they consist of. Hopefully once more data is available from more studies the picture will become clearer.
By the way, the DIY Dodecad 2.1 Wrapper has been updated to use also analyze Geno 2.0 and Ancestry.com files. It can be downloaded at http://www.y-str.org/tools/diy-dodecad-wrapper/
I was able to use a Geno 2.0 file and here are the results –
Geno Component Globe13 Component Geno Reults Globe13 Results
Northern_European North_European 18 16.36
Mediterranean Mediterranean 28 24.74
Native_American Amerindian 35 35.21
n/a Arctic 0 3.42
Northeast_Asian East_Asian 2 0.42
n/a Siberian 0 0.52
n/a Arctic 0 0
Southern_African Palaeo_African 0 0
Oceanian Australasian 0 0.37
Sub-Saharan_African West_African 6 5.98
n/a East_African 0 0
Southwest_Asian Southwest_Asian 10 4.04
n/a West_Asian 0 7.59
n/a South_Asian 0 0.26
Southeast_Asian South_Asian 0 0
East_Asian 0 0
Total Total 99 98.91
Basically the Globe13 North European, Mediterranean, and Amerindian will be very close your Geno 2.0 results. Your Globe13 West Asian and Southwest Asian will mostly be part of your Geno 2.0 Southwest Asian result.
Saludos,
Armando