Genetic Communities and the Lies we Buy Ourselves

With much aggravation taken as penance I have finally been able to put to rest a cluster of four supposed Ancestor Discoveries (Beta). Turns out the four were Grandparents of second cousin 3x named Jessie May Kelly, a grandchild of the Day family. I am fairly disappointed that they were none relatives to me and indeed come from a branch I was just pondering over regarding the percentage of shared DNA.

I have had a week recently where I have been contacted only by paternal relations to me, and those I think are outside of my x-inheritance. It makes me wonder then about the randomness of dna assignment and how much if any I can really trust Ancestry.com services. I am up to five genetic companies which have my dna sample and have rendered ethnicity lists. There is some major variation between them all, but yet also a seeming obvious underlying cause. Clustering populations together creates a wide variety. My initial thought was that where France goes so do I, but there is also a question of whether France itself is one unit. Northern France is flavored with Britain as much as east and south are peppered with other border countries. Ancestry.com is the only of five companies to declare any Irish, while FTDNA tells me I have recent Amerindian...on 23andMe my West African is disappearing by the minute.

The New Genetic Communities feature looks quite fun and I have wanted an aggregated list of birthplaces for a map for quite sometime. While I made my own by hand once or twice they never are very accurate when new data is arriving each day or even week. A major problem that seems to show up is first degree relatives getting drastically different communities. My Father for example, has at least four or five with many placed overseas or in the American south, think Confederacy. I on the other hand have only a single group for French settlers along the St. Lawrence, excluding regions relevant to my father. It seems quite odd to be sharing no groups with a parent and I have heard this from other users as well.

So what is the basis for communities? It would seem to be a macro version of dna circles, of which parents and children should share. Yet I am only seeing confusion and complaints. Given the differences in my parents ethnic background there should be more. There just isn't...

Third Party Confirmations!

As much as I work as a lone wolf so to speak in my genealogy work I do enjoy having others input. Recently I took the plunge and enrolled in a small tribal band of Metis-Cree people. As a help for others I explained the methods I had used so far in differentiating my own genetic markers for native blood. To my great surprise another member messaged me who was quite confident with using gedmatch himself. He is lucky enough to be a direct male descendant from Henri Membertou's band of Mi'kmaq. I was provided from him a three way match between him, myself and a even less diluted relation. The result was a positive match to the Mi'kmaq peoples. I would say to Henri himself but more likely it would be through one of his many contributing female relations.

Chr Start Location End Location Centimorgans (cM) SNPs
10 92446647 95958204 3.7 903
11 56368374 59769693 1.2 717
13 83990938 88896945 3.1 816
21 21293037 24043081 4.0 716
Largest segment = 4.0 cM
Total of segments > 1 cM = 12.0 cM
Estimated number of generations to MRCA = 7.1

Comparison took 0.05190 seconds.
Distant relation.


So this is a interesting factor in what I already know since I compose over four regions of native populations in the small sector I carry. Another cooperation comes from having limited access to my fathers genetic data. Being able to look at his smaller portion of native dna cleared up any possibility of my larger portion coming from him. Indeed my two largest portions of north amerindian and arctic do not come from him. My fathers tracers are for small portions of the south amerindian and mesoamerican. This I had already presumed but rather it is nice to have confirmed in mathematical calculation.

Another change from this last few weeks of research is discovering some depth to two of my fathers grandmothers lines. Along with that is some details on what was coined "Southern European" in earlier results. I now have confirmation through oracle population matching that my southern component is predominately Spanish with only a smattering of Italian. This follows at least the presence of late Italian ancestry in one portion of my tree but I have yet to connect the Spanish in genealogical records.

Statistical Work in Aborigional Ethnicity Prediction

Last year I set out to comb through the rough of my DNA sampling with many algorithms until I found the clearest percentages. In doing this I clustered Asiatic derivative groups under the heading of tricyclic populations as direct relation to the genealogical evidence of First Nations ancestry. However I did not at the time have a major justification for doing this except that these evidences were non compliant with the majority of my recent generational admixture. I also had my mtdna tested which is passed down from my farthest known matrilineal ancestor who came from Kirkgunzeon, Scotland. The result was the mtdna haplogroup H. I had not considered anything further about haplogroups until today.

Someone had also recommended to me that it could be useful to make a genealogical data map. I decided to place within it known birth locations as well as the locations of genetic populations I matched. The variability of genetic population was in some ways shocking. It did however follow the typical water trade routes in it's spread. The deep Asian line however and the Americas seemed not to be related with the exception of two locales of Aleut. Recently I provided information to the younger sister's of my husband about their own native history as Saginaw band of Ojibway. I did this with only name places, some stories and a map. It surprised me how much a map can say and so I thought of this again for myself.

My statistics produced the average of hard 19%, and soft 20% if one accounts for decimals, of Native genetic material per gene. This 20% however was also not a solid component. 53% was of Inuit derivative, 38% Americas proper including Central/South and lastly a unresolved 21% that was more closely related to Asian heritage. Some strict populations that occurred were Aleut, Inuit, Pima, Maya, Ecuador, Colombia, Peurto Rican. The more exotic was Samii, Tibetian, Han, Thai.

These places however share a major component of Haplogroup. Indeed in map review of the locales it is quite evident that the underlying mtdna associated with the portions of my DNA would be Haplogroup A. The great evidence of this is that Tibetan only hosts A of those also present in the Americas. Asian haplogroups are the predecessor of native american strands which is why it is proper to review them together especially in the absence of actual Asian heritage.

Since Catherine Dufour spoke French-Cree I have come to the assumption that she was Métis-Cree herself as other varieties of Metis would speak their own aboriginal language. Of the localized groups the Cree associated with the region she lived refer to themselves as the Innu. I would not however assume to be exact on which regional tribe of the Cree Catherine came from until we ever have the exact name of her mother. There remains as well the possibility that her father Peter was also of mixed ancestry. Within the Boussey family work I am lucky to have not only a exact name for Anne Marie dit Metisse but also the Indian settlement she lived. Port Royal, Nova Scotia is under the Bear River Nation of Mi'kmaq today. It would be lovely to have exacting answers but native american DNA methods are extremely limited.

Graphing FTDNA for Native Blood

It's common for people to find a unsubstantiated family myth regarding a long lost great grandparent of some tribe. It's part of the Rousseau ideal that native culture is somehow noble through its simplicity and to be admired. Indeed looking just at the surface of things at the time of first contact they had more advanced systems of hygiene and significant ecologically-sensitive agriculture. Those of us with genealogical evidence often find it scant due to the nature of the times regarding the personage and citizenship of a native person in the imperial Americas. As a Metis woman I am at a disadvantage that besides the myth and a lucky interview provided by James Sweinhart, a published journalist, that there is no physical connection I can grasp.

Cherokee Nation at Kansas City, 1994

As a anthropologist, I maintain a certain outlook from multicultural perspective but I have found it fails me when working with more traditional native american individuals. Often I am met with hostility which I believe is less to do with how I look, as some might presume, and more to do with the socioeconomics of tribal living. I certainly don't intend to be a bother and the only community I intend to pursue citizenship with would be the Metis peoples in my local Canadian/US sector. Due to the degree of mixture it simply seems idiotic to sequester anything more specific when indeed I am not specific. 

It is that mixture exactly that is the true topic of this post. After my 3rd party transfer to Family Tree DNA completed I uploaded it into GED Matches album to compare with ancestry original interpretation. It is notable to mention that ftdna actually reevaluated the chip itself not just the raw composite to my knowledge. For the most part I saw predictable and minute changes. Side by side I noticed that ftdna cited more africanized Moorish DNA as apart of my Iberian spectrum. When I reviewed MDLP 22 by chromosome it suddenly occurred to me to class out the asian decent populations and revise it for my native american blood. Trycyclic populations including the middle east are often misplaced amounts of plains native which is so intermixed that it more often registers as middle eastern.

On average I found that 20% of each chromosome was Asiatic derivative. Since the middle eastern connection is still debatable I also went a step further in removing the trycyclic and Asian components to search for individualized known samples of native american DNA. On average this left a 10-12% of each chromosome to be confirmed native american source which is above the threshold for founder markers. Each of the remaining populations confirmed in my chromosomes were set in known genetic regions allowing me to percentage out the confirmed strains by region. Of the 10-12% confirmed; 53% was Northern Hemisphere (Inuit derivatives), 38% was Americas proper, & 21% was trace matching aboriginal societies in the eastern hemisphere. Percentage peaks for the northern group rested on chromosomes 7, 16, 18 & 22. The Americas group peaked at 13 & 20, but only 20 was highly above the margin of error. Lastly the trace societies were highest peaked on 8 & 15 but not as impressively as the other groups. Overall it was North Amerindian which received the highest proportion spikes but it was also one of the least spread matches across the 22 chromosomes.

So I may be lucky enough to have numbers but it also comes down to personal experience and identity. Personal experience has always pushed me to look at aboriginal life in a learning capacity much like a child watching their parents. Even now I am writing my final paper for my degree program on Native American Studies. The expectations of that work is for me to be a participant but I've always been a participant even in times where I was so far away from the source. Part of my most loving experience was the short time I spent with the Cherokee Nation in Missouri. Despite outward appearance I was drawn in and allowed to participate culminating in a honorary membership. While it may have just been words it started me off on a journey I am still following now, as a Metis.

DNA Updates and Forthcomings

Well GEDmatch has finally finished with the re-tokenizing of my DNA pack. It seems that some of the simulations had errors which were pinging mostly in unrelated ethnic results but nonetheless gedmatch decided to rerun new editions for errors. This changes my results to a small degree. The smallest changes are to do with my largest portions. My European percentages are now farther east with 47% a solid mix of Fennoscandia and the Slavic countries bordering Russia. At this time my Celtic/British strains are only 6%. I am still left unsure which country provides the dominant amount of my 34% Mediterranean. Northern Italian and Iberian are both cited. The near east populations matches continue to tantalize me if only because I cannot find their source genealogically.  Roughly 10% of my blood is still Persian which I find amazing.

The more drastic changes have come in the form of my smallest amounts. While initially HarappaWorld and MDLP both cited my mesoamerican DNA from the Muscogee it is now something else. In a sense it could be a more accurate result. About 5% of my result is from South Amerindian and Arctic Amerindian. The latter would correspond to my mother's Mi'kmaq heritage. From what I've gathered online the south is most likely representative of Cherokee. I rather guess that does make more sense since there were only small amounts of exogamy in my direct relations. As the Cherokee are fairly wide, the bands can simply interchange children without looking in another region. Eurogenes still says Mayan...Ugh! But it also indicates a southeast Asia portion as well as oceanic which are barely above threshold.

What I do still find strange is my Samoyed and Berinigan DNA. They are very small amounts but I'm not sure which region to classify them under in my processing of this. I gather for the most part I work on a three category system; European, Asiatic & American. If it doesn't fit those groups than it's wasteful knowledge.

I also took a look at the new ancestry ethnicity predictor. I do find it helpful that they separated my Irish and British blood apart. That makes me confused however about whether Scottish is really the same as British. According to their diagrams it is separate and I have none. My new ancestry numbers are: 52% Great Britain, 21% Scandinavian, 15% Irish, 8% Iberian, 2% Grecco-Italian, 1% Eastern European and <1% Western European. At least that is more interesting than my original pie of British, Scandinavian and unknown. Still I find it very odd that Ancestry states my British as high as 50% when all other groups dictate that it is <26%. Consider for one that as of currently all the calculators for gedmatch place me at 6%. I think it makes ancestry look stupid. I still hope however to upload to FTDNA for my maternal test and correspond my admixture there. My tight budget might make it a Christmas gift to myself.

Unzipping My Raw DNA, Part Two

I don't believe that most people are prepared to make the jump in amateur genealogy to that of the intermediate especially with the inclusion of genetics. When the veil is lifted and you see who you are deep in your cells and atoms it can be either encouraging or feel as though all your work has been through colored glasses. Genetic testing is probably a positive for those whose ancestors were static and the acculturated generations match those that actually make up your being. I'm having the rather mirror-house experience where the recent does not match not only who I am as a person, but also is a falsification or at best over-simplification of the real genetic source.

The time it's taken to run the simulators and compile them into a single excel document however has given me enough pause to collect most of my thoughts. I have to remember that the exotic is mostly deep ancestry so it would not be relevant necessarily except for some quirks. Yet the mirror house experience is that, everything is exotic, at least in comparison to the records and beliefs of those still living today. Working first with admixtures there is a varying degree of interpretation still to the amount of 'white' that I am. Ancestry's estimate was 52% British isles/34% Scandinavian, on the whole roughly 80-90% carbon white.

Harrapaworld, Asia map, disputes this number with only a maximum of 48.5% European. 'Africa 9' on Dodecad echo's a similar opinion that I am 67.5% or less European (including Indo-European as strictly white). Dodecad World version 3 places this number closer to 52% with the rest highly africanized Iberian or west/southwest asian. Eurogenes perhaps gave the most honest admixture result due to it's definitive sub sects. North Sea, which includes water bound parts of Scandinavia as well as the British isles rests at only 20%. JTest states that of these European groups only 25.4% is wholly white and inadmissible as Jewish. I am only 3% however above the noise level (2.5%) for Azkenazi Jewish heritage so the direct line there is remote. Yet the large chucks of central and eastern European do echo the ancestry given my fathers genetic test.

Southern heritage is much more pronounced than Ancestry claimed with some rare alleles coming from Cyprus, Lithuania, and East Asia. Spain and Northern Italy are the founders of my southern heritage with only one so far found in the genealogical record. Dodecad World 9, for example places me at 72.5% Atlantic_Baltic, 25.5% Caucusus/Southern European and around 1.5% Amerindian. 1.5 is interpreted over and over as either Asiatic or Amerindian, notably however it is inferred as Mexican/Central American Indian when addressed. It also increases as high as 3-8% given the presence of oceanic, papuan, beringian and east asian derivatives. If this was a test I don't believe that 20-53% is ever a passing grade so why would I ignore the beautiful multicultural heritage I have.

So what about the estimations given by Ancestry.com? Well I think I figured that out partially inspired by a comment by the authors of GEDMatch's algorithms. "It not just the presence of a unique factor but also the presence of that percentage compared to like individuals." Thus unique factors on chromosomes could be viewed as noise and with strict interpretation ignored completely as genetic drift for known populaces. With the wars over centuries of the Germanic tribes then it would be easy to assume a known factor level of 'noise' and thus exclude it altogether. I believe this is how Ancestry.com applies their ethnicity predictor so it is most simplified and accurate for Europeans, no doubt their largest consumer at whole.

So what am I? Very roughly estimated in admixture...

• 26-40% Southern European (Spanish, Italian, Iberian, Sardinian) 
• 18-24% British Isles/North Seas (Great Britain, France, Scandinavia) 
• 12-26% Middle Eastern (Persian, Near East, West Asia)
• 8-10% Eastern European (West Germany, Balkan, Slav)
• 3-8% Asian/Amerindian (Mezoamerican, Beringian; East-South Asia, Oceania)
• 3-5% North African (Mozabite, Morroccan, Byaka)