Autosomal_NA_Analysis

BIOGEOGRAPHICAL AUTOSOMAL DNA TESTING

David K.Faux

The DNAPrintTest offered by Family Tree DNA and Ancestry by DNA provides a "most likely estimate" (shown as a red dot in the above triangle plot), based on an analysis of 172 "ancestral informative markers" known as single nucleotide polymorphisms (SNPs) on 19 of the 22 pairs of autosomes (the sex chromosomes are not included). By comparing a DNA sample to individuals in the four reference groups noted above, the test estimates the percentage of each group in the genetic makeup of a person.

The test appears to have some difficulty in distinguishing between the Native American and East Asian people, perhaps due to migration patterns and genetic similarity between these groupings. Therefore it is likely that a better estimate of the true ancestry of David K. Faux is 3% Native American. The black, blue, and yellow bands around the red dot are confidence intervals, and in this case mean that the Native American and East Asian components could be as high as 20% - however this is much more unlikely than the estimate where the red dot appears. This data becomes more meaningful when compared to other sources such as a well - documented genealogy. Comparing the two lines of evidence, it appears that the estimates are likely accurate. It should be noted, however, that lately the DNAPrint Version 2.0 has come under heated criticism based on findings that cannot be substantiated with other data (e.g., genealogical), and which make little sense based on known genetic principles. Therefore, it is entirely possible that the results of this test for myself (and others) are not valid - at least in relation to estimates of minority ancestry below 30% (approximately). Results for Greek with 9% and Pakistani with 30% Native Ameriacan does not inspire confidence.

One way to address the problems is to ensure that the testees have only one documented source of Native American and no Greek or Pakistani ancestry. It makes sense to assume that if there are statistical anomalies, that they could be "smoothed over" by employing a series of family members whose genealogy is known and who all descend from the same Native American ancestor. The broader picture provided by this approach may offer a more valid and reliable esitmate of the true contribution of this ancestor to the genomic make up of the family. For this reason I have requested selected family members to take the DNA Print Test - the results are available here -

YOUNG Family BGA Racial - Ethnic Testing:

1) AUTOSOMES: Tested via DNA PRINT

2) AUTOSOMES: Tested via new"Chip Technology"

To date I have had two genome scans involving approximately 1 Mb (1 million) SNP (single nucleotide polymorphism) markers by decodeme (http://www.decodeme.com) and 550 Kb SNP markers by 23andme (http://www.23andme.com). While the 22 autosomes (half inherited from my mother, half from my father) are promising, at present it is going to be very difficult to determine which haploblocks (segments of varying size, for example 3 Mb), of SNPs came from my father and his ancestors, and those which came to me via my mother and her ancestors. As research into Euopean ancestry accelerates it will soon be possible to answer some of these questions. For the moment it will be easier to search for blocks suggesting "minority" ancestry (non - European) from my mother's side of the family. This will be "encumbered" by the reality that my only documented Native American ancestor contributed (on paper) 1/256 to my genome.

It would be difficult to impossible to "find" the contibution of one ancestor from 8 generations back without a great deal of luck tossed into the mix. A hint at what might be found can be seen below in the decodeme "Ancestral Origins" snapshot of all chromsomes, showing European (blue), African (green), and East Asian (brown). It should be noted that an unknown part of the ancestry assessment will be "noise" or "false positives". This may be particuarly true for the African element since all people today trace their ancestry to Africa even if 60,000 years ago.

As noted above, the genealogical work of the author has shown that most of the true blocks of nucleotide bases attributable to specfic ancestry should be European, with only a small component of Native American. Since neither decodeme nor 23andme are able at this point to differentiate Native American from East Asian, only the latter category is used - hence there will be some degree of underestimation of true Native American ancestry due to failure to "recognize" some sequences from this source. This is particuarly problematic since the Native American reference samples are from the Mexican border (e.g., Pima) into South America (e.g., Surui). There are no reference samples for Native Americans of Eastern North America and so inferences must be made as to whether they are more likely to resemble other Native Americans, or the descendants of the original East Asian and Siberian parent stock. At this point the answer is unknown, except that in virtually all studies using various technologies and approaches, the Yakut of southern Siberia are the closest people genetically to Native Americans.

As seen below, chromosomes 2, 9, 10 and 14 look promising in the search for blocks that would take up about half of the right side of the chromosome (via my mother). It is difficult to separate out signal from noise here though since some apparent blocks could be just convergence (random) or hark back to an "Out of Africa" event with little relevance to the question of ancestry within a genealogical timeframe. In addition the autosomes, since there are two, pose difficulties since it would be unknown with a SNP having a CG motif, whether the C for example came from mother or father. This can only be resolved with powerful algorithms and programmes such as PHASE and PLINK which can phase data, placing each allele on the correct chromosome with very high accuracy (as assessed by comparing mother - son combinations). What is evident however, even with a diagram such as that seen below, is that if there is any Asian (Native American) autosomal DNA it will be from my mother's side and hence should be represented in a block with circa 30 to 60% Asian (showing as a spike or block extending half way across the chromosome).

Only preliminary work has been done to date. Neither of the two companies who do the testing provide in depth analysis to be able to compare a customer's genome to specific individuals in the worldwide HGDP-CEPH panel of 52 populations. It is unclear at this point what specific methodology is used by decodeme, but 23andme only compares an individual to three reference groups - a Utah population (representing northwest Europe), the Yoruba (representing Africa), and a sample of Japanese (representing Asia and North America). Clearly this analysis is going to result in a questionable output, and the blue (European), green (African), and orange (Asian) in their "Ancestry Painting" feature needs to be assessed with caution. This is far from anything representing a representative sample. Most individuals have primarily blue, and even those with a Native American grandparent have shown as 5% or less (when something closer to 25% would be expected). Undoubtedly in the future they will upgrade this component but for now there are likely going to be a large number of false negatives (e.g., those with Native American ancestry who show as 100% European). Thus the customer needs to complete the exacting work themselves if they wish anything more than a cursory analysis. Decodeme does, however, offer a browser where a customer can zero in on likely areas of minority ancestry, bracket it, and see the SNPs and exact position of the region. However, if it is Native American haploblocks one is seeking, they may not show on the display.

Anders Palsen assisted the present author by taking the raw data and subjecting it to an analysis by PLINK, a powerful program which separates the SNP pairs to their respective chromosomes (although it cannot determine whether each is from the father or mother - for this other family members need to be tested). It uses 657 Kb of SNPs (hence only a decodeme scan will provide sufficient data for comparison with this panel). The "bar" is set defining a match between two individuals in the dataset of 1049 persons (with the addition of my data into the mix) as a minimum of 1 Mb and 100 SNPs that are identical (shared). This typically results in a number of matches between individuals that is a function of the size of the chromosome (the larger the number of nucleotide bases in a chromosome, the greater the number of matches - except for the X with many times fewer matches e.g., 8, except certain groups such as Native Americans who are often members of one extended family and so have more matches. The search was restricted to only East Asian (any population, there are 15 groups designated as East Asian) and Native American (there are 5 groups) exact matches. All others were ignored. The analysis involved five of the 22 chromosomes as follows:

1) Chromosome 2: Result. 237 standard European matches with the exception of:

Maya (North America - Mexico) HGDP00875. About 2 Mb match (about 3 million nucleotide bases), 178 SNPs, between positions 105,084,214 and 106,996,424. It should be noted that while most of the matches of this individual were to Central and East Asians, and other Native American groups, there was one French, one Mozabite, and one Yourba match. In an instance such as this it must be demonstrated that the above block is truly Native American. In this case all of the neighboring blocks on both flanking sides involved matches only to other Native Americans.

2) Chromosome 5: Result:

Maya (North America - Mexico) HGDP876. About 2 Mb match (about 2 million nucleotide bases), 123 SNPs, between positions 94,563,767 and 96,442,155. Of importance is tha fact that this individual aslo matches a Sardinian and a French in a region overlapping this area. Also a Tuscan and a Basque also overlap in adjacent regions. This analysis, coupled with the knowledge that some of the Maya are admixed, means that this finding cannot with confidence be assigned to a Native American category and so will be dropped from the analysis despite this individual generally matching other Native Americans. It highlights why each match must be closely examined..

3) Chromosome 9: Result: 111 standard European matches with the exception of :

Yakut (Siberia) HGDP00953. About 2 Mb exact match (about 2.3 million nucleotide bases), 186 SNPs, between positions 86,538,538 and 88,868,351. It is important to note that this individual had no matches with Europeans or anyone except a Bedouin, other Yakuts, and only Central and East Asians.

4) Chromosome 10: Result: 94 standard European matches with the exception of:

Yakut (Siberia) HGDP00848. A match involving almost 4 Mb (million nucleotide bases), and 250 SNPs, between postions 91,473,081 and 95,469,789. Again, this individual generally matched only other Yakut or near neighbors, but there was one French and one Adegyi. In examining the block above, there was overlap for another Yakut and a Naxi (East Asian tribe), and end matches in both flanking areas to East Asians only..

5) Chromosome 14: Result: 50 standard European matches with the exception of:

Pima (North America) HGDP01060.
Pima (North America) HGDP01061. The matching here is just over 1 Mb and 101 SNPs, between positions 88,106,272 to 89,230,591. The Pima and most of the Native American tribes of the HGDP-CEPH panel tend to be closely related (often members of an extended family) and the variance between them is low. Almost all of the other matches for these two were other Native Americans, plus a small number of Japanese.

What is perhaps noteworthy is that of all the Native American groups the Pima (from Mexico near the Arizona border), and the Maya (also from Mexico, but Yucatan), are the closest geographically of any of the Native American reference samples to the Iroquoian speaking groups of Eastern North America including the Mohawks (although thousands of miles removed).

Preliminary assessment: The larger blocks may or may not tend to "show up" as Asian in the decodeme browser whereas the smaller ones are difficult to impossible to pick up out of a background of potential noise that may include even 20 to 30% Asian. Hence what is observed may not be found within areas that decodeme, in their browser, has indicated as relatively high Asian. Using the autosome profile in the browser as a guide as to where to locate valid East Asian and Native American matches appears to be questionable (in contrast to the findings with respect to the X chromosome as shall be noted later). The browser uses an unknown number of East Asian groups to make this statistical decision, and unless they have included Native Americans or Yakut then this blocking may have little relevance (unless one is part Japanese). 23andme at this point only use the Japanese in their "Asian comparisons" so it is doubtful that much of what is truly Native American is going to be accurately detected with the output of either company.

It appears that the analysis with PLINK devised by Anders is more reliable due to the specific search for all matches above 1 Mb or 100 SNPs which is an excellent compromise, reflected in the fact that each person of European descent generally has 50 or more matches to others, typically from Europe or the Middle East, in the database. This approach has high face validity. Hence when there is a match that is unique in the collection, and for example matches only those who are East Asian or Native Americans, then this appears to be a "clean" finding of admixture, especially when it can be linked to a known genealogy. However if there had been matches to, for example, a Melanesian from the Pacific, or a Cambodian, although both tend to cluster with East Asians, and in the case of the latter are East Asian, the specific match is not a good "fit". In this case the match must include certain features or it could be considered spurious. Hence it must either be Native American, or an East Asian group that research has shown is among the most closely related to Native Americans. These stipulations have all been met in relation to the author's findings, and to appear to accurately reflect the Native American ancestry shown clearly in the documentary sources.

Specific findings: In relation to the "expected" findings, some data can be obtained via a similation done by one of the testing companies for a customer. Basically the question relates to what percentage of "Asian" one might expect to "see" in relation to a Native American ancestor who is from two to eight generations removed from the testee. In other words, although one might expect that with one Native American grandparent that there would be a finding of 25% on the autosomes, this will of course vary from a theoretical zero to a theoretical 100%. A simulation by one population geneticist using in house data from Native Americans found that the actual figure at k=2 generations is about 14% (range 8 to 20%). It is not known why this figure is so low unless their reference sample was significantly admixed with European. The probability of finding 0% Asian here is zero. At k=4 with one great great grandparent who is Native American where the figure is about 4% (range 0.77 to 6.5%) and again 0% chance of seeing no Asian. A shift occurs at k=6 when the probability of seeing 0% Native American jumps to 16%. At k=8 (the author's situation) the likely percentage is 0.14 (range 0.00 to 0.79) and a 55% chance of seeing 0% Asian. It is not clear as to where they set the bar in terms of deciding what is and is not a match.

Another simulation via a population geneticist, using data from Native Americans (although the tribal affiliations are not specified), and looking just at the author's predicted ancestry (1/256) found that after 8 generations between two and four blocks could be expected, with two being the average. The empirical data via the assessment by Anders tallies remarkably with the above in that two Native American blocks were located (and two Yakut who are the closest Asian group to Native Americans). What is missing in all these analyses is reference data relating to the Great Lakes Region of North America where it might reasonably be expected that more direct and specifically applicable matches would likely be found.

The author's findings by chromosome are as follows:

Chromosome 2: 2 Mb of 243 Mb = .008%
Chromosome 5: 0 Mb of 148 Mb = 0%
Chromosome 9: 2 Mb of 139 Mb = .02%
Chromosome 10: 4 Mb of 135 Mb = .03%
Chromosome 14: 1 Mb of 106 Mb = .009%

An analysis in relation to all of the matches of the author and each of his four Yakut or Native American matches plus information on a European population for comparison purposes is found via a link here.

Conclusions: The average of the above is .0134 or roughly 1% - which is close to the finding of 3% for DNAPrint above, and compares favorably to the "paper number" (predicted from a knowledge of the genealogy) of 1/256 or .003%. It is unclear what would be found via exploring all 22 autosomes, or if the bar was set at .5 Mb, but it would appear that in this case the genealogy and the DNA findings are remarkably similar. While this is far from a "large percentage", the point is that there is a residual trace of up to a few percentage points remaining from many generations ago on at least 4 chromosomes. This deflates "common wisdom" that after 8 generations there will likely be no detectible contribution from an ancestor that far back. A reasonable conslusion is that each of our ancestors leaves an imprint of some description that could extend back to the "Out of Africa" event.

It is also noteworthy that the only non - European findings were exclusively with Native Americans or those in Asia who are consistently noted in the literature as bearing a very close genetic relationship to Native Americans since both likely came from the same ancestral stock. The East Asian tribe most closely associated with Native Americans are he Yakut who are the only East Asian tribal groups to match the present author when the level was set at a minimum 1 Mb and 100 SNPs.

Research design and statistical analysis
by Anders Palsen.

GENEALOGICAL EVIDENCE

NATIVE
AMERICAN
ANCESTRY

Although the author's "paper percentage" of Native American ancestry is low, the documentary evidence for what may be reflected in the DNA testing creates something of an "embarassment of riches". It is unfortunate that many North Americans have little more than family tradition or suppostion, this will not go a long distance unless supported by a paper trail that is secure. What follows is a listing of some of the documentary sources serving as evidence of specifics of the author's Native American ancestry, and of a selection of some of the details found in these sources:

1) The standard genealogical documentary records and sources were used to follow the lineage of the author back 8 generations to Lt. John Young (Johannes Jung) of the Six Nations Indian Department during the Revolutionary War, and his wife Catharine. They resided on the Six Nations Indian Reserve at the Mohawk Village and on a farm near Cayuga, Ontario. This file of information was submitted to the United Empire Loyalists' Association of Canada as part of a petition for membership approval in this lineage society. The application was approved in 1975.

2) In 1792 a Scottish traveller, Patrick Campbell, visited the Young family on the Grand River and recorded entries in his diary which was published in 1793. Here he stated that Mr. Young served as a lieutenant in the Indian Department in the "last war", and was "married to a squaw, sister to one of the chiefs of the Mohawke nation who succeeded Captain David" (p. 180). He further added, "Here I for the first time played cards with a squaw". At another location in the diary Campbell notes that Aaron Hill, "eldest son of the renowned chief, Captain David, whom everyone who knew him allowed to be the handsomest and most agreeable Indian they had every seen; he died about two years ago, and, what would be deemed hard by many, the son does not succeed to the honours and titles of the family, but they go in the female line to his aunt's son" (pp. 166-167). In sundry sources it can be seen that Captain David Hill's usual name was Karonghyontye, but that he also held the sachemship Astawenserontha - both names being included in a deed of sale to the Van Horne Patent in New York to Jelles Fonda, 6 July 1789. Other record sources (e.g., RG10, Indian Affairs Series, Archives of Canada) indicate that Hill was succeeded by his nephew Seth Hill Kanenkaregowa [son of John Widemouth, Seth's son and Mary Hill Katehriunigh], who became Captain Seth Hill Astawenserontha. Catharine's brother inherited both the military title "Captain", and one of the three Bear Clan (Tribe) titles among the Mohawks, Astawenserontha (the name translates as, "He enters with rattles on"). See sources outlined in Faux, 2002.

3) In a land deed of 1838 from the Six Nations in Council, which includes their signatures, the chiefs granted land to Joseph Young of Young's Tract on the Grand River, the youngest son of Lt. John and Catharine. The original deed is among family records, and a copy is registered in the RG10 (Indian Affairs Series) at the Archives of Canada. In the deed they specified that Joseph was to have a parcel of land adjoining one in the "possession of the said Joseph Young one of our people". The term "one of our people" clearly indicates that the chiefs recognized Joseph as a member of the Six Nations.

4) Joseph's nephew Col. Warner H. Nelles, was born 2 May 1799 on the Grand River Indian Reserve, (son of his sister Elizabeth (Young) Nelles), and is the first cousin of the author's ancestor Rachel (Young) Young. His 12 October 1896 obituary reads, "His popularity as a young man may be judged when, being only seventeen, he was made a chief of a tribe of Indians called the Beaver. Upon the death of the old chief, they saluted him as Tahanata, and the chain of wampum they threw over his head is still in the family. Up to the end of his death he was still recognized as chief" (Haldimand County Museum, Dorothy Hutton Collection N-1-4). Cross validation of the details of number 2 above can be obtained by a close inspection of each of the components noted in the obituary. First, a list of "Mohawks Principle Chiefs" created by Chief George Martin for the Indian Department 22 February 1815 (i.e., about two years before Nelles was made chief) included the names of the holders of the three hereditary sachemship titles for each of the Turtle, Wolf and Bear tribes (clans), and the associated war chiefs for each sachem (Archives of Canada, MG19, F1, Claus Papers, Vol. 10, p. 153). Number 3 in the Bear Tribe is John Johnson Astawenserontha (Captain Seth Hill died in 1808) with the associated "war chief" John Green Aronghyenghtha (a cousin of Catharine (Hill) Young). Oddly, here the Bear Tribe has four hereditary sachem titles with 3. Astawenserontha and 4. Tahatonne linked together via a bracket thus }. Since time out of mind there have been three Turtle Clan titles, three Wolf Clan titles, and three Bear Clan titles - and no more. Reg Henry, Cayuga, Six Nations linguist, translated Tahatonne to mean, "He is coming towards you saying over and over" (21 November 1980). Francis Cotter held the latter title in the Martin document of 1815. Two war chiefs appear to be associated with this sachemship, Adam Joghryogea and John Frazier. In 1877 "John Fraser" held the title Astawenserontha (Chadwick, 1897). The spelling differences (Nelles obituary - Tahanata; and Martin document - Tahatonne) apparently reflect a switching of the last two syllables in the name as remembered likely by his children or grandchildren in the late 1800s, all of whom were born and raised away from the Reserve in St. Catharines, a Canadian city. They also mistook Beaver Tribe for Bear Tribe in their reporting for the obituary. So evidently in 1815 there was a fourth Bear Clan title and associated with the ancient sachemship, Astawenserontha, created perhaps at at the beginning of the War of 1812. It appears that the entire Cotter family left the Reserve for the west about this time (1815) to take residence among the Wyandots (RG10, Indian Affairs Series, Archives of Canada). This likely ties in with the settlement of some Mohawks in Wyandot territory along Honey Creek Ohio between 1815 and 1817 (Sturtevant, 1978). This in turn may have been precipitated due to Francis Cotter, according to the Six Nations Chiefs, being among a group who acted in an "Unbecoming manner" by refusing to fight the enemies of the King, and so they recommended that he not receive any of "His Majesty's bounty" - annuity payments. Also on this list drawn up 9 January 1814 was Isaac Brant (Johnston, 1964, p. 219), a step-nephew of Captain Joseph Brant, who at this time appeared with his wife and four children on the Wyandot Reserve along Honey Creek where a Mohawk Village was established. Hence, as of about 1815, Cotter and others were in effect "shunned", and doubtless Cotter would have been "dehorned" - his chiefship removed and given to a more suitable candidate from within the matrilineal family. Apparently in 1816 or 1817 the Clan Matrons asked Elizabeth (Young) Nelles, whose mother was in the direct line of succession as Clan Matron, and would have been responsible for selecting any Astawenserontha successor had she not died early, to step in and select one of her eligible kin to fill the vacancy. As is typical, she chose one of her sons, in this case Warner H. Nelles. The obituary informants were accurate in their reporting of the manner (recalling of procedural details) of installing a new chief among the Six Nations. During the Condolence Ceremony, when the new chief is "raised up", the string or strings of wampum are used as a sacred emblem to "crown" the candidate chief, puting the wampum around his neck. The new chief then keeps the wampum strings as a certificate of office (e.g, Beauchamp, 1901).

5) Elliott Moses was a chief of the Delawares (his ancestry also included Lower Cayuga) who resided on the southern block of the Six Nations Reserve, and was perhaps at the time the most knowledgeable individual on the Reserve concerning the history and traditions of Six Nations and Delawares. While on a field trip with Dorothy Hutton a local historian of Haldimand County 23 June 1967, he pointed to the house on the hill where the above Warner H. Nelles was born and Hutton reported that Moses said that, "his grandfather or before knew that there was some or a bit of Indian blood in the Nelles' who lived on the other side of the River on the hill" (where the above Warner H. Nelles was born). Hence even up to modern times the link between the Young family and the Six Nations was still recognized by elders on the Reserve.

There is a lot more evidence, however it would appear that the above alone is sufficient to demonstrate a genealogical link to the Mohawk Nation. The Hills were members of the Lower Mohawk band at Six Nations (near Brantford) and Tyendinaga (near Kingston).

Mohawk Family Tree of the Author

X CHROMOSOME ANALYSIS

While the above analysis has brought forth autosomal sequences that can be linked to Native American (hence Mohawk) ancestors, the percentages are somewhat "uninspiring". Curiously the X chromosome has an entirely unique pattern of inheritance. Perhaps more substantial evidence of Native American heritage will emerge via an analysis of the DNA haploblocks on the X chromosome.

Link to the Analysis of X Chromosome Haploblocks