NC Myers and TN Myers. Weaving it together into a speculative whole.

I was looking through my Dad’s autosomal records, looking for Myers to ask to do a YDNA test, if they were not exactly in our line. I was hoping to find another Ross David Myers or Artie Myers, because DNA analysis doesn’t act vertically on the known relatives, it acts horizontally, it parallelizes the search because you find close relatives you never expected.

Neil O. Myers is a distant autosomal match of my Dad’s, a 7 cM segment of DNA shared with him. Not shared with me. But vastly important is that he is a direct descendant of John Myers 1712 who married Ann Bruce, and even more so, had a defined haplotype.

Neil’s haplotype is R-YP639, which is significant as it’s a subclade of R1a. I’m R-By15581, a subclade of R1b. R1a and R1b do not mix. So we now have definitive DNA evidence that John Myers of NC born 1942 (maybe) died 1826 in Mississippi is not part of the John Myers/Ann Bruce tree.

Neil, btw, has a book on the SC Myers, which Ross has purchased and looks pretty good.

In the meantime I heard from Rayedene Graves, Artie Myers’s relative and guardian of his DNA data, and it was really great stuff she had to say.

Artie is a descendant of George Myers. George Myers is the link to NC, and his father is, by her telling of the tale, the son of Captain William Myers, the one who dies in 1923 in Greene Co TN. That makes him the William Myers Sr so often spoken of in Green Co lore (see comments on this article). Charles Myers is thus related to George Myers who is on record as saying his dad was Captain William Myers. And the William Myers’s are all in Greene Co. And comments by other folks that they moved most likely as a family group are perhaps more realistic now.

If you try to look up William Myers d 1823 on Ancestry trees, you find a ton of folks who mention a William Myers b 1770 Somerset Co (that county again) and have him die in 1823. I guess they missed the 1815 death notice.  If the 1770 figure is correct, it’s not for William Sr, but rather for William Jr.

So please indulge me a little bit. Let’s suggest that “our” Charles is literally the Charles Jr mentioned in William Myers Jr’s will, and that he is born in 1789 in Somerset County PA. That would make the putative father, William jr, about 19 years old. Charles Myers Sr would be a brother of William Jr, who because of 1840 US census data (Van Buren Co, MO), had to be born between 1781 and 1788.  Then Charles Jr is the one who marries Rebecca, in this telling. What isn’t clear is the relationship between William Sr and Jr, and it is possible that a brother of William Sr is the father of William Jr.

And if this is all true, then the connection between the Williamses is proven by the Y37 match between Artie, myself, and my dad.

This isn’t the only way these data can be interpreted, but with more facts, the story just seems to get better and better.

Update – my assumption, that “our” Charles is Charles Myers Jr,  can be disproved by data from Pleasant Hill Cemetery in Cass County Missouri.

Phasing DNA. Pros and Cons.

I had my dad get an Autosomal test at ftDNA (whatever their merits or demerits, Family Tree DNA keeps your DNA sample for 25 years, so you can update testing anytime you want), and I have transferred that test to Gedmatch. His kit number is posted on Wikitree ( which is another genealogical resource you might find useful). But to repeat them here, Ysearch kit is MMB63. His autosomal kit on gedmatch is T141705. Anyone who has a gedmatch login can do ‘one to many’ comparisons with his gedmatch kit.

That said, with my Dad’s kit there, I can phase my own DNA kits, splitting them into a maternal and paternal side. Using these, I can now tell whether a DNA match is matching from the maternal side or paternal side. So, when trying to dig up dimly understood ancestors, this is quite valuable.

On the other hand, the admixture of my maternal side is a bit fanciful..


And the answer is, no I don’t believe it. The sanest of the admixtures so far seems to be 50% Scots/Irish/British, 30% Scandinavian, and 20% Southern Europe. That’s what I tend to believe, as of the moment.

And yes, they do warn you that phased data and admixture are not the best of ideas. Compare with my admixtures in previous posts.

gedmatch admixture testing.

Gedmatch allows you to test autosomal data against a variety of autosomal testing tools. The value of this is to see how different reference populations , and different DNA tools, affect the outcome. Some sets of reference populations give very different results when tested.

This is a result using the Eurogenes program and their most popular population model.


This is the result when I use the Harappa World model. This is a population set that favors people of African, Middle Eastern, and South Asian ancestry.


Note that if my father were Lebanese and my mother came from the Sudan, this would have been a far better model to pick. I’ve noted in my reading that Americans of African descent often prefer the Harappa World model to the others in the gedmatch set.

The point is, depending on where you come from, the commercial admixtures may or may not best represent what your genetics are. In particular, we’ve already noted there are marked differences in how someone with Micronesian bloodlines is handled by the major commercial services versus Nat Geo’s Genographic Project.

So you’re mostly Micronesian and want to know how Asian and European you are?

My wife is half Chamorro and half Japanese, so on her side of the family they generally have two questions. Is the paternal ancestor European or Chamorro, and how much European blood do they have? On top of that, a sane admixture result (how much Japanese, how much Chamorro, how much European) was desired, so that reasonable guesses could be made as to how far back the European blood was introduced.

We ran into the issue first with Ancestry DNA. They had her mostly Asian, and a hodge podge of other assignments. Dissatisfaction with the admixture results from Ancestry led us to export the data from Ancestry and put a copy of this data up on gedmatch. Testing different Admixtures with the gedmatch tools made it clear that without a reference population of Southeast Asians/Oceania, then any real attempts to separate these two was not going to work. Ancestry DNA evidently has a Polynesian reference but not really a true Austronesian reference, which really is needed for any people populated by the Austronesian expansion.

Nat Geo Genographic was pretty excellent. They have a Southeast Asia/Oceania reference population, and it has my wife with about 53% Asian, 36% Oceania, and 9% European.

Upon transferring that data to Family Tree DNA, that separation has been lost. She’s now 80+% Asian, 9% European, and 1% Native American.

So, in short:

Ancestry – probably OK if you’re Polynesian. If you’re Melanesian, Micronesian, Filipino, Vietnamese, from New Guinea, Ceylon, or Madagascar, probably not.

Net Geo Genographic – the best of the three, with a sane reference population

Family Tree DNA – in terms of admixture percentages, not worth it if you’re descended from Austronesian people.

Note: my wife, in various gedmatch tests, often turns up substantial percentages of Siberian blood. The Siberians are the population from which Native Americans came, so the 1% Native American says that there are still traces of genes found in Native American populations.