Documenting my family's past for future generations. My family tree includes the Smith/Mansell families of Alabama and Oklahoma, the Castle/Day families of Kentucky and Oklahoma, the Wheat/Ming families of Texas and Oklahoma, and the Bell/Roberts families of Mississippi, Tennessee, and Oklahoma.

Monday, August 19, 2013

Where in the World Am I From?

I’ve been waiting a couple of months to upload my Family Finder results from Family Tree DNA to Gedmatch.  Gedmatch lets you compare DNA results from several companies, as well as offering you different tools to help you interpret your results.  Gedmatch exists on donations and has been down for a while upgrading its server to handle the ever-increasing numbers of DNA results uploaded by eager researchers, trying to understand what they mean.


It takes Gedmatch a few weeks to chart all your matches, but while you’re waiting, you can play with some of the tools.  You can match one-to-one with someone, as long as you know his/her kit number.  Since my brother’s Family Finder results have just come back, and I’ve uploaded them to Gedmatch, I could compare where we match on each chromosome.  That’s not so useful with my brother’s results since I already know how we’re related, but it does help me become familiar with how Gedmatch works so I can use the tools when my other matches come up.  I’ve got a couple of DNA gurus among my Huff cousins that I hope will teach me how to use all the tools, once they are all functional.

What’s really been fun is working with the Admixture (ethnic percentages) tools.  I’m not an anthropologist or a mathematician, but I’m guessing that the creators of the different admixture models are.  It’s really interesting to see how they differ with each other.  I’m guessing that the differences exist because the creators of the models compared against different populations or used different mathematical formulas to determine the percentages of ethnicity.  What’s fascinating is how closely they do match.

In every model, I’m at least 50% Northern/Northeastern/or Western European and from 26 to 30% Mediterranean.  The Mediterranean is sometimes further differentiated as Neolithic, which doesn’t surprise me, as my mitochondrial haplogroup, T2, came from northern Italy in the Neolithic period.  As most of my known ancestors can be traced back to England, Wales, Scotland, Ireland, or Germany, the North/Northeast/West Europe percentage does not surprise me either.  I was excited to see on most models that I am at least 1% Native American.  Combined with other Asian percentages, as Roberta Estes does in calculating Native American heritage, it is even more.  On one model it is broken down as .49% American, .53% Beringian, and .23% Siberian.


Ancestry.com also does its own "Genetic Ethnicity Summary."  Mine shows 44% Central European, 33% British Isles, 21% Scandinavian, and 2% Uncertain.  Again, not too surprising, except for the large percentage of Scandinavian.  This appears to be a flaw in the Ancestry.com formula for calculating ethnicity, as has been discussed online by Roberta Estes and other genetic genealogists.

In my next post I’ll talk about comparing my brother’s Family Finder results with mine and using FTDNA’s Chromosome Finder to pinpoint the chromosomes on which relatives with particular surnames may occur.  I hope when we are both fully up on Gedmatch, we will find even more to help us extend our family tree.  

Saturday, August 3, 2013

Will the Real Ancestor Please Stand Up?

Two years ago I gave my brother a DNA test for Christmas.  In my defense, 1) it wasn’t his only present, and 2) even though it seems like I bought a gift for myself, we really do have a mutual interest in finding out more about our Smith line.  I was born a Smith, and my brother is still a Smith, making him perfect for yDNA testing.  

I have always loved genetics.  I remember learning in school about Gregor Mendel, the monk who tracked the inherited traits of peas and became the father of the science of genetics.  I even liked doing Punnett squares, those little graphs that predict the traits of offspring of a particular set of parents.  When I became interested in genetic genealogy, I had to do homework again--about what tests are available and what they can tell you about your family tree.  The field of DNA testing has new developments all the time, and I’m still learning about testing and interpreting test results.

The first test I took myself was a mitochondrial (or mtDNA) test.   Mitochondrial DNA passes from mother to child, uncombined with the father’s DNA.  Only daughters can pass it on to their children, so what mtDNA does is trace your matrilineal line, your mother’s mother’s mother’s mother, back into the distant past.  Since a woman’s surname usually changes in each generation, it can be difficult to use mtDNA to find ancestors by name.  What you get instead are haplogroup matches—others who show the same mutations from the original mother of us all, who the geneticists call “mitochondrial Eve.”

My haplogroup is T, one of the 7 or 8 groups that are really common in Europe.  One of the first companies to test mitochondrial DNA for the general public was Oxford Ancestors.  You may have heard of its founder, Dr. Brian Sykes, because he wrote The Seven Daughters of Eve, in which he explains mtDNA testing and describes the seven clan mothers from ancient Europe who still have descendants today.  My clan mother is Tara, who lived about 17,000 years ago in northwest Italy.  (Geneticists look at how Tara’s ancestors dispersed and how many mutations have occurred in her line to help them estimate where her clan began and how long ago they lived.)  Really cool, but not especially useful in helping you fill out your family tree.

Back to my brother’s Christmas gift.  His paternal line consists of our father, Jack Francis Smith; his father, Weaver Harris Smith; his father, Stephen Albert Smith (for whom my brother is named); and his father, John A. Smith.  Yes, that’s right—John Smith.  That’s as far as we know, and all we know about John A. Smith is that he was born about 1805 in Virginia and died about 1850 in Pike County, Alabama.  We hoped that my brother’s yDNA test would help us differentiate our John Smith from the many others with the same name and lead us to his parents, grandparents, etc. 

While mtDNA traces the matrilineal line of both men and women through mitochondrial DNA, which both sexes have, yDNA traces a man’s patrilineal line, his father’s father’s father, etc. through the genetic information encoded in his Y-chromosome.  While a man can take an mtDNA test, a woman cannot take a yDNA test for the very important reason that she doesn’t have a Y-chromosome to test.   yDNA is a lot more useful for identifying ancestors because surnames remain the same from generation to generation. Hopefully, we would connect with another Smith (we joined a surname project called Smiths Worldwide) who matches my brother genetically, and maybe he knows who John A. Smith’s father was.  Mystery solved.

Except my brother’s closest yDNA matches—he matches one of them on every one of 37 markers—are three men named Banks.  What??  I assumed the worst—what genetic genealogists call a non-paternity event (NPE).  I figured that sometime, somewhere, a female married to a Smith was unfaithful with a Banks man, and the result of that union took the surname Smith.  I even looked for Banks families that lived near our Smith family in Pike County, Alabama, although I don’t know that that is where the alleged NPE took place.  (There actually was a Banks family in Pike County, but now what do I do with that information?  The family tree of that Banks family doesn’t match the family trees contributed by the Banks men who matched my brother on the yDNA test.)  So, there the results sit, on the Family Tree DNA site, just waiting for additional information that makes sense of them.

Until this week, when one of my Internet cousins suggested I read a blog by genetic genealogist, Roberta Estes, at www.dna-explained.com.  In fact, I have read the entire archive of her fascinating blog posts, but one post, “Surprise Y Matches—What Do They Mean?” has been especially helpful.  According to Roberta, there is more than one reason that one man might have matches with another man with a different surname, and this occurrence is not that uncommon.

The fact that my brother and Mr. Banks match at 37 markers means that they have a common ancestor at some point in the past.  However, it’s possible to test 67 and even 111 markers.  If my brother and Mr. Banks were to upgrade to more markers, some mutations might show up that suggest that their common ancestor is not recent but way further back in time.  Testing of other cousins might prove either line and show that what we are dealing with really is an NPE.  Roberta also suggested what I had already done—look for Banks families in proximity to our Smith families in locations where we know they lived.

Other reasons for the mismatch in surnames?  1) Someone just decided to change his last name.  (What better name to change to than Smith?)  2) A step-father raised a boy that took his name, rather than his biological father’s.  3) A boy took his mother’s maiden name, perhaps because he was illegitimate.  4) Someone was unfaithful.  And something I actually had not considered—that the NPE, for whatever reason, might have occurred on the Banks side, not the Smith side.  Before DNA, those circumstances might never have been suspected, and the answers would never be found.

The recent sale at Family Tree DNA ($99) has encouraged my brother to purchase the Family Finder test.  I’ve already had the FF test done but have yet to find any conclusive Smith or Banks cousins.  Family Finder finds matches through a third type of DNA—autosomal.  This is the DNA that most of us are familiar with from those Punnett squares, the kind that is contributed by both father and mother and combines in different ways to make each of us the unique person that we are.  

The test produces a lot of data that I am still learning to interpret, but Family Finder does a lot of the work for you by listing your DNA matches, typically cousins and their suggested relation to you—2nd-3rd Cousin, 4th Cousin, etc.  Our hope is that my brother’s test may give us some more options to trace our Smith line.  Because of the nature of autosomal DNA, my brother may have inherited more on our father’s side and will show up with more conclusive cousin matches that will give us some clues about the ancestry of John Smith (or, I suppose, prove that we aren’t related to him at all.)