DNA Testing for Genealogy – Getting Started, Part One

Posted July 18, 2012 by Geni | 31 Comments

We’re excited to bring to you a special guest blog post by genetic genealogist CeCe Moore. Some of you may recognize her from her blog YourGeneticGenealogist.com. For the next few weeks, she’ll be providing a great overview about DNA testing for genealogy. So without further ado, here’s part 1! Enjoy!

Interest in DNA testing for genealogy has reached an all-time high thanks to its increasing sophistication and the resulting visibility in the media. We hear about what we can learn from DNA testing from popular genealogy television programs, news stories and in advertising.  As a result, many family history enthusiasts have expressed their desire to venture into the fascinating world of genetic genealogy, but don’t know where to start.  If you are one of these people, then I am writing this for you. In a series of four posts over the next month, I will explain the three different types of DNA testing currently used by genealogists to discover more about their family trees. I will endeavor to help you determine which test or combination of tests would be best suited to your interests.

The first step is to determine your testing goals.  Do you have a general curiosity about genetic genealogy or is your focus more specific? For example, please consider the following questions:

  • Are you primarily interested in researching your surname?
  • Are there specific brick walls that you wish to target with the use of DNA testing?
  • How far back in your family tree are these brick walls?
  • What is the ancestral pattern back to these brick walls, i.e.- mother’s mother’s    mother or father’s mother’s mother’s father?
  • Are you ready for a long-term project or do you desire quick answers?
  • Are there adoptions in your family tree that you would like to explore?
  • Is your primary interest receiving a percentage breakdown of your overall ancestral origins or “ethnicity”?

Currently, there are tests geared at isolating three different types of DNA that can address these questions and others. They are Y chromosome DNA (Y-DNA), mitochondrial DNA (mtDNA) and autosomal DNA (atDNA).  Y-DNA has been in use the longest and, as a result, has the best track record for helping genealogists demolish those proverbial brick walls, so let’s talk about that one first.

Focused on your surname?

Y-DNA refers to the DNA found on the Y chromosome. Only males inherit the Y chromosome, so this test can only be used to trace the direct paternal line.  Simply speaking, a man inherits his Y chromosome from his father who inherited it from his father who inherited it from his father and so on. Testing of Y-DNA can reveal the ancient origins of a person’s direct paternal line. It can also be of great genealogical value since the Y-DNA traditionally follows the same inheritance pattern as surnames. As the surname is passed down from father to son, so is the Y-DNA. Because of this, a male can often identify which direct paternal ancestral line he belongs to by matching with others of the same or similar surnames. For those with ancestral roots in societies with traditional surname inheritance, this type of testing can sometimes help identify common ancestors from several hundred years ago. Though far from guaranteed, an adoptee may even find success in discovering his biological surname by testing his Y-DNA.

If you are a woman who is interested in discovering more about your maiden surname, then you will need to locate a male of that surname from your family. You can test, for example, your father or brother (full or paternal half), a paternal uncle or a paternal male first cousin. If needed, you can even test a more distant cousin as long as he carries the surname of interest and shares a common ancestor along your direct paternal line.  Remember, you cannot test your son for this purpose, since he carries the Y-DNA of his father!

Although this test is limited in scope, it can be used in ways not immediately obvious. For instance, you can identify any specific surname of interest in your family tree, follow that surname down the branches toward present day and use these same rules to try to locate an appropriate male for testing.  This way, the usefulness of Y-DNA testing in your family research can be expanded beyond this one ancestral line.

Perhaps the best news about Y-DNA testing is that there are thousands of surname-specific DNA projects headed by volunteer genetic genealogists to help you with your DNA research. With the birth of the first commercial genealogy DNA testing company, Family Tree DNA, in 2000, surname projects immediately began springing up. Some have already had great success in determining the origin(s) and family branching of their particular surname of focus, while others are still searching for enough participants to reach solid conclusions.  You can participate and assist in focused research on your surname by joining the appropriate project. To locate your surname project, you can try Googling  “(Insert surname) + DNA Project”. If you find that one does not exist, you may want to consider starting your own and recruiting those genealogical cousins that many of us are already in contact with through our family history research.

As in all genealogy, I do want to point out that immediate gratification is not guaranteed. Some participants have not yet found success with this type of testing due to the lack of matches to their DNA in the databases. However, time should rectify this problem for most with the growing popularity of testing and the increasing numbers of those contributing their DNA to this research. For cultures that did not utilize inherited surnames in the traditional way, Y-DNA testing will not be as effective, but can still be useful in identifying probable origins and migration patterns.

The Y-DNA tests on the market currently offer anywhere from 12 markers to over 100 markers. I suggest testing at least 37 markers to start with (around $150). In the simplest of terms, “markers” are specific spots on the Y chromosome for which numerical values are assigned based on each individual’s DNA signature. So, for example, if you order a 37 marker Y-DNA test, this means you will generally receive a string of 37 numerical values. A man’s closest matches may be found by comparing his string of numbers to all of the other results in the company’s DNA database.  As a general rule, the men who have the most markers that match each other are the most closely related to each other.  The approximate number of generations back in time that two men shared a common paternal ancestor is estimated by determining how closely the individual strings of numbers match each other.  (Don’t worry; the testing company will automate these calculations for you.)

Since this test exclusively addresses your direct paternal line, remember the results are only informative of a very small portion of your family tree. If you are interested in a more general overview of your ancestry, please check back for my post in two weeks on autosomal DNA tests.

Next week, we will take a look at mitochondrial DNA testing. (Update: Read part 2 here).

CeCe Moore is a professional genetic genealogist and writes the popular blog Your Genetic Genealogist, She is the administrator of several surname DNA projects and the ISOGG DNA Newbie Yahoo Group. CeCe serves as an “Ancestry Ambassador” to 23andMe, on the Advisory Board of the Mixed Roots Foundation and is a member of Mensa. Her favorite genetic experiment is her seven-year old son, Nicolas.

Share:

  • Andrew Colman

    What an admirably clear and sensible introduction to Y-DNA genealogy! I look forward to the sequels. Thanks CeCe!

    • CeCe Moore

       Thank you Andrew!

  • israelp

    I am trying to get a handle on testing specific degree of cousinhood. For instance, I have two people who I think are brothers. We are testing the second great-grandson of one (67 markers) and the third great-grandson of the other (37 markers). Results are in for only twelve so far and the match is perfect. Is it possible to get results which will demonstrate conclusively that these two are in fact brothers and not cousins or uncle-nephew?

    • Ersatzharry

      Conclusively?  Probably not at the 37-marker level or even the 67-marker level, though the answer depends on the mutation rate of the set of markers tested and the number that are tested.

      To illustrate (and simplifying away a lot of things), suppose the mutation rate is 1/200 for the sake of argument.  Then the probability of a son matching his father perfectly over 37 markers is (1 – 1/200) raised to the 37th power, or about 83%.  The probability of a second son matching perfectly is also about 83%, so the probability of both sons matching the father perfectly is 83% x 83%, or about 69%.

      You can use the same logic with a pair of paternal cousins, and the probability of them both matching their paternal grandfather perfectly over 37 markers would be about 48% using the 1/200 mutation rate.

      If prior information led you to believe that there was a 50% chance that the two people were brothers and a 50% chance that they were first cousins, then incorporating the test results would move that 50% up to about 59% in favor of brothers rather than first cousins.  I don’t think most people would call that conclusive.  Crank up to 67 markers, run through the same calculations, and the result would go to about 66% instead of about 59%.

      You can vary these numbers a bit, but the answer to your question will likely remain no.

      • CeCe Moore

         Thanks for your input Ersatzharry!

      • israelp

        At first glance, your name looked like Esterhazy, who was the real spy in the Dreyfuss case.

    • CeCe Moore

       @israelp:disqus  – I gather from your explanation that the two testers are third cousins once removed. Y-DNA will not give you a conclusive answer on this question. Unfortunately, this relationship is just beyond the point at which autosomal DNA can be quite conclusive. Beyond second cousins the genetic inheritance gets more random and thus more difficult to predict exact relationships. For third cousins the chance of having an autosomal DNA match is 90% so for third once removed it is a bit lower. Third cousins once removed should share on average .39% of their DNA. Fourth cousins once removed would be expected to share very little DNA if they match at all. So, if they matched at a level around .39% (or higher) you would have some support of the closer relationship, but it would not be conclusive since we sometimes see higher sharing than expected in more distant relationships. If they do not show a match, you could not reach any conclusions based on that since even at the 3rd cousin level one in ten will not show a match at all. I know this isn’t what you were hoping for, but I wanted to explain this since autosomal DNA testing may have occurred to you as a possible option to explore for answers.
      I believe the autosomal DNA article will be posted here on August 1st, so check back if you are interested in learning more.
      Thanks for commenting!

      • israelp

        Thank you and to ErsatzHarry. I must digest this.

        We are talking putative fourth cousins once removed, not third.

        We have also done some autosomal testing – including aunts of the two Y-testers, who came out as “suggested second cousins.”

        • CeCe Moore

          Hmmm…you wrote “we are testing the second great-grandson of one and the third great grandson of the other”. Do you mean great great grandson vs great great great grandson? That = 3rd cousins once removed. The reason this really matters is because you tested the older generation on autosomal, which is GREAT news! If the men are actually third cousins once removed, then the aunts are 2nd cousins once removed, thus getting us into the reliably predictable area of autosomal DNA. If they are predicted to be second cousins, I would say that you have very good confirmation that the men in question were brothers and not cousins, etc… May I ask how much DNA the aunts have in common? (% at 23andMe or total cMs and longest segment from FTDNA) If it is enough, it may give you a conclusive answer.

          • israelp

            If grandsons of brothers are second cousins then great-grandsons of brothers are third cousins, then great-graet-gransons of brothers are fourth cousins. So these should be fourth cousins once removed.

            The aunts share 141.64 with the longest piece 37.09. FTDNA calls them third cousins.

            We also have another pairing of one aunt with a first cousin of the other aunt – 186.54 shared, 41.98 longest. FTDNA calls them second cousins.

            Obviously I have no idea how to read all of this.

          • CeCe Moore

            Okay, got it now. I was missing the brother part and just focusing on the gggrandson part. Sorry.

            That’s a lot of DNA! You’ve got me curious. (Can you tell I love this stuff?!) Please add up only the segments over 5 cM for each and write to me at yourgeneticgenealogist@gmail:disqus.com. (You don’t have to pay a consulting fee.) :-)

          • israelp

            If I should be talking to you off-site and paying you a consulting fee, don’t hesitate to say so.

  • Don

    Thanks for this great article!  I’m very interested in doing my first DNA test but I’m very conflicted on which company to use.  I would think that it’s very important to use the company with the largest database, because a small database of users will not produce as many matches.  Do you agree?  And if so, is there any marketshare data to help me narrow down my choice?

    • CeCe Moore

       @Don – Yes, I definitely agree. If you are thinking about taking a Y-DNA test, then Family Tree DNA has by far the largest database and is where most host their surname projects. If you are considering an autosomal DNA test, it is a bit more complex than just the size of the database. 23andMe has the largest database, but many in it tested for health purposes and are not genealogists. Don’t get me wrong, I’m a huge fan of 23andMe. I do receive many inquiries in regard to family history and there is a very active genealogy community over there, but it is a bit like comparing apples to oranges when you try to compare the database size to FTDNA”s Family Finder’s. I will have an article here covering this type of testing on August 1st, I believe. You can write to me at yourgeneticgenealogist@gmail.com if you need more specific advice before then.

      • Don

        CeCe, thank you so much for this helpful reply and also for the additional articles you’ve posted.  This is great stuff!

    • CeCe Moore

      Sorry I missed this, Don. Yes, the database size is very important. For Y-DNA, Family Tree DNA has the largest database by far and I definitely recommend them. For autosomal DNA, 23andMe has the largest database, but many there are not genealogists, so it is a bit difficult to compare their database versus Family Tree DNA’s “Family Finder” database, which is primarily genealogists. I really like both tests for different reasons. If you are at all interested in the health side, 23andMe is the way to go and you can then transfer your raw data to FTDNA’s database for a reduced cost. AncestryDNA’s autosomal database is growing fast, but it is only offered to subscribers by invitation.
      I hope that helps.

  • T. H. Roderick

    “Y-DNA has been in use the longest and, as a result, has the best track
    record for helping genealogists demolish those proverbial brick walls,
    so let’s talk about that one first.”

    Actually mitochondrial Eve has been around longer in the scientific and genealogical literature.  The mutation rate of the Y chromosome STRs is higher than that for the mtDNA markers, making the Y chromosome approach better for determining relationships in historical and pre-historical times
    T. H. Roderick

    • CeCe Moore

      Thanks for the clarification, Thomas. Although I wasn’t referring to usage in literature, to make my point more clearly I should have written, “Y-DNA has been used most frequently by genealogists.”

  • kjaffa

    For Ydna when calculating the MCRA, is the mutation relate thought to be the same for each marker? What is the best way to estimate your MCRA if you only differ say by 2 out of 67 markers?

  • Carol

    CeCe enjoyed reading your article. But am still a little confused, is there a DNA test that could help my mother ID her birth father?

  • hosmer

    I am interested in finding out how much native american blood i have in me . have been told i am cherokee, blackfoot , iroquois, and mohawk . being a woman can i do this on my own as i want to learn more about my indian heritage.

  • http://www.facebook.com/shirley.tomlin.98 Shirley Tomlin

    About one out of six females get a y from their father. They are the female phenotype as they are totally insensitive to testosterone. Therefore in their hormones they are completely estrogen. They do not have ovaries. Some have a uterus. So it is wrong to say that only males get Y from their father. However I don’t know if any xy females have ever had their y tested for genealogy purposes.

    • Jo Jesty

      Your incidence is wrong by more than 3 orders of magnitude. On average less than 1 in 10000 phenotypic females are androgen-insensitive, i.e. they have Y chromosomes.

      • anoldlady

        you might be right. My information of one out of six came from a retired ob/gyn doc. I thought it was was high.

  • cc

    Hello

    I am writing my family newsletter this month may I put your column in it?

  • rebecca

    i sent my dna samples a while back, so now what?

  • Suzanne Saxton-Frump

    I am totally at a loss of what to do. I know I want y chromosome testing, my father is quite old and the only remaining male relative. I have brick walls in England starting around the 1400-1500s and going back. One surname is Hinckley, last governor of Plymouth Colony, another is Saxton. How should I proceed. Wondering if I need to have testing done by a British company or where to turn. Thanks so much. Suzanne Saxton-Frump

  • Jo Jesty

    Great stuff! If you have time I’d appreciate your input, which may be relevant to many…

    I have about 1100 names in my family tree so far, back to the 1600s in the S of England and now spread fairly far and wide. Except for short daughter lines of 1 generation, it’s all one surname and an unusual one at that (see below). However, I also have a large island of about 300 Canadians (same unusual surname) known to be from a single, known founder (with his wife of course!) in Newfoundland in the 1850s, and purportedly originally from England. He’s my brick wall, 5-6 generations ago. Given Canadian and English samples (say an older man from each), and using Y-chromosome markers, what sort of probability/significance might I expect for (i) confirming that there is a relation, (ii) refuting the idea if there’s no evidence found, and (iii) if there is, giving the most likely separation generation (assuming no removals)? I know already that it must be more than 5 generations ago, if it’s true. And how many markers might you suggest, if this is in fact the way to ask the question?

    Many thanks.

    Jo Jesty

  • yolande

    My son and his girfrien are doing a dna to find out if the child is his , who is the third party ?

  • Lane

    I had a Y-DNA-37 test and my brother’s grandson had a Y-DNA-37 test. The results are identical!

  • pgmac01

    CeCe, I recognize the value of having my DNA analyzed, but at the same time wonder if it would be beneficial to have my elderly mom’s DNA assessed as well.