Combining Tools – Autosomal Plus Y-DNA, mtDNA and the X Chromosome

Sometimes, there’s nothing worse than a little bit of knowledge to get us into trouble.  If you need proof of that, I can show you a picture of one of my first quilts which has thankfully disappeared someplace and was known semi-affectionately as “The Ugly Quilt.”  I even entered it in an “Ugly Quilt” contest and it wasn’t even good enough, or is that bad enough, to win that!!  Fortunately, things have improved!  I’ve learned a lot.

Combine a little knowledge with people who desperately want answers, and you have a situation ripe for mistakes, misinterpretation and misunderstanding.

That’s what sometimes happens when you combine the results of two different genetic genealogy tools and you don’t really understand their differences, their application to the specific problem at hand, or what the results are really telling you.

I’m talking about combining autosomal testing with haplogroup based testing, both Y-line and mitochondrial DNA.  This comes in two flavors; generic and specific.

Generic Matching – 23andMe

At 23andMe, your match results are displayed in a list along with information which may or may not be relevant to you and your match.  Shown below are my 8 top matches at 23andMe.  I know who these people are – they are my relatives, so there is no question of interpretation here.  Let’s take a look at the information provided.

combining 23andme

I have omitted the name column which is first.  The second column is their relationship to me.  The top row is me.  Everyone has the option to enter geographic (blue tab) and surname information (green tabs,) which I have done.  Not everyone does that as you can see by the information shown for the others.

Note the different haplogroups here.   For mitochondrial (pink tab), you have 7 different haplogroups out of 8.  That’s because these people, other than my son and I, don’t share a common maternal line.  If they did share a haplogroup, it would be coincidence, or very far back in time, because we know the pedigree charts of all of these people and they do not share a known maternal ancestor.

Looking at the Y-DNA haplogroups, you’ll notice that there are 4 men and of those 4, three share the same haplogroup.  That is because, in this case, they are cousins who also share the same surname.  If I was an adoptee and made this discovery, I’d be in 7th Heaven, because this would be a very large hint.  However, if these men shared a haplogroup but didn’t share a common surname, again, it could be coincidence or a common ancestor very far back in time.

I put those words in bold because recently I’ve seen the tendency to jump to conclusions about the relevance of common haplogroup information related to autosomal testing.

Let’s use an example. At 23andMe, you are provided with what is considered an extended haplogroup.  Most of the time, these are correct except when the haplogroup designation involves insertions and deletions or reversions which can’t be detected reliably by this type of testing, only by full sequence or SNP testing.  Let’s not go there and let’s presume these are absolutely accurate for purposes of this illustration.  I happen to know my haplogroup listed at 23andMe is out of date.  It is listed as J1c2 and it is actually J1c2f, but that actually enhances the point I’m about to make.

Using the Behar paper supplement to “A Copernican Reassessment of the Human Mitochondrial Tree From its Root,”  the common ancestor for haplogroup J1c2 lived approximately 9700 years ago (plus or minus 2010 years standard deviation).  Therefore, my common ancestor with anyone sharing this haplogroup is anyplace from the current generation (my children or parents) to nearly 10,000 years ago – clearly not relevant for genealogy.  However, looking at my extended haplogroup, not determined by 23andMe, but found in my Family Tree DNA full sequence information, the common ancestor of J1c2f lived about 1900 years ago (plus or minus 3100 years standard deviation).  Clearly that makes about an 8000 year difference, which narrows the window, but it still isn’t necessarily genealogically relevant.

Furthermore, at 23andMe, haplogroup information is provided, but personal mutations are not, for either Y-DNA or mitochondrial.  This is why I referred to this type of match at “generic.”  For specific Y-DNA or mitochondrial matching, you’ll need to go to Family Tree DNA.

Specific Matching – Family Tree DNA

At Family Tree DNA,  Y-DNA, mitochondrial DNA and autosomal results require different tests.  The results are shown on different tabs on your personal page.

combining ftdna

Each tab provides you with a significant number of pages of information about each test and displays your results in different ways.

For both Y-DNA and mitochondrial (mtDNA), one of the options is “Matches” which shows you your personal matches at several levels.  For mtDNA, the levels are HVR1, HVR1+HVR2 and Coding Region, which equate to the three levels of tests that you can take – basically introductory, intermediate and advanced.  For Y-DNA, the levels are 12, 25, 37, 67 and 111 markers.

My match results are shown below, again, with the first column, names, removed.

combining mt matches

SmartMatching is important here, because Family Tree DNA has already done you the favor of removing anyone who is not a “true match.”  Notice that the first column shown here includes the envelope icon, a notes icon, a pedigree chart icon, and following that, the level of testing taken by this person.  I’m showing my full sequence matches here, so everyone has taken the FMS or full mitochondrial sequence test.

These are the people who also share the extended haplogroup of J1c2f.  This means our common ancestor lived sometime between now and about 2000 years ago (plus or minus the standard deviation.)  When you look at the oldest ancestors and the matches map that goes along with this test at Family Tree DNA, you can see how widely spread these “most distant” ancestors are.  You can also see that one person has listed their grandfather, which means they were confused.  A most distant mitochondrial, maternal, ancestor cannot be a grandfather – so this also calls into question the accuracy of their geographic information as well, shown in the Czech Republic, below.

Combining map

Two thousand years ago (give or take) the common ancestor of all of these people was one person, and their direct descendants, their children, all lived in the same place initially.  You can travel a long way in 2000 years.  My oldest ancestor, the white balloon is found in German and my closest match is found in Norway.

To understand how to use combined tools, you have to understand each individual tool first.

Family Tree DNA does provide a combined matching tool called “Advanced Matching” for Y-DNA, mtDNA and autosomal (Family Finder) tests.

Advanced Matching

Advanced matching allows you to combine test types and filter on specific fields.

combining advanced options

The most common advanced matching for autosomal DNA is the combination of the Family Finder test plus either mtDNA or Y-DNA results.

As they say, “your mileage may vary” and much of this variance will depend on two things.  First, how many people tested at which testing level of the mtDNA and Y-DNA tests and second, the relative rareness of your haplogroup.  Said another way, if your mtDNA haplogroup is H and/or if your Y-DNA haplogroup is R, you’re very likely to have a lot, many, low level matches because those haplogroups make up about half of the European population, respectively.  However, if your haplogroup is J1c2f, meaning that your base haplogroup is much less common than H and that you’ve taken the full sequence test, you’re going to get a lot fewer and a lot more meaningful matches.

At the haplogroup H level, which is the most common HVR1 results, your common ancestor lived between 12,000 and 30,000 years ago, depending on whose estimates you use.  Compare that to J1c2f’s 1900 years.  Big difference.  But is it big enough?  It’s a clue, just like any other clue.

What Matches Don’t Mean

Let’s say that on the advanced menu you selected two tests, the Family Finder and the FMS (full mitochondrial sequence) test.  The result is no matches.  IF you had a match at this level, it does NOT mean that your common autosomal match is on the maternal, mitochondrial line.  This is a very common mistake in logic.  It means that you should continue to include this line in your search and maybe you want to focus there.

Let’s look at why.  Autosomal testing reaches back in time to recent ancestors and measures how much of their DNA you share.  In the past 5 or 6 generations, you likely share some DNA from all of your ancestors.  After that, some of your ancestors DNA gets so diluted that it becomes in effect, washed out, or is present in such small quantities that we can’t effectively attribute it’s source.  Mitochondrial DNA however, is never admixed or divided.  Therefore time in terms of recent generations, unless we’re talking about when mutations occurred, like the mutation that set apart haplogroup J1c2f some 2000 years ago, is irrelevant.  Mitochondrial and Y-line DNA both measure back in time to your earliest ancestor in that line.

The best use of both mtDNA and Y-DNA with autosomal is to eliminate possible lines.

What Matches Do Mean

Let’s say I select Family Finder and the HVR1 level and show only people I match in both tests.

At this point, especially if you are haplogroup H, you’re going to get a long list of matches and people get very excited at this point. Don’t.

combining mt no match

Above is an example list.  Here’s also the problem.

Problem 1 – Most people only tested at the HVR1 level.  For haplogroup J, this means the common ancestor lived about 35,000 years ago, plus or minus 5,000.  What this really means is that if these people were to take the full sequence test, chances are they would no longer match you.  There are more than 100 subgroups of haplogroup J and chances are very good that the tester would fall into one of them.

Problem 2 – Some people have tested at the HVR2 level or the FMS level and don’t match you at that level, even though they matched you at the HVR1 level.  Look at the first result, the second column, the X.  This means they did test and they don’t match you.  This means that you’ve just eliminated this direct maternal line as a possible autosomal match, barring a mutation in the past few generations which is not impossible but extremely unlikely.

However, when people are desperate for any shred of evidence, they interpret this as “I match on the HVR1 level so this must be my common line with this person.”  That is flawed logic and is outright wrong in the situation where the person has tested at a higher level and does NOT match.  In fact, it’s just the opposite, you’ve just disproven this line.  Now I think this is a good thing, because that means you can focus elsewhere.

This same logic holds for Y-DNA matching as well.  Finding someone you match with at the 12 marker level in haplogroup R, especially R1b1a2 (M269) is quite common.  Finding someone you match at 67 or 111 markers and autosomally might be quite another matter.

A Third, Neglected Tool

There is a third tool that can be added to the mix here, but it’s not nearly as convenient as Advanced Matching.

Both 23andMe and Family Tree DNA test your X chromosome when they do their autosomal testing.  Family Tree DNA does not report these matches.  23andMe does show them on their Family Traits, Ancestry Composition and Family Inheritance: Advanced results.  Both companies include the X information in their raw data downloads (as does Ancestry), which you can then upload to www.GedMatch.com.  GedMatch is the great equalizer here, although they are temporarily (as of July 2013) not accepting any new uploads for the next month or so.

The X chromosome has a unique inheritance path which is different for men and women.  If you recall, women inherit an X from both Mom and Dad, but males only inherit an X from Mom.  They get the Y from Dad which makes them male.  If you match someone on the X chromosome, or you don’t, that too is powerful information.

Blaine Bettinger originally published some wonderful X inheritance charts on his blog, The Genetic Genealogist, in December 2008 and January 2009 documenting how to use the X chromosome for genealogy.

The chart below shows the male inheritance path for the X chromosome via the colored locations.  Because males and females both inherit the X from their mother, the maternal inheritance path of the X chromosome, the right half of this chart, is the same for men and women.  In this case, we’re particularly interested in the mitochondrial DNA path as well, which is the furthest right pink line on the chart, shown with the arrows along the edge.

combined x match

Including the X chromosome matching, here are your three possible outcomes.

  • If you match autosomally, you match at the deepest (full sequence) haplogroup level and you match on the X chromosome, you may indeed have a solid lead in the direct maternal line.  It’s a lead, nothing more.  It’s not confirmation of a common autosomal ancestor in that line.
  • If you match autosomally, you do not match at the haplogroup level, but you do match on the X chromosome, then you know it’s NOT the direct maternal line but it IS one of the other lines where you share an X chromosome.
  • If you match autosomally and you do not match at either the haplogroup level or on the X chromosome, you know that you can eliminate the direct maternal line and your match is probably on a line where you don’t share the X.  I say probably because like any other DNA that is shared in an autosomal fashion, meaning divided by approximately 50% in every generation, it’s possible after several generations to not show as a match on the X but to still be descended from those lines.

Jim Turner created some nice X chromosome inheritance pedigree charts that are easily printable which you can find here.

Take Away

What’s the take-away in all of this?  These are very powerful tools, but they only tools and they provide clues.  Some clues eliminate possible connections, some clues suggest them.  It’s only through multiple tools like triangulation and old-fashioned genealogy research that we confirm them.

We’ve gotten spoiled with the relatively easy Y-DNA answers.  A man tests and if he matches other men with the same surname with few mutations, we call it family and all is good.  Women don’t have that luxury and neither do adoptees, although male adoptees clearly have the advantage of a potential solid Y match.  Other types of DNA testing and analysis just aren’t as straightforward or easy, but that doesn’t mean the answer isn’t there.  Perseverance is key.  Common sense, understanding the tools and removing emotion, as much as possible, from the equation are critical.  If you’re in doubt, get help.  It’s a lot better to pay for an hour or two of consulting than to make a critical error in logic that can introduce errors into your family tree or cause you to waste time chasing the wrong lines.

Unraveling the secrets your DNA has to tell you is much like that game of Clue that we played as kids – accumulating pieces of information that, cumulatively, hopefully, lead to an answer.  Miss Scarlet did it in the ballroom with Professor Plum.  Or was it Colonel Mustard, or Reverend Green?

combined clue

33 thoughts on “Combining Tools – Autosomal Plus Y-DNA, mtDNA and the X Chromosome

  1. Roberta, if I am understanding this correctly, I should not depend upon my brother’s FTDNA to check for MtDna. I should also be tested. Would it be best to use FTDNA, 23 and ME, or should I do both? There are so many matches I do need to narrow down the search. Most of my genealogy is back to the 1700s.

    • You and your brother, assuming you have the same mother, should share exact mtDNA results, so yes, you can use his. It’s the X and autosomal where you won’t necessarily share the same DNA. If you’re going to test for mtdna, the full sequence at Family Tree DNA is the way to go because you get your personal mutations and matches on top of your full haplogroup designation.

      • My brother’s DNA has been tested to the full mtDNA and to 111 on the male side. It seems like I need the autosomal for myself. I taught myself how to research genealogy, but this is much harder . Thank you so much for your blogs!

      • Yes, you need autosomal for yourself, but autosomal for your brother will help too. You will each match some different people because you inherited from your parents differently.

  2. Loved this blog. I can relate to this because when I first tested with ancestry.com, what a huge mistake, I was matched with everyone. Then when I tested with Family Tree I received a lot of matches that did not make sense so I tested with Nat Geo 2.0 to confirm my mtDNA and it still was off. It was when took FSM and then emailed my results to Mr. William Hurst that the haplogroup was revealed. Mr. Hurst and his friend ran my mtDNA through a program that his friend developed and they found that I am K1e. So, now the tool will help me find the right path. I spent more money but the clues make more sense and it was worth every penny. Thanks, Roberta for all your hard work. It helps us to navigate that road map of genetic genealogy.

  3. In your first screenshot, you can click where it says something like “2nd to 3rd cousin” and then enter the correct relationship. I assume you know the correct relationship since you said that you know their pedigree charts.

  4. Since my FTDNA test was already uploaded to GEDMATCH, I went there and brought up a long list of matches in a chart titled “DNA Matches to Kit # [my number]. It says to click on the “X” to get a comparison, but every one that I tried had the same result “No shared X-DNA segments found.” One of these is my maternal uncle; others are known autosomal matches. There are zeroes in all columns under X-DNA. Is something not working in the Gedmatch program OR is it possible that I don’t share an X-chromosome with any of my autosomal matches?? Also, the kit numbers all start with a letter, ie. F###### or A###### or FN##### or M######. Are these a code to the company which did the testing??

    • F is Family Tree DNA, M is 23andMe, A is Ancestry and I don’t know what FN stands for. Given what you’ve just said, I suspect GedMatch is not working correctly. You would definitely share large and multiple X segments with a maternal Uncle.

      • Janet as I recall I had to upload the X results from family tree DNA separately from the other chromosomes. It sounds like yours are not there yet. When uploads are re-enabled follow the ftDNA instructions for uploading your X

    • Janet as I recall I had to upload the X results from family tree DNA separately from the other chromosomes. It sounds like yours are not there yet. When uploads are re-enabled follow the ftDNA instructions for uploading your X

      A is for ancestry, F for family treeDnA and M for 23andme

      • Thanks for the suggestion. Gedmatch isn’t doing any uploading again until Aug. 15, but I’ll try it again after that date.

  5. Following the Y-DNA. If you no longer have a father and cannot test him anymore can I use a distant cousins DNA that is on Family Tree to trace my ancestor if I know that his DNA is a conclusive match?

  6. Great article. I suspect some of the less than bright questions I have posted on the DNA Newbie list lately might have prompted it! If not, it was just serendipity and I thank you immensely, Roberta.

  7. In the article you say that your FMS matches give a time to your most recent common ancestor of 2000 years (plus or minus). I was under the impression that a FMS match gave much shorter time to a common ancestor. The ftDNA website suggests (faq=2140) a time of 125 to 550 years the common ancestor with a perfect match. Perhaps the 2000 year figure has some degree of mismatch?

    • My point was that there is no way to tell. The outer border is 2000 years based on the haplogroup age, which is the 2000 years, plus or minus. There is a great deal of disagreement about the mutation rate in mitochondrial DNA. So a perfect match could be within that range, but we know it’s within the 2000 year bracket because of the haplogroup age. Very few people can go back even 550 years on mtdna. That would be about 1463, before the discovery of the New World by Europeans.

  8. I have a general question and hoping you might do a post sometime soon covering this.

    I have FTDNA and GEDMATCH accounts for my family AT kits. I have found a nice match which breaks down a brickwall. except it is possible we match somewhere else in the tree. I would like to find other people that match us in the same location to see if they connect to my brickwall family or to a different family. How do I do this? I don’t find any tools that let me enter specific chromosome and segment data and look for matches. Am I missing something or is this not possible with our current set of tools? Do I have to do it the hard way, looking through my 250 odd matches one by one?

      • I explained I have my results at GEDMATCH.
        Where is this feature found there?
        “I don’t find any tools that let me enter specific chromosome and segment data and look for matches.”

        The DNA tools option sound interesting for other reasons, but it seems much quicker to find what I need using FTDNA Chromosome browser to turn on 4 matches at a time. If there is nothing in the general area on the specific chromosome I am interested in than move on to the next four. I found it does go pretty quick.

      • “Find people who match with you on a specified segment” is the tool at GEDmatch

  9. Pingback: Friday Finds – 07/19/13

  10. I am LOST!

    I went to the DNA Tools site and registered, then to the FTDNA tab and “Get Data” which puts a compressed file on my machine. I uncompress it, but it only contains data for my kit. There are other options on this FDNA page but they are grey and never become active.

    Download Raw ICW Spreadsheet
    Download Raw Chromosome Browser
    Download Family Finder Matches

    The last one appears to be what I want, but how do I make it available?

    • I also went to the DNA Tools site and registered for an account. When I clicked to open the .zip file, there were three files there. One file name is the Raw chromosome matches from FTDNA, and I assume the other is a ICW spreadsheet and Family Finder Matches. If you got three files, you should have what you’re looking for.

  11. I have 2 cousins that match me in the same segments on Chromosome 3 in FTDNA. When I do matches in common with Cousin A, Cousin B does not show up and vice versa. How can I match/share this segment and the cousins not be related to each other?

    • They may both match you, but not match each other on a large enough segment to show up as a match. You can compare yourself to both and see if the segments where they match you are exactly the same. If they are offset, that’s likely why.

  12. Thanks for the very informative article Roberta. Just wanted to mention, it looks like there’s a typo when you said “about 1900 years ago (plus or minus 3100 years standard deviation)”.

    • Hi Bart. I wish it was a typo, but it isn’t. The mitochondrial DNA dating is taken from the paper titled A ‘‘Copernican’’ Reassessment of the Human Mitochondrial DNA Tree from its Root by Doron Behar et al published in 2012. It is the most recent tool. Unfortunately, sometimes the standard deviation is higher than the age which of course, on the low, end would make the haplogroup not yet born, which can’t happen. So, just chalk it up to being a “math thing” and knowing that on the low end, it’s not possible, but it is on the higher end of the scale. Good eye!!!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s