In this blog, I’ll write about false DNA matches: what they are; how to find them; examples – some from the Frazer DNA Project I am working on.
What Are False Autosomal DNA Matches?
False DNA matches are those that are not Identical By Descent (IBD). Perhaps you have heard the term. It basically means that the match is not from a person that is your ancestor. That sounds like defining something by what it is already. And it kind of is. A false match is also called Identical By State (IBS) or Identical By Chance (IBC). These are two different names for pretty much the same thing. It basically means that when the computer generated your match it wasn’t from an ancestor. ISOGG has a good article on the subject.
How Can I Tell If I Have a False Match?
There are several ways. I’ll list a few. I will give examples later in the Blog. The first list is more sure fire, but involves additional testing of parents or other relatives.
FINDING FALSE MATCHES BASED ON ADDITIONAL DNA TESTING
- If a person matches you but doesn’t match your mother or father’s DNA results, that is a false match. As you got all your DNA from your parents, this has to be a false match.
- Conversely, if you match someone else but don’t match their mother or father’s results, you have a false match.
- This is similar to the above. There is a way to phase your own results if you have had one or both of your parents tested. If you do not match on the phased (that is maternal or paternal) portion of your results, then it is most likely a false match.
- The last method has to do with chromosome mapping. I have written some about this in the past. If you have mapped your DNA to one grandparent, and the match is in the same area of your chromosome, from a different grandparent, then that has to be a false match. I’ll give an example later. There are 2 ways to do this mapping. One way is to test a lot of relatives and map their results to a common ancestor. Another way is if you have 2 siblings tested in addition to yourself, it is possible to figure out from which of your 4 grandparents your DNA came from. This method has been pioneered by Kathy Johnston.
WHAT IF I DON’T HAVE PARENTS OR OTHER RELATIVES TESTED?
- Testing parents is the best way. Then it is good to test other relatives. If that is not possible, then one may look at statistics. Many of the statistics are at the ISOGG article I mentioned above
- 15 cM or greater match – these are considered to be all good matches
- less than 5 cM – very few at this level will be considered true matches. ISOGG reports that about 85% of matches at this level are false. So it’s better not to go there.
- Triangulation – this is a way to determine true matches (or IBD). I have read that any match 5 cM or greater that triangulates will likely be a true match. In my experience only the larger cM matches tend to triangulate, so for me, this is a self-fulfilling prophecy. I won’t get into the triangulation aspect much in this blog.
My False Match With Deb
Deb was one of the first false matches that I was in touch with. I had thought that perhaps we had colonial ancestry. We shared many colonial ancestors including some of the Pilgrims from Plymouth, Massachusetts. She mentioned that she had her parents tested also. This would have been helpful to find out which side we were related on. However, I matched neither her mother’s nor father’s results. So it had to be a false match. Here is how we show to match at Ancestry:
It looks legit. It even says that Deb and I have a Shared Ancestor Hint. But in this case it is a bad hint. Another clue that this might be a false match is that the match is fairly low. At Ancestry, they use a filter and the match was only 6.0 cM. Here she is on my spreadsheet.
The matches in my spreadsheet are above the thresholds for FTDNA and Gedmatch. The lower number is phased to my father’s side, so one would think that the match would be good. However, my paternal phasing is based on a test with my mother. These phasings are not 100% accurate apparently. Deb also matches with my 2 sisters. In addition, she matches my two sisters on the X Chromosome. Apparently, these are all false matches. I have also read that many female X Chromosome matches are false. I suppose these are two examples. The bottom line is that I don’t match Deb’s parents and my sisters don’t. So these cannot be real matches.
Another False Ancestry Match
I have another example that I just thought of. I have another Shared Ancestor Hint. This one is on my mother’s side. It is based on an AncestryDNA match between Kay and myself. Kay also matches my sister Heidi but not my sister Sharon. So Heidi shows this same False Shared Ancestor Hint.
This match is down to 5.4 cM at Ancestry with their fancy filtering methodology. Unfortunately, Ancestry apparently doesn’t have the technology to check that even though my mother tested with AncestryDNA, my mother doesn’t also match Kay – at least not by DNA. However, Ancestry clearly shows that Kay and my hint’s line to me is through my mother. So this is a false match. Ancestry is wrong again. However, they do have a lot more money than I do.
Frazer False Matches
I have perhaps more experience with the Frazer side of my DNA than other DNA having worked on the Frazer DNA Project for a while. There are also false matches within that project. Here are a few false matches on Chromosome 7 between my two sisters and Frazer DNA Project Members. My sisters are SH and HHM.
HERE’S HOW I FIGURED OUT HOW THESE 2 FRAZER MATCHES WERE FALSE
Jane and MFA are in the Frazer DNA Project. In fact they have great matches elsewhere and even triangulate. So why am I calling these matches false? The main reason is the Chromosome Mapping I have done. This was done by a method I have described in previous posts. Three siblings are compared (my 2 sisters and me). Crossover points are determined. Here is what my Chromosome 7 looks like.
I have a cousin on my mom’s side who tested (in green). Her match at 56-75 with my 2 sisters and me ensures that the maternal side is on the top of the 3 DNA bars. This is because at that location (56-75), there is only one color that all the siblings share (green). That means blue and purple represent DNA from my paternal side. Blue is from our Hartley grandfather and purple in this case represents my Frazer grandmother. The numbers represent matches with relatives who I have had tested that are related to two of my four grandparents. In this case, the relatives matched my mother’s father (green) and my father’s father (blue).
SHARON’S FALSE FRAZER MATCH
My sister Sharon’s DNA is represented by the first horizontal bar. She has blue Hartley DNA from the beginning to point 129,000,000 (or 129 for short). At that point from 129 to the right end, the DNA from Sharon’s Frazer grandmother takes over. 129 is the crossover between where she gets her Hartley DNA to where she gets her Frazer DNA on Chromosome 7.
I have that Sharon matched Jane from the Frazer DNA Project from 98 to 107 for 7.6 cM. However, this cannot be a Frazer match as Sharon got all her DNA from the beginning of her paternal side to point 129 from her Hartley (non-Frazer) grandfather.
HEIDI’S FALSE FRAZER MATCH
Likewise, my sister’s match with MFA of the Frazer DNA Project is also false. Her bottom bar is all blue which means she has all [non-Frazer] Hartley DNA. There is no room for her to match MFA from the Frazer DNA Project from 130 to 135. In fact, Heidi has a match with her reference Hartley relative from 134 to 139. What the map shows above is that you cannot get DNA from 2 different paternal (or maternal) grandparents at the same location. It has to be either one or the other.
Interestingly, these false matches happened in the places where they could not have happened. If they were to have been real matches, they could’ve happened with me (Joel) as I have more purple area on my bar above. Or MFA could’ve had a true match with Sharon where she had some purple room, rather than with Heidi – which is a false match.
So Are False Matches Good Or Bad?
They are neither good nor bad. However, if you have a match that is false and you think it is true, then that could be bad. That would lead to wrong conclusions. Notice that in the above example, both the matches were just above the Gedmatch 7.0 cM cutoff. Just because a match is above the cutoff, doesn’t mean it is a real match. That level was chosen because there are probably more true matches than false matches at that level, but there are still a lot of false matches around 7 cM. Gedmatch and testing companies don’t generally like to filter out matches that could be true.
Summary
- It is good to be aware of (and wary of) false matches
- Just because a match is above a threshold doesn’t mean that it is a true match
- Matches below a threshold could be true also, but the odds are against that
- False matches do not triangulate
- False matches do not match either of your parents’ DNA
- Neither do they match either of your matches’ parents’ DNA
- False matches may match a phased kit of your own DNA as phasing a parent based on another parent’s testing is not 100% accurate
- If a match doesn’t match your paternally or maternally phased kit, it can be considered false
- A low match level means high likelihood of false matches; a high match level means a high likelihood of true matches
- At about 15 cM there should be no false matches
- Don’t blindly accept AncestryDNA Shared Ancestor Hints.
Well, I don’t understand all that you said in the article because I’m not that advanced on DNA studies. However, I do have a question/comment for you that I’d love for you to address.
Sometimes in families secret things happen that maybe only 1 or 2 people know about.
For example: there is a woman who has 3 children. 2 boys & 1 girl. All of them believe they have the same mother and father. Why would they think otherwise?
However, in truth they do not all have the same father. The mother had a secret affair one time with the last baby. And it’s father was somebody different. But she let her husband think that this was his child. So nobody knows & she’s never going to tell. That girl is going to grow up and maybe get her DNA done some day & have a big surprise.
Secret adoptions also mess things up. Most adopted people are told when they are young that they are adopted. But not everyone is. Sometimes you get parents that never want the child to know – not even when they are an adult. If they did DNA they’d get a big shock…and have lots of questions.
Then you have an instance where a couple has 4 kids all about 2 years apart. Then late in life they have a 5th child. A girl. Everybody just thinks they are lucky to have had a child at their age & don’t really think anything about it. However, the truth is that this baby really came from one of their daughters. She was pregnant out of wedlock at the age of 17, & her parents took the baby and raised it as their own & the real mother & the grandparents never told a soul. Not even the child when it grew up. She lived her entire life never knowing that her older sister was really her mother.
These kinds of things happen all the time in life. There are all sorts of reasons why your DNA might not match your parents or siblings or other relatives.
This is why nobody should ever just go by surnames alone when doing genealogy. Nor can you always ‘assume’ your DNA is going to match your other siblings. Or that your DNA will match that of an Uncle or Great grandparent.
You can go back many generations & somewhere in that line (possibly 100’s of years back) there is going to be some instance of when some of the things I’ve mentioned above have occurred.
The problem arises when we, today, do not know of these things from years ago. It can mess us up whether we’re doing regular genealogy or DNA matching.
At least that’s what I’m guessing. How would you account for situations like this? What if a person thinks they have one of these ‘false’ matches – but it’s actually due because there is some unknown information about paternity that is missing?
Hi Christie,
One good way to solve paternity issues is through YDNA testing. This is accurate going back hundreds of years. However, it only works for the father to son to son, etc. line. For fairly close relations, autosomal testing is accurate in showing if the relations are what they are supposed to be.
Joel
I agree with you on this about the paternity issue. Families didn’t talk about that back then.
I don’t understand how to figure out the DNA items, nor do I pretend to, I enjoy seeing the match ups bases soley upon it.
Some of the names in that show up on my DNA are right on, known relationships, and others are surnames I have never heard of.
Perhaps this is because of a relationship of not knowing that they had a different father or mother or were adopted.
I love the searching genealogy end of matching them up.
Thank you so much for clearing up some of the things I had about DNA matching. You explain things so clearly and to the point!
Are there any books that you would recommend that help explain DNA matching?
Again, thank you!
Hi Jenni,
Most of the good DNA information is on the internet. The Segmentology website has been helpful to me. Roberta Estes and others have instructional blogs. Roberta has many blogs, but has organized them by category.
Joel
Well that explained something for me. I have had my grandmother, my son and myself tested through ancestry, one of the results came back as related to my grandma and my son but not me…. Makes sense now…
Yes, some things seem obvious when pointed out or realized. I was thinking I was related to Kay by DNA. Kay shows an Ancestry Tree related to my mom. But my mom doesn’t match Kay by DNA. That shows something is wrong.
Joel,
I find what you have to say very interesting. In my case, all of my direct ancestors, i.e. parents, grandparents, Great Grandparents & their siblings are all deceased, and never had their DNA tested. I do have a sibling, but she has not expressed interest in doing her DNA. (Might be something she might do.) We do have many cousins but having or doing their DNA has not come up because we are not that close, in fact some I have not seen since since grade school over 60 years ago.
Willy
I asked my 2nd cousin to test and he was glad to do it. I’ve probably seen him only a few times in my life. I also asked a cousin I found via genealogy and she tested for me. The same was true with my father’s first cousin.
So I am trying to get this right. I have my Mum tested and have phased maternal and paternal kits at gedmatch. I understand that true matches will have both Mum and I matching to be maternal, plus her phased kit based on my results. For my paternal side however I’m not quite sure. If the phased paternal kit and I match its all good clearly. But if me and my son match and Mum doesn’t, but there is no match on a phased paternal kit, even though it is a good segment size, I would have though it was IBD but are you saying they both could be false matches?
Hi Veronica,
Thanks for writing. My understanding of a phased kit, is that the computer is sorting out the DNA you got from you mom from the DNA that you got from your dad and making 2 kits out of one. If you have a match with your maternal phased kit, it should be a maternal match. If you have a match with your paternal phased kit, that match will be on your paternal side. You can check also if the match is with your mom on the maternal side match also, but it is probably an unnecessary step. I believe that you have the second part right. If you and your son match, but you didn’t get that DNA from either of your parents, something went wrong somewhere as you had to have gotten your DNA from your parents. Remember that you and your son both have maternal and paternal matches. You may be showing a match with him that isn’t really a match. What I mean is that you may be ‘matching’ your son in the same area of the chromosome where he has DNA from his father. If you had phased kits for your son, it may clear things up. Then you would know which side he is matching on also. It is a bit confusing!
Hi Joel – I’ve enjoyed reading many of your blogs and this one in particular. They are very helpful to me, I appreciate it.
I’m still at the beginning stages of Chromosome Mapping using Kathy Johnston’s method. When you did the graphical match between two people – did you set a minimum match somewhere 10cM or 15cM? I am wondering because your matches have far less segments than I have – I can simplify mine down to 13 crossover points which seems like a lot.
For my mother’s family, I have her and 2 siblings tested. On their paternal side I have an aunt and two first cousins tested. On their maternal side, I have an aunt, a first cousin and a first cousin 1R tested. The two aunts – even though on paper they are not related for at least 100 yrs before their birth, no idea where they could be related – but they show up genetically related as third cousins. Nearly all of their ancestry from 1800’s came from the same remote islands off Scotland: Barra, S. Uist and Benbecula. How does Endogamy affect Kathy Johnston’s chromosome mapping method?
Hi Kathy,
Thanks for the nice comment. I was using the gedmatch defaults for the Kathy Johnston Chromosome Mapping. However, based on her advice, I went to a lower threshold to get some of the more subtle crossovers. I go down as far as 3 cM. That way, I can decide whether I want to use all the small matches or not – depending on whether they appear to make sense.
To tell you the truth, I don’t know how endogamy would effect the mapping. I would guess that you could run the utility at Gedmatch for ‘are my parents related?’ That would tell you the areas to watch out for when mapping. If you run the utility and don’t get that your parents are related, I wouldn’t worry about it.
Thanks for the tip. I hadn’t thought of phasing my son but thats a great idea. Now I have to find the matches where we both matched and sort them out again!
Hi Veronica,
I’m glad my comment helped. Sometimes I’ll get a comment from someone or read something that will start of spark of understanding in my head.
How often are larger matches incorrect? I supposedly share 350+ centimorgans with a big family from another part of the country. I was trying to be open minded to the possibility that there was controversial event to account for this, but not even that is making sense. Turns out we’re mostly from different countries, and no paths cross on the one plausible connection 5 gens back, which is distant for this type of match, right?
I’d like to know if genetic genealogists do consults on small, specific cases. Any thoughts?
Thanks, all-
Normally, if there is no endogamy (marriage of cousins), a match of 15 cM or more should be a ‘real’ match. Certainly 350 cM would be a cousin. A lot of work has been done on trying to figure out for the number of cM what relationship that might be. 350 cM would be in the higher range of a 2nd cousin or the lower range of a 1st cousin once removed.
Genetic genealogists do consults.
Joel
Hi Joel,
I have a strange one – after having my 23andMe kit done, I uploaded the data to MyHeritage. My results showed a 1st cousin twice removed (4th cousin) that if true, would be extremely unexpected. His family history is well documented and from a different country than any of my ancestors. To follow up on this, I ordered kits for my mother and my paternal uncle (dad passed years ago). This mystery relative does not show up in their results – only mine. And we are positive that dad and his brother are legitimate.
I would write the whole thing off as a false match, except for your suggestion that anything over 15cM should be a true match.
MyHeritage gives the following match quality: 0.6% shared DNA (46.7cM), 5 shared segments, largest segment 17.9cM.
I’ve been wrestling with the idea of contacting him – if we’re indeed related it would be a heck of a story for my side, not so much for his. Given your knowledge on this subject, do think this sounds like a false match?
Hi Wendy,
I would upload your results to Gedmatch.com to double check. I’m not so familiar with how accurate the MyHeritage results are. Perhaps this match has uploaded to Gedmatch. Or if this person has tested at FTDNA, it may be possible to transfer your results there for free. Also a 1st cousin twice removed would be similar to a 2nd cousin, so something doesn’t see right with what you mentioned about 4th cousin.
Joel
I have not found MyHeritage to be very accurate. A known relative with significant common cMs have matched with my sister, uncle and me, but not with my mother.
The strangest case on MyHeritage has my father matched to his cousin’s husband. She is a verified cousin through our known shared lineage fortified by tests with 23andMe, which have been uploaded to gedmatch.com. Further, she is of primarily African extraction from Jamaica and her husband it Italian!
It seems to me that MyHeritage is still very much in the experimental stage at the moment and will take a minute to get it right.
I have been told Ged match has been wrong before but I dont know how true this is. I am not finding anyone on my fathers side. I was told to run are my parents related tool and I did and it comes up that they are. Then talking to people I was told my mom got around and had sex with her brothers and her father was so so mean to her. I don’t know what to think any more. Friends tell me to get on to them for things. What do I believe I don’t know where to go from here I am so sad and angry at the same time. I feel bad for my mom but they all lied to me well I think? I mean friends and family members of my birth mom said she was prostituted out and things so it has to be true right?
I thought I would share an incident I had with a false match on GEDMatch in May 2018. I uploaded the raw data for both my Mom and myself from FTDNA to GEDMatch back in January, and was surprised to see an interesting result this morning.
As background, when I loaded our raw data in January and looked at the one-to-many match results, there were no surprises. My mom was at the top of my matches list with over 3,500 cM in common, and I was the top match in her results. Moreover, we both matched with the same group of people we matched on FTDNA’s Family Finder. Nothing unusual.
I was eventually contacted by a gentleman who wanted to know what my relationship was to his New Zealand relatives. I do not have any relatives in New Zealand, but decided to take a look on GEDMatch this morning.
Lo and behold, there was an entire group of New Zealanders in my one-to-many match results. Most curiously, one lady matched me at over 3,500 cM total and estimated a 1.0 generation gap. In fact, she matched me the exact same total cM, longest cM block, AND total/longest X-chromosome block.
So what scenario does this fit? Does my mother have an identical twin? What else could it possibly mean?
When looking at the one-to-many results through my Mom’s DNA, I found none of the same matches. So were the New Zealanders relatives on my Dad’s side? Well, considering I didn’t get my X chromosome from my Dad, I don’t see how that’s a possibility, baring some amazingly coincidental genetics.
Then I tried a one-to-one comparison between my DNA and the New Zealander, and also between my Mother’s account and the NZ’s. GEDMatch showed in both instances that we both share no blocks over 7 cM with the NZ lady.
So I’m left scratching my head – but it seems clear that there’s an issue with the GEDMatch algorithms; I’m currently awaiting a response from GEDMatch to troubleshoot the case. There is a polar opposite difference between the one-to-many and the one-to-one results.
Just a word to the wise – SCRUTINIZE your DNA matches. Make sure you know what you’re looking at before interpreting the data. If something seems fishy, it just may be.
Note: I meant to say that the New Zealand lady “…matched me the exact same total cM, longest cM block, AND total/longest X-chromosome block” AS MY MOTHER. Sorry for the confusion!