Sometimes DNA can reveal an ugly truth. A reader, let’s call her Patty, asked me for help explaining the unusual looking comparison of her DNA test results with her uncle Bob’s results. The surprising thing was the large number of green bars that she saw in the GEDmatch one to one comparison, indicating fully identical segments (FIRs), almost as many as a full sibling would have. How could that be?
Of course my first thought was that Bob is actually her full brother, that her mom, Janet, had a child out of wedlock who was raised by the child’s grandparents, Mona and Dick, as their son. This has happened in many a family. But that was not the backstory. Janet was a small child when her brother Bob was born. Bob and Patty also share just one segment of 27 cM on the X chromosome, which, of course, would be normal for a maternal uncle but low for a brother. Have a look at the comparison image from GEDmatch for chromosomes 1-22:
Thoughts? Usually only full siblings or double first cousins will have numerous fully identical segments, so what could this be? Obviously Patty’s dad would also have to be a close relative of Bob’s for there to be so many FIRs. A full sibling would usually show even more of them, however.
Here is what Patty told me:
“I am trying to determine if my mother’s father is her biological dad (which we have been told he isn’t). Unfortunately my mother has passed away. My mum’s brother just did a DNA test to help me. Here’s the thing, I have now found out that possibly my father is not my biological father and that my Mum’s stepfather could be. How many cMs could I expect to see from my uncle who could be my half brother as well ?”
Oh my! Her uncle Bob is likely both a maternal half uncle and a paternal half brother! This would certainly explain the DNA picture that we see above and those FIRs. Patty’s mother Janet is only the half sibling of Bob (different dads) while Patty shares her biological father Dick with Bob … Part of me is hoping that Dick knew he was not the biological dad when he had relations with his daughter. Needless to say this is a very difficult situation and I really feel for Patty.
I checked Patty’s kit at GEDmatch and her parents were not related. That confirms the fact that Patty’s father was not the biological dad of her mother Janet.
A half brother is expected to share about 1700cM (range: 1300-2100*) and a half uncle would be about 850cM (range:540-1172*) thus a total of about 2550cM for this relationship. However some of that will be fully identical which does not show in the totals. Also due to the randomness of DNA recombination, the amount of shared DNA can vary widely. Below are the numbers: they share 1,938 cM which works as a sum of the low ends of both ranges. Notice that there are 6 segments larger than 100 cM which is indicative of a very close family relationship.
I informed Patty that the DNA results were consistent with the expected two half relationships. Since this was such an unusual case I wanted to share these results so we could all learn from them. She gave me permission to publish this as long as no real names were mentioned. So here are the results.
|Chr||Start Location||End Location||Centimorgans (cM)||SNPs|
Largest segment = 156.4 cM
Total of segments > 7 cM = 1,938.4 cM
37 matching segments
Estimated number of generations to MRCA = 1.4
* cM ranges for close relatives from the work of Blaine Bettinger as listed on the ISOGG wiki
A small box chart and some (made up) given names would make this easier to follow.
Thanks Israel. Not sure I could even come up with one but I will think on that!
I was very interested in the unusual relationship aspect… I have an interesting situation as well – possibly a 3/4 male sibling. I am VERY new to this. and am so blessed to have found a brother – and he is as tickled as I am! We still have massive digging to do, but it is very convoluted.
I will email you so I can have a look. The most common 3/4 relationships, obviously, are children from a mother and 2 brothers or a dad with two sisters …
Hi! I have a situation where my dads brother may be my sisters father. I uploaded my results to GEDMatch. Can you help decipher it?
Savannah – we were in contact before and I looked at your results, you have the same father, your dad’s brother is not her father. Or is this a different sister?
But I loved the suggestion of using fake names to make it easier to follow so I just made that change, thanks!
Did you post the family tree chart with pseudonyms yet or did I misunderstand what you were doing?
One of the great things about this hobby is that nearly everything can be represented graphically. I have trouble loading sequences of text in my brain, but charts and graphics almost always lead to an instant recognition of what is happening. Or at least I feel like the information is under control and I can move on.
With every adoption puzzle I have worked, even the simplest to solve are always “complicated” in some way. Even when the reunion is welcoming and joyful, there has always been a something in the family we don’t want to talk about. Wait…. Actually, that might be true of everyone! Nevermind.
Thanks for the wonderful work. It makes a difference.
I’m quite new to family genetics learning as much as possible by reading helpful articles like this, thank you. The bar chart shown is very pretty but is it correct to assume with this kind of chat that the solid blocks of colour, yellow and blue here, are the shared DNA? I’ve seen of GEDMatch a resource showing comparisons with ancient DNA found in various parts of the world. The result is presented in a similar bar chart but until now I was not sure how to read it.
The closest post of mine that I found that explains some of this is http://blog.kittycooper.com/2014/10/using-gedmatch-for-my-ancestry-com-cousins
The bottom blue chunks show where on the chromosomes two people match. Since you have two chromosomes at every position, the top line shows green only when both chromosomes match. Red is not matching
If you do a one to one comparison at gedmatch there is an explanation at the top of the page for each color
So Janet’s mother had two relationships, one with Janet’s father and one with Bob’s father. Bob’s father then had a relationship with his stepdaughter Janet and Patty was the result?
Emily, not quite.
Patty is the daughter of Janet and Janet’s step dad while Bob is the son of Janet’s mom and Janet’s step dad
This sounds like one of those written math questions…guess in a way it is.
Is there a good resource for finding the cM length of a segment based on the start location and end location of the segment? It would be nice to be able to calculate the length of the FIRs.
23andme gives these lengths as does the Genesis beta site at GEDmatch. I requested this feature from them yet again…
@Jason Lee If these were your comparisons, you could run the graphical one-to-one at GEDmatch with the full resolution box checked. With that output, you get tick marks below the chromosome to show you indicators of the Mbp.
This is amazing! That’s all I can say.
One MUST NOT get too upset when doing a DNA test. and the results may be surprising, possibly under the heading of “SHOCKING”.
We are, NOW, who WE are! Cannot be changed.
just digest the information, and get on with your life.
We CANNOT condemn our ancestors for the way that they lived and or how they lived.
Times were very hard back then. “You Had To Be There”, as the movie name tells. ( think it was with Peter SELLERS.)
Maybe do the Lazardous utility (Phasing) and create a parent’s missing DNA.
Then run the Diagnostic Utility, “Are Your Parents Related”, from the Home page. On those newly created kits.
May tell you how or who are related to who.
a fellow novice researcher once exclaimed, “My Gosh, my grandmother did NOT know how to READ nor WRITE. How appalling. Did not know she was SO IGNORANT”.
Definitely, one of those “You Had To Be There” type events. People today have no clue at all, what our ancestors lived, struggled, worked their life to the bone and more.
Thank HEAVENS for them, or We would not be here today.
Still struggling, grasping, clawing out a living, to keep the wolf from the door”. Some more than others.
Definitely a GIVE ME GIVE ME GIVE ME, world. ALL FOR FREE.
(Free as in who do they think are paying for it?)
The Best Always for your continued research.
What a great analysis Kitty!
Thanks all, I made a diagram to help everyone understand these relationships and added it to the post plus I gave more fake names
This is not on the order of a “reply” but could I get some guidance? Where should I post this long-form question that reveals my lack of knowledge and progress?
Kitty, I admire your knowledge and the fact that you explain your reasoning. I’m still stuck on the very most basic moves to learn connections and possible matches. I just don’t have time to concentrate on getting smart enough to begin any type of analysis. I’ve read much of this string and find it fascinating. I can almost follow the techniques, just not the content. How very delicate one must be in rounding up the folks in this dilemma. I don’t even suspect anything like these problems in my research, but have brick walls I feel could be broken down if I could just go beyond the actual spending $$ and poking around “results”. I have had Ancestry tests for myself and my son, and FTDNA for my husband and male cousin on my father’s side. I have an additional FTDNA kit I could use but now feel like it would be redundant, after asking around to get opinions from others. I’m looking for great great grandparents (whose names I know but can’t find/identify) – my grandmother’s grandparents on her father’s side; Then, my father’s great great grandparents on his father’s side. My father is no longer living, so I got my last remaining male cousin to take the FTDNA test. Is there a chance I can find what I’m looking for? I’ve said that if I could just get a detailed list or flow chart of “moves” I might be able to stop spinning my wheels. I can’t sneak time away from my caregiving duties here to get smart! So frustrated.
Yes you can possibly find those great great grandparents with DNA. You need to get a bit lucky and have some people descended from their parents or grandparents who test DNA and match you. To improve your chances of finding those matches get as many cousins descended from that couple to test as you can. Today is the last day of the sale for
Ancestry.com DNA testing!
Also build as robust and deep a tree at ancestry as you can (and be sure your DNA is linked to it)
Ancestry.comhas a brilliant DNA to tree matching system. See how I used it for a similar case here:
As far as learning and understanding, this all takes time. Good reading to get smart are Kelly Wheaton’s series of lessons on Genetic Genealogy:
Also Blaine Bettinger’s latest book, The Family Tree Guide to DNA Testing and Genetic Genealogy. See my article on Basics here:
mskitty, I’m looking at 302 pages of matches! That’s on one of the tests (mine)! For my husband, I’ve only had 3 matches and one of them looks so similar I don’t know how the term works, how to describe, but comparing the numbers they are so similar – yet I’ve heard that things are not always what they seem and sometimes similarity of numbers doesn’t mean much. Oy! Also, I saw a Kindle ad for the FT publication, but, wondered if there haven’t been more and better discoveries and methods since it came out in 2016? Otherwise, it did seem that I was following the info pretty well, for a change. Just need the basics, ma’am! I’ll get it if it isn’t outdated, but how to know? There’s no way to do the duties here at home and meanwhile try to read and absorb all of this – but I’m just so curious!
One way to learn is by doing. Stick to the top matches and try to figure out how they are related to you. As you go along you will get better at that, read Kelly’s lessons, also try googling your specific questions and see what comes up. I have lots of posts here … and enjoy the process!
Thank you! I don’t know where or how to start, but will keep poking.
Hi Kitty, Am totally new to this. Do you know if it would be safe to download my dna kit from Ancestry to GEDMATCH. I need a x-chromosome test to show if my half sisters share the same dad or not. We all have our dna on ancestry and it shows we could be either half sisters, Aunt, or first cousins.
What do you mean by safe? At GEDmatch anyone with your kit number can use the tools there to look at your DNA results. See
If that is a concern for some reason, perhaps upload to family tree DNA instead where you can also compare the X
I m very new atnthis and feel inadequate to help another person. However, I recently came across a person who shares 170 cMs and 8 segments of DNA. Both Ancestry and Blaine Bettinger tell me that this person and I are much closer than third cousins and likely are 2nd cousins. This person was adopted in a closed adoption. When she looks at my ethnicity she doesn’t see Iberian which is on her chart. She was told her birth mother was Irish and she herself has none according to her autosomal test on Ancestry.com. I’m a lot of Irish and almost 98% of my family on both sides either have traces or percentages of Iberian. On GEDMATCH I do too. What can I do for her? She’s a little cautious and finds it difficult to believe she’s found possible blood family and I’m nervous that I may throw caution to the wind. I have suggested that she upload to GEDMATCH and/or join for a trial subscription so she can learn more. I didn’t mention that she is the first cousin who has traces of Eastern Jews. I find a mininimum of one to two out of six weekily starting at third cousin level. Is there an adoption group that I should recommend to her. What level of matching do you think we share. I love your blogs and you’re very kind to newbies while members have so patience with us at all. Thank you.
You are helping her already by being so accepting and being so willing to help. If you are 2nd cousins, then one set of your great grandparents are also hers. So a way for you to help is to build a good family tree that she can use.
The ancestry compostion always needs to be taken with a grain of salt as there is wide variation in the predictions at different companies so don’t set that much stock by it.
There is a $49 Ancestry Insights membership that is for trees only but she has to call in to get it
See my page of advice for adoptees for more suggestions: http://blog.kittycooper.com/dna-basics/help-for-adoptees/
I’ve found something unexpected in my results and was wondering if you could guide me. I’ve found a connection that doesn’t match me at all on paper. She’s not on my mother’s side (no shared DNA). I’ve found DNA connections to her great great grandparents on her father’s side, so my line is likely the iffy part. My brother’s YDNA matches my father’s line (names almost all match ours), and I’m waiting for his family finder results, which I added later. My mother and my brother have been perfectly happy to provide their DNA for testing (and my father passed away many years ago), so I’d be a bit surprised if it were my father that was the issue. I haven’t brought it up though because I want more info before I do. Where can I go from here? The GEDMatch comparison for my new relative is below. =)
Total cM 949.2 Largest cM 164.8 Gen 2.0 XTotal 28.7 Largest 21.2
Thanks so much, Melinda
Well that amount of DNA is likely either a first cousin or a half niece/aunt. So you have other matches to her gg-grandparents line? Which of your paternal lines have no matches? That is likely where the unexpected parent would be.
Could your Dad’s father not be the expected person? These tips from Michele can be helpful for helping narrow down where the break might be.
It’s not likely to be my grandfather since my brother’s Y matches are as expected. So it’s likely my father or my grandmother’s line. My brother’s autosomal results are due in a couple of weeks, so I’ll know more then. I have also retested through ancestry since I find it easier to follow lines and find line breaks there. My father and I were super close and I’m his baby girl no matter what (was always told I look like him and I definitely have his character :), so even if that’s the break, daddy is still daddy. The biggest issue once I sort this out is deciding who to tell and when or if I should. That will be the hard part. Small town USA, so once one person knows everyone will.
Anyway thanks for the response. I’ll update once I get my results in case anyone else finds themselves with similar results.
I administer tests for a brother (B) and sister (S). They were always assumed to be full siblings, but they only share 2.152 cm on Gedmatch. B’s daughter shares 1251 cm with her aunt (S). There were rumors of an affair between the siblings’ mother and their father’s brother (U). U’s grandson shares 335 cm with B, but 1215 cm with S according to Ancestry. U’s grandson is thought to be 1st cousins 1x removed to B and S. Is the extremely high match to S conclusive evidence that U is the biological father of S?
Shouldn’t B and S share about 2550 cm if their fathers were full brothers? Does the lower cm match suggest that the two men may have been half brothers?
I posted to a group about this, and the suggestion was that GedMatch does not count fully identical regions twice, and may therefore underestimate the cm count. Is there an alternative tool that I can use to solve this mystery?
The amount shared can vary fairly widely, so no 2125 is just on the low side. If S and B are tested I can often tell. It is the other matches that will prove or disprove your theory. Can you test more known descendants of U?
I will email you.
I have seen a similar case of a double relationship, half-sibling and half-nibling with 2002 shared cM. I believe you could add the FIR green segments in your GedMatch results to get a value for the total shared segments closer to the average value of 2550 cM you might expect for the two individual relationships.
Exactly right Richard and, as you pointed out in your email, this is defined as a 3/4 relationship by Wikipedia.
Yes, Patty and Bob would be 3/4 siblings.
Hi Kitty! I have sent you a few emails tonight via your contact form with a question on this topic. I am trying to find out if my brother is a half sibling as well as another close relation to me- we share 9 identical segments over 100, the highest being 263.32. We share half match on each of my x chromosomes, 83 and 30 cM; and ancestryDNA reported us being 2688 cM match (36%) and 23andme reported us being 50.9% cM match. I know 23andme can be more than others but that seems like a large difference. Can any of this help confirm some questions I’ve had either way? Thank you in advance.
Apparently my contact form broke with recent updates at my host or of wordpress … I noted that and put my email on my contact me page
and answered your question, full brother
I may have a 3/4 sibling. Could use some help. Nephew matches 1519
That is a totally normal amount of DNA to share with a nephew. You can see that if you use the calculator at https://dnapainter.com/tools/sharedcmv4
Love your blog! Sister and I share 52.4% dna, 3901 cm via 23andMe. Have maybe identified cousins that maybe 3rd on dad’s side and 4th on mom’s via Ancestry (about 3 so far). Does this sound any out of the ordinary? Thanks so much if you can answer this silly question. Thanks.
I have been trying to understand the dna of 3/4 siblings however I’m not sure how to determine. My brother and I share a Mother but because of rumors over the years I found out that my Fathers full brother could potential be my biological Father. Unfortunately my Mother, Father and Uncle are deceased. My brother and I took the Ancestry test and share 2406 cM shared. On Gedmatch our (FIR) are the following
Largest segment = 49.6 cM
Total Full-Match segments (FIR) = 563.4 cM (15.709 Pct)
23 shared segments found for this comparison.
430821 SNPs used for this comparison.
66.947 Pct SNPs are full identical
Also my brother shares 1604 cM with my Daughter, not sure that helps but I really would like to know. Any help would be greatly appreciated.
Wanted to add to my previous post also the Gedmatch Estimated number of generations to MRCA = 1.3 not sure if that helps as I’m new to learning DNA and understanding it. Thank you
Those numbers indicate a 3/4 relationship, as you were expecting. Adding the FIRs to the HIRs gets about 2969 cM which is too low for full siblings but just right for 3/4.
Read this post
Thank you so much for your help and sending the link which really helped me to understand by adding HIR and FIR’s. I wanted to ask about Ancestry results of 2406 shared cM vs Gedmatch with the HIR results being:
Largest segment = 124.2 cM
Total Half-Match segments (HIR) = 2464.8 cM (68.727 Pct)
Estimated number of generations to MRCA = 1.3
59 shared segments found for this comparison.
Total Full-Match segments (FIR) = 563.4 cM (15.709 Pct)
If I add the HIR and FIR’s from Gedmatch it is 3,028.2 cM. Would that still be 3/4 th siblings which I highly expected? Just want to make sure I’m understanding correctly if Ancestry and Gedmatch do shared cM differently. Your help is greatly appreciated.