A new approach to problem solving is to appeal to social media. So here is my request to “crowd” collect statistics about X chromosome matches. If you have multiple close family members DNA tested then please help me by filling out this form for each relationship at http://goo.gl/forms/4294xjjhqq
I have already put this request on DNA_NEWBIES and facebook so the form is now, hopefully, well tested and I will list some of my preliminary results at the end of this post. I have to give credit to Blaine Bettinger for this data collection idea; he is collecting autosomal statistics this way, see my post about that and the results so far.
Now why did I decide to collect these statistics aside from my own curiosity? To help an adoptee of course.
Recently I got an interesting question on one of my X chromosome posts. I was asked if the matching X could determine whether the uncle or the aunt of a female first cousin was the parent of an adoptee. The autosomal match with yet another first cousin identifies the shared grandparents and the female cousin is descended from them via her father. Neither potential parent, the aunt or uncle, has any known progeny nor are they available to be tested. Well?
My initial reaction was no way, X is too weird and undependable; but then I thought about it some more.
Since men only get one X chromosome, they pass the one they got from their mother intact to their daughters. Thus this female cousin has an unrecombined X from her Dad, half of her grandmother’s X. Therefore if this X chromosome match looks look like two brothers or a brother and a sister, it would be the uncle who is the father of our adoptee.
But how much X DNA do siblings, where at least one is a male, share? Is this predictable? Is it the same roughly 50% as they share in their autosomes?
In my experience, male or male/female siblings usually share about half to two thirds of the X so 90-120cM out of the total of 196cM (at Gedmatch and Family Tree DNA) or 182cM (at 23andme) with either one or two recombination points
Here is the match:
|Chr||Start Location||End Location||Centimorgans (cM)||SNPs|
The total shared cMs are about 70. This would suggest the aunt is the mother since that is not as much shared X DNA as would be expected if the uncle is the father. The expected amount of shared X between these first cousins, where one is getting her related X from her father and the other from her mother, is about 25% or about 65 cM as I show in my inheritance picture.
However I was not sure that the small sample of data that I have worked with was adequate for this conclusion. Also I wondered whether my observation that shared X usually exceeds the expected mathematical percentages was valid, so I decided to collect some more statistics via crowd sourcing.
Here are a few preliminary statistics from about 50 and 100 data points. As you can see the expected amount for an uncle to his sister’s child fits our case.
The collected numbers for a brother and his sibling show that they usually share about half of the X which is statistically the expected result. And would also be the expected result if our adoptee’s father was the brother of this cousin’s father.
But these are just likely amounts and there are plenty of outliers. So I suggested that our adoptee test another sibling of the cousin to get us more input. Plus I am hoping for more data from my crowd sourcing effort.
The form is at at http://goo.gl/forms/4294xjjhqq – thank you for helping out.
I’m going to fill this out for my family members who have tested, but you may want to add another field to the form to ask about other information that may be helpful. With all branches being Ashkenazic Jewish, the amount of shared DNA on the X chromosome will likely be skewed higher than would be expected.
In my experience, the endogamous nature of a population has little effect on very close relationships. That effect is seen in further relationships.
You still get 50% from a parent.
And what I am studying is very close relationships
But you talked me into adding a checkbox for endogamy
I’m in agreement with Lara. I will be filling out some forms to help with this, but believe your selections might need explanation. I will enter my sister’s daughter, but it’s my half sister, so would think it would be helpful to know that when compiling your data. So a comments box seems as if it would help. I’ll wait to see if you can supply that feature.
Half sister’s daughter if you share a mother is like an uncle to sister’s child, but best to leave that one out
GEDmatch has an X matching tool that uses the sort of logic you are following. I have found it of great assistance for matches that are further out. It shows who can’t be involved and puts probabilities on who might.
Yes that is their “Relationship Tree projection” – I will be sure to give them a copy of the results of my study to help with that and of course I need to try it out and blog about it, which I have not done yet
I submitted all the permutations of my large family (5 brothers, 2 sisters, 6 1st cousins and 4 grandchildren) that matched your criteria for inclusion. Even though the family is heavily skewed to males, there were enough of we females to make it interesting.
I just submitted several for you. Hopefully I remembered to check the endogamous box each time.
My brother to me and to my daughter. My dad to his brother and to my daughter. My uncle to me.
How do I appeal to social media? I have left questions on message boards before but have usually found the links to be old, no one responds or the response has not been helpful? I am a totally right brained person with not a logical bone in her body that needs to understand how I can use my DNA results to find my gg grandfathers parents. After searching for 4 years on the Prince surname, I am told that William Prince was actually fathered by a Rose and born to a Prince female. I have daddy’s DNA with ancestry, gedmatch.com and also a Y-DNA descendant with family tree and gedmatch. Even the websites help line don’t seem to have anyone that can interpret the findings so that you know what you are looking for. Any ideas? Lisa Montgomery
There are a number of great facebook groups for genetic genealogy try this one:
There is a steep learning curve with this stuff. Test with 23andme also, get more cousins to test (more data can help with this), and perhaps take a class at DNAadoption.com
Has the Y descendant done a Y test at ftDNA?
You could of course pay some one to help you, perhaps an hour of someone’s time would do it … but learn more first perhaps
I would love to help, but my closest tested is a known second cousin. I don’t even see many names that I recognize at FTDNA and even less at Gedmatch. Sounds like a good study though.
I could provide x data for me and mom via ftdna, and my paternal half-sister via ancestry. would that help your project?
Jenny, You and your mom are fine to contribute but not the half sister. If you are all on GEDmatch then please also contribute your half sister’s match with you to my other project, see http://blog.kittycooper.com/2016/08/how-to-tell-the-relationship-from-the-shared-dna/
I see your X form, however, I do not see where you ask for the size of the X on that form. Did I miss it or do you only want those that have some amount of X matching, no matter how much?
Thanks for all you do. This is great work!
I am not sure I understand your question.
Did you click through to the form?
It asks for the amount shared in centimorgans (that is the size), total segments, and largest segment. Zero is an acceptable answer.
So you are asking only for total cMs NOT the amount of shared X?
I am asking for the total shared cMs on the X chromosome only
Can you clarify the grandmother/grandhild relationship on your X-chr form? Do you just want cases where a maternal grandmother is involved? I inputted just now one relationship w/ female grandchild, and Paternal grandmother. Of course they share a full X — duh. I’m messing up your data!
Also, what about relationships between half-1st cousins whose mothers are paternal half-siblings? Can that go under “other first cousins with expected shared X” ?
You are absolutely right, I changed it to maternal grandmother.
Also that case is fins under other first cousins.
Hello Kitty. After submitting a DNA Test to Ancestry.com, I have found a match who shares 1,742 centimorgans and 73 DNA segments with me. My daughter shares 3,391 centimorgans and 95 DNA Segments with me. The first would be a Paternal match. My Father was of Puerto Rican descent… and so is this match. I would think this would be a Half Sibling. How can I tell? I have tried to understand GEDMatch… but I just don’t get it. Can you help me figure it out? Unfortunately, this person has not logged in to see the many messages I have left. I AM SITTING ON PINS AND NEEDLES!!! I MUST KNOW HOW WE ARE RELATED!!! HELP PLEASE!!!
He could be a half sibling or an uncle. If his DNA results are on GEDmatch, you can figure it our from the segment sizes since he is paternal. If he is only on ancestry then if it is a lower number of segments probably a half brother. Read this post about that very problem:
As far as getting him to respond to your messages, he may not be seeing them, read this:
If his name is not too common try looking for him on Facebook or google for his name or username… lots of tips in that article and the article it links to by Blaine Bettinger.
Thanks Kitty for all of your Great work.
Question: With the X-Chromosome, many researchers content that only segments that are over 20cMs are valid for searching and matching. However, are small segments 2 to 5 cMs valid for attributing the match to the maternal family or are these segments IBS? I have a match with a person that had a total of approximately 20 cMs with these very small segments. How should I look at the total segments ( with small values) that are less than 20 cMs?
Male to male segments as low as 3cM can be valid
Male to female segments 7-10 cM
Female to female 15 cM
Since X can last a long time, segments less than 20 cM are not worth pursuing without autosomal matching as welll
Also those numbers are not from scientific research, just from my own experience and from talking with others. Lots more reading available at ISOGG see https://isogg.org/wiki/X-chromosome_testing
Are you still collecting x-dna data?
Yes I am still collecting
My brother’s X chromosome varies wildly with his three full sisters. They share anywhere from 13,850 SNPs to 4,450 SNPs to 2,000 SNPs. That range doesn’t exist on any of the other chromosomes.
His extreme variation on the X chromosome also occurs with two male first cousins who descends from sisters. The similarity matches range from 404 SNPs to 5,853 SNPs. Since they all get just one X chromosome from their mothers and those women are full sisters, shouldn’t the X chromosome be far more similar?
We prefer to discuss DNA segments in terms of cMs.
The reason his range varies so with his full sisters is that he had only ONE X chromosome from his mother while they have two, one from each parent. Every other chromosome he shares with his sisters can be from either or both parent(s).
As for his cousins there is no reason to expect more similarity. They each get very different cobinations of X from their mothers some of shich was from the unrelated side.