That Unruly X….Chromosome That Is


Something is wrong with the X chromosome.  More specifically, something is amiss with trying to use it, the way we normally use recombinant chromosomes for genealogy.  In short, there’s a problem.

If you don’t understand how the X chromosome recombines and is passed from generation to generation, now would be a good time to read my article, “X Marks the Spot” about how this works.  You’ll need this basic information to understand what I’m about to discuss.

The first hint of this “problem” is apparent in Jim Owston’s “Phasing the X Chromosome” article.  Jim’s interest in phasing his X, or figuring out where it came from genealogically, was spurred by his lack of X matches with his brothers.  This is noteworthy, because men don’t inherit any X from their father, so Jim’s failure to share much of his X with his brothers meant that he had inherited most of his X from just one of his mother’s parents, and his brothers inherited theirs from the other parent.  Utilizing cousins, Jim was able to further phase his X, meaning to attribute portions to the various grandparents from whence it came.  After doing this work, Jim said the following”

“Since I can only confirm the originating grandparent of 51% my X-DNA, I tend to believe (but cannot confirm at the present) that my X-chromosome may be an exact copy of my mother’s inherited X from her mother. If this is the case, I would not have inherited any X-DNA from my grandfather. This would also indicate that my brother Chuck’s X-DNA is 97% from our grandfather and only 3% from our grandmother. My brother John would then have 77% of his X-DNA from our grandfather and 23% from our grandmother.”

As a genetic genealogist, at the time Jim wrote this piece, I was most interested in the fact that he had phased or attributed the pieces of the X to specific ancestors and the process he used to do that.  I found the very skewed inheritance “interesting” but basically attributed it to an anomaly.  It now appears that this is not an anomaly.  It was, instead the tip of the iceberg and we didn’t recognize it as such.  Let’s look at what we would normally expect.


The X chromosome does recombine when it can, or at least has the capacity to do so.  This means that a female who receives an X from both her father and mother receives a recombined X from her mother, but receives an X that is not recombined from her father.  That is because her father only receives one X, from his mother, so he has nothing to recombine with.  In the mother, the X recombines “in the normal way” meaning that parts of both her mother’s and her father’s X are given to her children, or at least that opportunity exists.  If you’re beginning to see some “weasel words” here or “hedge betting,” that’s because we’ve discovered that things aren’t always what they seem or could be.

The 50% Rule

In the statistical world of DNA, on the average, we believe that each generation receives roughly half of the DNA of the generations before them.  We know that each child absolutely receives 50% of the DNA of both parents, but how the grandparents DNA is divided up into that 50% that goes to each offspring differs.  It may not be 50%.  I am in the process of doing a generational inheritance study, which I will publish soon, which discusses this as a whole.

However, let’s use the 50% rule here, because it’s all we have and it’s what we’ve been working with forever.

In a normal autosomal, meaning non-X, situation, every generation provides to the current generation the following approximate % of DNA:

Autosomal % chart

Please note Blaine Bettinger’s X maternal inheritance chart percentages from his “More X-Chromosome Charts” article, and used with his kind permission in the X Marks the Spot article.

Blaine's maternal X %

I’m enlarging the inheritance percentage portion so you can see it better.

Blaine's maternal X % cropped

Taking a look at these percentages, it becomes evident that we cannot utilize the normal predictive methods of saying that if we share a certain percentage of DNA with an individual, then we are most likely a specific relationship.  This is because the percentage of X chromosome inherited varies based on the inheritance path, since men don’t receive an X from their fathers.  Not only does this mean that you receive no X from many ancestors, you receive a different percentage of the X from your maternal grandmother, 25%, because your mother inherited an X from both of her parents, versus from your paternal grandmother, 50%, because your father inherited an X from only his mother.

The Genetic Kinship chart, below, from the ISOGG wiki, is the “Bible” that we use in terms of estimating relationships.  It doesn’t work for the X.

Mapping cousin chart

Let’s look at the normal autosomal inheritance model as compared to the maternal X chart fan chart percentages, above, and similar calculations for the paternal side.  Remember, the Maternal Only column applies only to men, because in the very first generation, men’s and women’s inheritance percentages diverge.  Men receive 100% of their X from their mothers, while women receive 50% from each parent.

Generational X %s

Recombination – The Next Problem

The genetic genealogy community has been hounding Family Tree DNA incessantly to add the X chromosome matching into their Family Finder matching calculations.

On January 2, 2014, they did exactly that.  What’s that old saying, “Be careful what you ask for….”  Well, we got it, but “it” doesn’t seem to be providing us with exactly what we expected.

First, there were many reports of women having many more matches than men.  That’s to be expected at some level because women have so many more ancestors in the “mix,” especially when matching other women.

23andMe takes this unique mixture into consideration, or at least attempts to compensate for it at some level.  I’m not sure if this is a good or bad thing or if it’s useful, truthfully.  While their normal autosomal SNP matching threshold is 7cM and 700 matching SNPs within that segment, for X, their thresholds are:

  • Male matched to male – 1cM/200 SNPs
  • Male matched to female – 6cM/600 SNPs
  • Female matched to female – 6cM/1200 SNPs

Family Tree DNA does not use the X exclusively for matching.  This means that if you match someone utilizing their normal autosomal matching criteria of approximately 7.7cM and 500 SNPs, and you match them on the X chromosome, they will report your X as matching.  If you don’t match someone on any chromosome except the X, you will not be reported as a match.

The X matching criteria at Family Tree DNA is:

  • 1cM/500 SNPs

However, matching isn’t all of the story.

The X appears to not recombine normally.  By normally, I don’t mean something is medically wrong, I mean that it’s not what we are expecting to see in terms of the 50% rule.  In essence, we would expect to see approximately half of the X of each parent, grandfather and grandmother, passed on to the child from the mother in the maternal line where recombination is a possibility.  That appears to not be happening reliably.  Not only is this not happening in the nice neat 50% number, the X chromosome seems to be often not recombining at all.  If you think the percentages in the chart above threw a monkey wrench into genetic genealogy predictions, this information, if it holds up in a much larger test, in essence throws our predictive capability, at least as we know it today, out the window.

The X Doesn’t Recombine as Expected

In my generational study, I noticed that the X seemed not to be recombining.  Then I remembered something that Matt Dexter said at the Family Tree DNA Conference in November 2013 in Houston.  Matt has the benefit of having a full 3 generation pedigree chart where everyone has been tested, and he has 5 children, so he can clearly see who got the DNA from which of their grandparents.

I contacted Matt, and he provided me with his X chromosomal information about his family, giving me permission to share it with you.  I have taken the liberty of reformatting it in a spreadsheet so that we can view various aspects of this data.

Dexter table

First, note that I have sorted these by grandchild.  There are two females, who have the opportunity to inherit from 3 grandparents.  The females inherited one copy of the X from their mother, who had two copies herself, and one copy of the X from her father who only had his mother’s copy.  Therefore, the paternal grandfather is listed above, but with the note “cannot inherit.”  This distinguishes this event from the circumstance with Grandson 1 where he could inherit some part of his maternal grandfather’s X, but did not.

For the three grandsons, I have listed all 4 grandparents and noted the paternal grandmother and grandfather as “cannot inherit.”  This is of course because the grandsons don’t inherit an X from their father.  Instead they inherit the Y, which is what makes them male.

According to the Rule of 50%, each child should receive approximately half of the DNA of each maternal grandparent that they can inherit from.  I added the columns, % Inherited cM and % Inherited SNP to illustrate whether or not this number comes close to the 50% we would expect.  The child MUST have a complete X chromosome which is comprised of 18092 SNPs and is 195.93cM in length, barring anomalies like read errors and such, which do periodically occur.  In these columns, 1=100%, so in the Granddaughter 1 column of % Inherited cM, we see 85% for the maternal grandfather and about 15% for the maternal grandmother.  That is hardly 50-50, and worse yet, it’s no place close to 50%.

Granddaughter 1 and 2 must inherit their paternal grandmother’s X intact, because there is nothing to recombine with.

Granddaughter 2 inherited even more unevenly, with about 90% and 10%, but in favor of the other grandparent.  So, statistically speaking, it’s about 50% for each grandparent between the two grandchildren, but it is widely variant when looking at them individually.

Grandson 1, as mentioned, inherited his entire X from his maternal grandmother with absolutely no recombination.

Grandsons 2 and 3 fall much closer to the expected 50%.

The problem for most of us is that you need 3 or 4 consecutive generations to really see this happening, and most of us simply don’t have data that deep or robust.

A recent discussion on the DNA Genealogy Rootsweb mailing list revealed several more of these documented occurrences, among them, two separate examples where the X chromosome was unrecombined for 4 generations.

Robert Paine, a long-time genetic genealogy contributor and project administrator reported that in his family medical/history project, at 23andMe, 25% of his participants show no recombination on the X chromosome.  That’s a staggering percentage.  His project consists of  21 people in with 2 blood lines tested 5 generations deep and 2 bloodlines tested at 4 generations

One woman’s X matches her great-great-grandmother’s X exactly.  That’s 4 separate inheritance events in a row where the X was not recombined at all.

The graphic below, provided by Robert,  shows the chromosome browser at 23andMe where you can see the X matches exactly for all three participants being compared.

The screen shot is of the gg-granddaughter Evelyn being compared to her gg-grandmother, Shevy, Evelyn’s g-grandfather Rich and Evelyn’s grandmother Cyndi. 23andme only lets you compare 3 individuals at a time so Robert did not include Evelyn’s mother Shay, who is an exact match with Evelyn.

Paine X

Where Are We?

So what does this mean to genetic genealogy?  It certainly does not mean we should throw the baby out with the bath water.  What it is, is an iceberg warning that there is more lurking beneath the surface.  What and how big?  I can’t tell you.  I simply don’t know.

Here’s what I can tell you.

  • The X chromosome matching can tell you that you do share a common ancestor someplace back in time.
  • The amount of DNA shared is not a reliable predictor of how long ago you shared that ancestor.
  • The amount of DNA shared cannot predict your relationship with your match.  In fact, even a very large match can be many generations removed.
  • The absence of an X match, even with someone closely related whom you should match does not disprove a descendant relationship/common ancestor.
  • The X appears to not recombine at a higher rate than previously thought, the previous expectation being that this would almost never happen.
  • The X, when it does recombine appears to do so in a manner not governed by the 50% rule.  In fact, the 50% rule may not apply at all except as an average in large population studies, but may well be entirely irrelevant or even misleading to the understanding of X chromosome inheritance in genetic genealogy.

The X is still useful to genetic genealogists, just not in the same way that other autosomal data is utilized.  The X is more of an auxiliary chromosome that can provide information in addition to your other matches because of its unique inheritance pattern.

Unfortunately, this discovery leaves us with more questions than answers.  I found it incomprehensible that this phenomenon has never been studied in humans, or in animals, for that matter, at least not that I could find.  What few references I did find indicated that the X seems to recombine with the same frequency as the other autosomes, which we are finding to be untrue.

What is needed is a comprehensive study of hundreds of X transmission events at least 3 generations deep.

As it turns out, we’re not the only ones confused by the behavior of the X chromosome.  Just yesterday, the New York Times had an article about Seeing the X Chromosome in a New Light.  It seems that either one copy of the X, or the other, is disabled cell by cell in the human body.  If you are interested in this aspect of science, it’s a very interesting read.  Indeed, our DNA continues to both amaze and amuse us.

A special thank you to Jim Owston, Matt Dexter, Blaine Bettinger and Robert Paine for sharing their information.

Additional sources:

Polymorphic Variation in Human Meiotic
Recombination (2007)
Vivian G. Cheung
University of Pennsylvania

A Fine-Scale Map of Recombination Rates and Hotspots Across the Human Genome, Science October 2005, Myers et al
Supplemental Material



I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.

Thank you so much.

DNA Purchases and Free Transfers

Genealogy Services

Genealogy Research

224 thoughts on “That Unruly X….Chromosome That Is

  1. Another superb article Roberta–THANKS for sharing your knowledge and experience with us all! Sent you an email with a few more examples. Enjoy! 🙂

  2. Great article.

    My one known cousin match on the X-Chromosome is between my father and his second cousin. From the common ancestral pair, dad is descended F-F-M, while his cousin is M-F-M. According to FTDNA, they X-match in one big block, 40.27cM / 2742 SNPs. That seems pretty reasonable for a 50/50 split along the way, although I haven’t worked through all the math.

  3. Maybe this will help with one instance. Me and my wife are cousins 3rd 2xremoved. we share only one set of grandparents and they are my 2nd Ggrandparents (and yes its an Xdna Path). They are her 4th Ggrandparents.
    According to FTDNA FF we share 31.47 of the longest block on the 9th cH.
    On the Xdna we share 2 segments one is 12.2 and the other is 22.56

    Hope that helps in someway.

  4. You say X is just under 200cM long, so on average one would expect 2 recombination events per maternal transmission, and, assuming a poisson distribution, about 13% of transmissions will have no recombination. Many of the other chromosomes are of similar size* so the rate may be similar in other chromosomes as well, but I suspect for genetic genealogy we rarely think about just one autosome and one transmission.

    * Morton reported a range of 68-392cM, which would give a range of 50%-2% without recombination, depending on the autosome.
    Morton, N. E. (1991). Parameters of the human genome. Proceedings of the National Academy of Sciences, 88(17), 7474-7476.

    • It may be an oversimplification, but I like to think of X, for the purposes of genetic genealogy, as just like another autosome — except that it isn’t passed from father to son and it can’t recombine when passed from father to daughter. There’s really nothing special about X beyond that.

      Autosomes can be as unruly as X.

      If I focused on a single autosome, e.g. Chromosome 22, I might conclude that my brother is probably unrelated to my sister and me. In the case of Chromosome 13, we look like distant relatives (only 18 cM shared). But looking across the whole genome, a more predictably typical pattern emerges.

      • Um… yea, it’s still special. without the x chromosome at all, there’d be alot of messed up males since the x chromosome is responsible for many human processes. The y chromosome is only responsible for making a human male.. as long as it has the sry gene, that is. A female with a missing x chromosome (usually missing from the father’s contribution ) is still a female. 🙂

  5. So, if I understand this correctly, two sisters should share one identical x-chromosome that they received from their father. Can this information be used to phase the X? (and separate
    the maternal allele from the paternal allele?)

    • If you have one of the parents, yes. The problem is that they also share some of the X of both parents and without one of the parents, you can’t sort it out. Or at least I don’t know how you would.

      • Thanks Roberta.

        What about phasing using the X-chromosome from a brother in comparison with the X-chromosome of the sister? I am trying to ferret out the paternal X-chromosome
        from the sister.

  6. Great article! It explains a X match with a 3rd cousin once removed. I shared around 58 cM (out of the total of 196 cM which X has) with that person and using standard inheritance rules I could not figure out why this region was so large.

  7. Roberta, does this have any implications for autosomal DNA. I know autosomal does not define gender like X & Y but if these X results reveal things we never knew before, are the experts considering revisiting what they think they know about atDNA?

  8. I was able to test my grandmother’s atDNA at Her results are being uploaded to Gedmatch. After reading this article, I look forward to comparing our DNA and our X chromosome.

    I’ve also tested my father and am excited to review our X data as well. Thanks for this article 🙂

  9. Are those “% inherited…”s percentages or decimal proportions of the wholes”?
    Yes. I am nitpicking at an excellent blog.

    • They are decimal proportions of the whole which is why I gave the example of how to interpret them. Plus, there wasn’t enough room at the top of the column for the full name:) I actually agonized quite a bit over how to label the columns.

  10. Roberta, as the oldest child of a family with six children and a mother with 5 siblings, I had many cousins I knew. I noticed when I was young that many cousins and siblings (descendants) were more like a grandparent than a parent or sibling. Sometimes it was the hair color, sometimes the hands, sometimes the sense of humor, sometimes the height, etc. Thanks for all you do.


  11. Just when you think you “know” how DNA behaves…there’s that old “empirical evidence” that it doesn’t really do what you thought it did. LOL Never a dull moment.

    Thanks for a great article!

  12. This is very interesting and reminded me of the Barr bodies we studied in genetics 30 years ago when I took several genetics classes in college and grad school. I don’t remember if they talked about crossing over in X chromosomes at time but they sure knew that one X was inactive in every somatic cell in the female body. I wish I had kept up with all the changes in knowledge over the years but after grad school I only had to use my knowledge of genetics a few times and didn’t keep up with the literature. A lot has changed in the past 30 years!

  13. This really confused me: Not only does this mean that you receive no X from many ancestors, you receive a different percentage of the X from your maternal grandmother, 25%, because your mother inherited an X from both of her parents, versus from your paternal grandmother, 50%, because your father inherited an X from only his mother.
    I thought you were referring to us men, but then I guess you mean in general, men or women, it doesn’t matter. But that a female would inherit 50% (of the X) from her paternal grandmother. Or is it 100% since what a female inherits from her father’s X is directly what was passed onto him from his mother.

  14. Thanks for all the information Roberta. I kept reading all the postings on the X Chromosome but didn’t really understand it until I used Pierre’s Charting Companions Chart. Aleda told me how to get the program and I love it. I appreciate all the information you share with us.

  15. “If you don’t match someone on any chromosome except the X, you will not be reported as a match.”

    This explains why my X Chromosome matches at FTDNA are suspiciously paltry! But this makes absolutely no sense to me at all. I can see no reason for it.

    They could say, “Well, if you match someone on Chromosome 21, we won’t report it unless you match that person somewhere else as well.”


        • Odd.

          My husband has about 270 autosomal matches and only 2 X matches…but a few days ago, he had 6 X matches.

          Hmmm.  Did FTDNA change their threshold for reporting X matches?

      • Upon further contemplation, I would imagine they’re doing this to conserve computational resources. If this is true, they need to reconsider. As it stands, their “X-Matches” report is misleading and incomplete.

        This news about their process for finding X-Matches is upsetting in light of their dubious explanation for my small number of X-Matches, i.e., “It is not unusual to have very little or no X matches, and for the total shared X chromosome DNA to be very small. The X chromosome drops off very quickly in the inheritance process.”

        The fact that they’re not even looking at X Chromosomes outside of my group of autosomal DNA cousins is a much better, albeit disappointing, explanation.

    • rightklik is correct upthread about ANY of the paired chromosomes behaving in the same fashion: it’s possible to inherit all (or nothing) from a grandparent. It’s been reported many times, but it really jumps out at us when we focus on one specific chromosome. I have an example in my family where the entire length of chromosome 4 was inherited from a grandparent through at least four transmission events. This diagram explains how it happens: four gametes (eggs or sperm) are created, only two of which are recombined

      For FTDNA to report matches *due to* an X segment they would have to recompute their entire database, an enormous computational load because of all the pairwise comparisons that are necessary. They would also need to establish different parameters for declaring a match on the X. 23andMe did this empirically based on their simulations, as Roberta reported above. This blog post describes their general approach and has a link to the technical article

      The fact that females are getting so many more *small* segments exposes the difficulty in using genotypes (where you don’t know which allele came from which parent). Males with their single X chromosome automatically have a haplotype (we know their results are all on one chromosome, inherited from the mother).

      • Thanks for the article Roberta and thanks for the links Ann, but is it true that only 2 of the 4 chromatids have the potential to recombine? I thought that any of the chromatids could recombine with any other chromatid, be it a sister chromatid or a non-sister chromatid. The recombination of 2 sister chromatids does not produce any difference to the original DNA sequence because the 2 sister chromatids are identical. It is only when 2 non-sister chromatids combine that the original parental DNA sequence is altered. There was a post about this on the ISOGG Facebook page or Mailing List some time last year I believe, but a scientific article would be a welcome clarification if you anyone has one. Cheers, Maurice

  16. In the two charts with the column heading of Generation, the top two rows have an extra Generational “G”. The charts do not list GGG-Grandparents.

    GGGGG-Grandparents 1.56% should be GGGG-Grandparents.
    GGGG-Grandparents 3.12% should be GGG-Grandparents.

  17. I always use my own Ashkenazic background as a filter when reading these things. According to FTDNA I am 100% Middle Eastern Jewish, as is my maternal uncle and my paternal first cousin. Your explanation here contributes to my amazement and confusion when trying to do genealogy in this endogamous population.

    The sentence that really blew me away in your article is this one: “The X chromosome matching can tell you that you do share a common ancestor someplace back in time.” Currently, my paternal first cousin has 418 X chromosome matches; my maternal uncle has 523. I have 903. Other than one other relative that I have had tested, we cannot find a genealogical match with a single one of these people.

    Thank you for continuing to educate with your thorough and understandable posts. But I despair that this genealogical “tool” is practically useless for those of us trying desperately to reconstruct our lost Jewish family trees.

      • But for people who suffer from pedigree collapse, it’s not useful at all. Not to mention the X matches, due to their rate of recombination, is actually not as useful unless both matching parties have an extensive tree and/or their MRCA was very, very recent. Most people are better off using autosomal matches.

      • @Kalani

        Every chunk of DNA has its own story to tell. Yes, people from endogamous communities do have to be especially careful, particularly when making seemingly reasonable assumptions — as I have discovered from my own small town genealogy. (With one 60 cM cousin I found autosomal DNA from one line and X DNA from another.)

        I hope you don’t give up on X matches. In my case it has provided a more complete story. Unfortunately, it seems 23andMe is the only company that provides complete X DNA matching, but if you have the information, please use it.

      • @rightklik…..back in June while reading a lot about the X inheritance patterns, it was the same time when other people starting asking me because we have large amounts shared on the X. The lowest, a paternal cousin of mine – 6cM, versus others who share 20.9cM, 28.5cM & 32.2cM, and these people have been separated from our people for over 500 years. But because we lack genetic diversity, have multiple common ancestors (which makes it even more challenging), and our people have gone through multiple bottle necks, it becomes challenging, but we’re all learning from all of this. I don’t know how or if any Jewish person has been successful using the X. I did read recently about someone’s success story on the X, but hers was unique, and it was a recent ancestor, versus someone sharing less than 20cM or so. I definitely would like to hear more success stories and figure out how, if at all possible, it could be helpful to us.

        A few weeks ago I realized that my paternal aunt and my mother do share a total of 32.4cM. That’s pretty high considering both my mom’s ancestors and my paternal aunt’s ancestors (and mine of course) weren’t known to be one group of people but are in regions that are separated by several hundreds of miles. By plane it’s about 2 – 3 hours, so not sure what that is in miles, plus there were seas to cross as well. Yet they share 32cM on the X, which at first I thought was odd, but as I recently keep reading more about the X and getting input from those skilled in the field of genetics, it’s no surprise since these still could be classified as IBS, but at that amount? Still new, and we have a lot to learn about all of this.

      • @Kalani

        Yes, apparently there’s a lot left to learn about the nuts and bolts of X recombination, but to the extent that the X Chromosome fails to recombine as often as we might have expected, it will allow us to peer deeper into the past.

  18. Roberta,

    Thank you for this wonderful post! I’ve taken both the FTDNA Family Finder and mtDNAPlus tests. The whole DNA for genealogy subject is fascinating…and kind of confusing at the same time. =)

    My autosomal test results showed that I have Native American ancestry, specifically Mayan. My great-grandmother was from Chiapas. I wrote a blog post about this find explaining that this test result proved something my step-grandma had said in the past. She told my mom that my mom had Mayan ancestry. And, she was right!

    I want you to know that your two of your blog posts are listed in today’s Fab Finds post at

    Have a great weekend!

  19. Love your articles Roberta. Not trying to be dismissive but wanted to provide an illustration of why I think the idea that something special is going on with the X when it comes to recombination/crossing over is a little premature given the examples you provided.

    I just did some quick autosomal DNA comparisons on 23andme as I have both my parents and my children’s DNA to compare 3 generations. I am in the middle generation and XY.

    My son inherited all of my father’s chromosome 4, 7, 11, 15, 16 and 21. (6)
    My daughter inherited all of my father’s chromosome 4, 12 and 13 (3)

    My son inherited all of my mother’s chromosome 14 (1)
    My daughter inherited all of my mother’s chromosome 1, 8, 9, 20, 21, 22 (6). She also inherited all of my mother’s X as would be expected.

    Overall, my son shares 26.6% of his DNA with his paternal grandfather and 20.9% with his paternal grandmother. My daughter shares 23.2% of her DNA with her paternal grandfather and 26.5% with her paternal grandmother.

    In the case of my two children, to whom I gave 22 autosomal chromosomes each, no recombination occurred between my parents DNA in 16/44 (38.6%) of the autosomal chromosomes.

    I do not think the failure of the X-chromosome to recombine in a woman when it has a chance is likely to be much different than the failure of any particular autosomal chromosome to recombine.

    See user rightclik’s post above which is well worded and correct –

    “It may be an oversimplification, but I like to think of X, for the purposes of genetic genealogy, as just like another autosome — except that it isn’t passed from father to son and it can’t recombine when passed from father to daughter. There’s really nothing special about X beyond that.

    Autosomes can be as unruly as X.

    If I focused on a single autosome, e.g. Chromosome 22, I might conclude that my brother is probably unrelated to my sister and me. In the case of Chromosome 13, we look like distant relatives (only 18 cM shared). But looking across the whole genome, a more predictably typical pattern emerges.”

    • Fascinating! I love it when all the citizen scientists comment. I am working on a study involving 84 people of both Jewish and non-Jewish origin. While not specifically geared to three generations, I have 2.7 million individual segments (including the X) which I might want to take a look at regarding how things are passed on.

      • I’m still not sure what to do with matches who are not immediate family but here’s an analysis of Mother-Son-Granddaughter. The cM values on the left are what the Mother shares with her Granddaughter and the cM values on the right are what the Son shares with the Granddaughter (his daughter). It looks like the granddaughter got virtually all of Chr 1 from her grandfather, and virtually all of Chr 7, 13, 15, 16, 19, 20 and 22 from her grandmother. Chr 99 is the X and behaves as expected in this case.

        CHR Parent cM Child cM
        1 18.99 267.21
        2 197.39 253.06
        3 149.73 219.1
        4 127.21 206.75
        5 131.62 199.6
        6 91.24 188.9
        7 178.91 180.79
        8 118.66 161.33
        9 65.65 160.36
        10 90.7 176.25
        11 125.62 155.78
        13 126.48 126.48
        14 66.79 111.66
        15 117.21 117.83
        16 127.78 131.9
        18 50.71 119.39
        19 91.14 99.07
        20 104.2 104.2
        21 29.48 58.99
        22 52.53 53.03
        99 195.93 195.93

      • Here’s another one. Mother-Daughter-Grandson

        Grandson inherited virtually all of Chr 11 from grandfather and virtually all of Chr 2, 13, 14, 15 and 20 from grandmother.

        CHR Parent cM Child cM
        1 97.88 267.21
        2 253.06 253.06
        3 107.28 219.1
        4 97.97 206.75
        5 125.34 199.6
        6 124.76 189.14
        7 106.75 180.79
        8 97.43 161.76
        9 51.76 160.36
        10 76.64 176.25
        11 8.94 155.78
        12 115.35 167.39
        13 125.91 125.91
        14 111.66 111.66
        15 101.08 118.07
        16 74.43 131.9
        18 38.25 119.39
        20 104.2 104.2
        21 18.23 58.99
        22 34.39 53.03
        99 139.97 177.43

  20. Love your articles Roberta. Not trying to be dismissive but wanted to provide an illustration of why I think the idea that something special is going on with the X when it comes to recombination/crossing over is a little premature given the examples you provided.

    I just did some quick autosomal DNA comparisons on 23andme as I have both my parents and my children’s DNA to compare 3 generations. I am in the middle generation and XY.

    My son inherited all of my father’s chromosome 4, 7, 11, 15, 16 and 21. (6)
    My daughter inherited all of my father’s chromosome 4, 12 and 13 (3)

    My son inherited all of my mother’s chromosome 14 (1)
    My daughter inherited all of my mother’s chromosome 1, 8, 9, 20, 21, 22 (6). She also inherited all of my mother’s X as would be expected.

    Overall, my son shares 26.6% of his DNA with his paternal grandfather and 20.9% with his paternal grandmother. My daughter shares 23.2% of her DNA with her paternal grandfather and 26.5% with her paternal grandmother.

    In the case of my two children, to whom I gave 22 autosomal chromosomes each, no recombination occurred between my parents DNA in 16/44 (38.6%) of the autosomal chromosomes.

    I do not think the failure of the X-chromosome to recombine in a woman when it has a chance is likely to be much different than the failure of any particular autosomal chromosome to recombine.

    See user rightclik’s post above –

    “It may be an oversimplification, but I like to think of X, for the purposes of genetic genealogy, as just like another autosome — except that it isn’t passed from father to son and it can’t recombine when passed from father to daughter. There’s really nothing special about X beyond that.

    Autosomes can be as unruly as X.

    If I focused on a single autosome, e.g. Chromosome 22, I might conclude that my brother is probably unrelated to my sister and me. In the case of Chromosome 13, we look like distant relatives (only 18 cM shared). But looking across the whole genome, a more predictably typical pattern emerges.”

    • The difference is that we can use statistical inference on the autosomes and we cannot on the X, because of the inheritance pattern, and that’s without any recombination issues. I would love to see a much larger study. That would put this issue to bed one way or another.

      • Correct! But also the big difference that is hardly ever mentioned is due to the rate of recombination being much slower than autosomal chromosomes, the amount of centimorgans on the X needs to be significantly high in order to find a decent match to a recent common ancestor. Other than that, one could share an X ancestor going back many generations.

      • @Kalani

        It’s a very interesting question… How many generations back can a big segment of DNA go? Over what range does “plausible” fade into “extremely unlikely”?

        Of course, as others have noted here, the big disclaimer in this discussion is “individual results will vary.”

        I wonder if anyone has looked at the extent to which this varies from autosome to autosome. I would imagine that some of the smaller autosomes might be in some ways comparable to X in this regard.

        This is where my ignorance in genetic genealogy shines.

      • I don’t think I have seen any papers that talked about the how long a 7.7cm IBD segment might remain intact on the X. Clearly it’s going to be longer than on the other chromosomes and I think I recall reading that a 7.7cm segment might survive 14 generations in very extreme cases, IF so could perhaps a 7.7cm X segment survive 20 generations??? (guessing for now) It might be a challenge to identify which one of our potential 1,048,576 x18 great grandparents in might have originated from if it could survive 20 generations! Hopefully there will be some good research coming out on the subject in the future.

  21. Right but we can only use statistical inference on the autosomes as a whole not as individual autosomes, as my example illustrates. I agree that a larger study would be interesting and more conclusive. Surely there is some data on the likelihood of cross-over/recombination occurring each generation in humans. I’ll try to find something. It’s probably something that is variable but how much variation is the question.

  22. Pingback: Friday Finds – 01/24/14

  23. I wouldn’t be surprised if Poisson statistics can show some of the basics of the detection of X chromosome crossovers or lack thereof. However, as usual in biology, the reality is much more complicated than the model. (Ann Turner’s illustration of recombination, posted above, is extremely simplified.) In reality, there is “crossover assurance” (at least, assurance of pre-crossover structures), since these structures are REQUIRED for accurate chromosome segregation at the first meiotic division. (Based on this, one would expect to see at least one crossover at EACH female meiosis. How might this be missed in our SNP test results?) There is also crossover interference, there is partitioning of double-strand break repairs into crossovers vs noncrossovers, et cetera. There are various mechanisms for many of these effects, and it is at least conceivable that certain of these mechanisms differ for the sex chromosomes vs autosomes. See for example:
    “The choice in meiosis – defining the factors that influence crossover or non-crossover formation”
    “xnd-1 Regulates the Global Recombination Landscape in C. elegans”
    (“We show that xnd-1 [X non-disjunction factor 1] functions independently of genes required for X chromosome-specific gene silencing, revealing a novel pathway that distinguishes the X from autosomes in the germ line…”)
    “A major recombination hotspot in the XqYq pseudoautosomal region gives new insight into processing of human gene conversion events ”
    (“These findings allay early concerns that conventional COs might not occur in this human-specific PAR. …. the appearance of RAD51/DMC1 foci in the mouse PAR, which mark the DSBs that mediate pairing and successful exchange, are temporally delayed compared with the rest of the genome, possibly being predominantly induced by Spo11α rather than the conventional Spo11β isoform.”)

    I would add that there MAY conceivably be cases where two crossovers have occurred, but in our array-based SNP test results the situation looks like none has occurred, because there are too few tested SNPs between the endpoints of the crossover segment, and the error threshold for declaring a match allows a few mismatches.

    • @ RT — Yes, there’s “crossover assurance” during a stage of recombination, but the final result of four gametes includes two that maintain the composition of the original parental chromosome. Isn’t that adequate to explain the phenomenon we’re seeing?

      • Ann, I’m just learning the details myself, but apparently it is not exactly true that half of the results maintain the composition of the original parental chromosome. See (and see Figure 4)
        “The DSBR pathway commonly results in crossover, though it can sometimes result in non-crossover products… Whether recombination in the DSBR pathway results in chromosomal crossover is determined by how the double Holliday junction is cut, or “resolved”. Chromosomal crossover will occur if one Holliday junction is cut on the crossing strand and the other Holliday junction is cut on the non-crossing strand (in Figure 4, along the horizontal purple arrowheads at one Holliday junction and along the vertical orange arrowheads at the other). Alternatively, if the two Holliday junctions are cut on the crossing strands (along the horizontal purple arrowheads at both Holliday junctions in Figure 4), then chromosomes without crossover will be produced.”

        So apparently it’s not random; the resolvases control which strands of the double Holliday junction are cut, in such a way that non-crossover is reduced. I’ve had a hard time finding the ratio of COs to NCOs after DSB repair in human or mammalian meioses, but I’ve seen one estimate of ~2 to 1 in mice (ref 7 in ). This is not as extreme of an imbalance as I was envisioning based on the wording of Figure 4 in the wikipedia article.

        And “crossover assurance” apparently only requires one CO per bivalent, though I was thinking it meant one CO for each pair of homologous non-sister chromatids. So you’re right, it’s not surprising to find nonrecombinant chromosomes. Thanks for clarifying that.

      • Thanks both RT and Ann for this explanation – it addresses the question I raised earlier in the thread. Cheers, Maurice

  24. Roberta –

    What if there is a double crossover of identical segments in both parents?

    mnopbcdefgvrsfg chromatid for 1 parent
    fghubcdefgqrsty chromatid for the other parent

    mnop bcdefg vrsfg
    \/ \/
    /\ /\
    fghu bcdefg qrsty

    The likelyhood in a large population would be low that such a segment exists and would persist, but in smaller populations, it might persist with even distant inbreeding,


  25. I have been experimenting with doing one to one X matches in I am using the raw data that I up loaded from FTDNA. With gedmatch I am showing significantly more matches than I am getting with FTDNA. I have no idea why.

    • Chr Start Location End Location Centimorgans (cM) SNPs
      X 22,279,098 23,221,690 4.6 292
      X 45,532,886 47,329,929 3.4 239
      X 67,791,107 69,583,640 3.1 281
      X 87,800,614 93,364,480 4.7 397
      X 108,936,674 112,056,371 4.3 299
      X 118,418,147 119,812,040 3.8 242
      X 128,545,620 133,541,438 4.5 516
      X 141,024,884 142,296,998 3.3 277
      X 150,817,675 154,545,424 7.8 565

      Here are the results from gedmatch for one person that did not appear in my FTDNA listing.

      • GEDmatch will give you results that have less than 500 SNPs and FTDNA will not. That’s the major difference.

  26. My initial reaction to the article is leaning towards the camp that thinks we are not seeing anything different with the X that we don’t already see on the other chromosomes (outside the obvious known differences on the X). Like a few others have now I tested a set of 4 grandparents, 2 parents and 1 grandchild and as others have posted saw several instances of large chromosomes not being recombined and surviving intact. I have also seen that % inherited cm is much more swingy than I might have guessed even over all of the chromosomes never mind just looking at one of them. Here is the total cm the grandchild inherited from each grandparent in our case.

    2110cm Maternal grandfather
    1778cm Paternal grandfather
    1629cm Paternal grandmother
    1314cm Maternal grandmother

    The maths says 1700cm is the typical expected per ISOGG wiki.

  27. Pingback: Neanderthal Genome Further Defined in Contemporary Eurasians | DNAeXplained – Genetic Genealogy

  28. Pingback: Ollie Bolton Estes Robbins (1874-1955) and the Wrath of a Woman Scorned | DNAeXplained – Genetic Genealogy

        • That means they share ~25% which is c/w half siblings. Other possibilities could be Grandmother-Granddaughter; Aunt-Niece or double first cousins. I’m assuming you know the age of the person who matched and it is similar to your mother so that rules out GM-GD relationship and may make Aunt-Niece relationship less likely. If you suspect half-siblings knowing their mt-DNA may help you determine if they share the same mother or the same father. If the mt-DNA is different it has to be same father. If the mt-DNA is the same it is likely the same mother.

          • Thank you so much for that. I have the mt-DNA for the suspected half-sister but no my mum. Looks like I may need to upgrade her testing level. Thanks again.

    • Yes. You have same mt-DNA as mom. If your mt-DNA does not match your Mom’s potential half-sibling then they are related through father. Can you completely rule out double first cousins or Aunt-Niece relationship though?

      • The are 14 years apart in age and I can’t see any way that they could be related as either Aunt-niece or double first cousins. I did think that my granddad’s brother may have been the father but that would make them ordinary 1st cousins and the doesn’t seem likely now.

  29. Wow, lots of comments you have on the topic. I have another one that hope can be answered with your generational inheritance study. I have a chunk of Iberian dna (23andme) on my X chromosome. It shows on the Conservative, a bit more on the Standard and quite a big chunk on the Speculative version. My parents have two distinct ancestries: one is French Cdn with some documented Iberian, and maternal is all Irish, with one line Ulster Scots. The French side has been fully documented back to the early 1600s and the Irish back to the early 1800s. I thought I had it pegged that my Iberian dna came from my ggg grandmother, Marie Esther Rodrigues. b 1818. She’s in the pink shaded part of the X-chart and the Iberian ancestry is through her father. Her paternal gggg grandparents were born in Lisbon, ca 1616, documented. But it can’t be because while Esther’s father is in the blue-shade part, his mother is in the pink-shade part and she’s French, all the way back, documented. So this Iberian X-dna can’t come from her.

    The thing is, is that there are no Iberian ancestors in my tree for at least 7 generations (for sure) at all, let alone those that qualify for X inheritance. They’ve all been traced back to France. There might be some 10 or so generations back – hard to say because their names were changed to French. I am aware of a FamilyTree Project which is trying to prove that some French Canadians descended from Sephardic Jews. Quite a few have middle eastern haplogroups so it seems to be true.

    As far as for my maternal side – it’s all Irish and Ulster-Scots – nothing else. Unless prior to say 1800 I hade an Irish ancestor who went to Spain to serve there in the 17 or 18th centuries and married a Spaniard. I suppose it’s possible, but…

    So with the info I have, I wonder just how long this X segment of Iberian dna has been around to be inherited by me. Seems likely an impossible long time. And I’ve checked and rechecked my genealogical records. Family Tree also confirms I have some Spanish dna. Amazing.

    • Any “paper trails” data we have is suspect, because as a Jewish neighbor lady told us, “only the mother knows who the father was, and sometimes she is not sure”. In previous days, many people took in their neighbors or siblings or other relatives children and raised them as their own, after the parents died, or if the parents were too poor to raise their children. Very few records were kept of these “adoptions” until recent times. I was a bit relieved to find that a few men with the surname Ingalls do appear to match my YDNA data!

      Jay Ingalls

      • Thanks Jay. Well, that must be probably it. Funny, it never crossed my mind till now. It’s the only possibility, isn’t it? Unless this X dna can carry on a .very long way. I’ve been going far back using those charts and found a couple, 9th great great-grandparents (Abel Turcotte& Marie Girou) who might fit the bill. They fit into the blue/pink squares. Abel is Hebrew and Turcotte comes from the word Turk or Moor. They were living in LaRochelle, Fr where it is suspected many Sephardic Jews lived as Conversos. Sounds like their names are Jewish and they were living in the LaRochelle, Fr. area where it is suspected many Sephardic Jews lived as conversos. But that’s a long, long way back. I’d be more inclined to go with your assumption though. .

  30. I was trying to find some info on the x-chromosome with regard to proving sibling relationships. We think that we may have found a brother but normal DNA testing proved inconclusive and it has been recommended that we do an x-chromosome test. We know that if he is our brother, we share the same parents and grandparents as he is a full brother, he resembles my maternal grandfather so some extent…..we are all in such a state and really don’t know if we should go to the expense of the test being done in Germany.

  31. I still find this mind boggling. I am redoing my Blevins spreadsheets since they added the X.. Out of the first five that I put in 2 of us share an X chromosome but not the same segment. So my first thoughts is that this could be two different granddmothers. The one I know who it is. Elizabeth Wells Ward born about 1740 and we know her ancestors back about 100 more years. The other is the mystery. I know her line only back to about late 1700. We know that these men all share the same DNA but the “story” was there were 6 brothers. I know three sons for Elizabeth Wells Ward Blevins. I just do not know what to do with this information, This gal also matches two of the more of the five but not on X. I know that one descends from one of Elizabeth’s son, the other gal son 2 and me son 3 but we do not have this other match beyond 1800. I am beginning to think that there is a 4th son for sure. And help with a plan of attack very much appreciated.

  32. I find this very interesting and am struggling to understand it. My father, myself, and my son have all used FTDNA and recently were surprised at the X chromosome results as well as the new MyOrigins results. Previously we knew via mtDNA testing that my father is A2e, a northern Mexico/southwest US lineage most likely as he is of hispanic descent from Mexico, but naturally, as he cannot pass that on, no % of native dna showed up in anyone else’s results. MyOrigins changed that and showed him as 3% Bering group, and my son at 2%, myself at 2%. Before MyOrigins results, neither my son nor myself showed any % of native blood, none at all, which I expected. I wondered where did that 2% that wasn’t there before come from, and showing in the Bering group. My son’s father shows 100% European and my extensive genealogy back to the early US colonists, in many cases, does not show any native lineages, in spite of the occasional oral tradition. All this led me to look at X, and sure enough all three of us share 32.99 cMs in the exact same location on the X chromosome. My understanding at this point is the possibility that this represents the location of a small segment of genes passed down from our A2e ancestress, and that my father passed this segment to me (recombined or otherwise, I don’t know) and that my son received that also from me. I have no way to no if this is correct or not, but it seems to be a logical inference. And I’m also open to the fact that I could be dead wrong and there is another explanation.

    • Hi Karen,

      I’m going to be publishing another article about autosomal in the next few days. That may help you. The bottom line is that these small percentages of admixture are very fluid.

  33. Reblogged this on Study by Night and commented:
    Wow – I knew the X-Chromosome combined differently to the rest of our DNA, I didn’t realise it would be This different. Certainly casts a new light (or shadow) on the various DNA matches I have that appear to match on the X chromosome.

  34. Pingback: Lost in X – DNA that is | Family history across the seas

  35. Is a match of 33 cM on the X chromosome with only one other match on another chromosome of 14 cM reasonable? All the other matches (12 in all) are 3 cM or less. This person knows virtually nothing about his maternal background and is looking for reliable clues to follow.

  36. Do brothers usually inherit the same X from their mothers to pass on to their daughters? I have a potential first cousin or half-sibling. I am not sure which of two brothers was my father but one of them has a daughter who was tested with me and she shares an x with me and 26 % dna. I am female too. Could she still be a first cousin or does this mean she is a half-sibling.

  37. According to, several people match me for 15-20 cM on the X chromosome, but have 0 matches on any autosomal chromosomes. Is this very likely or can it be a bug? Other people match both X and autosomes.


      • Thanks for your insights. So far no common ancestor that I know of, but I have some holes in my tree. Need to finish analyzing the matching segments as you suggest. It just seemed intuitively weird–but that’s X for you I guess.

  38. Pingback: Identifying Possible Common Ancestors Utilizing Multiple Tests | DNAeXplained – Genetic Genealogy

  39. I am new to this DNA genealogy but have tried to figure out where certain medical problems come from. So this is my K I S S assumption and tell me if I am wrong. A males DNA does not change at all. What he receives from his father on the Y-chromosome is what he passes unchanged to his son. What he receives from his mother on the X-chromosome passes unchanged to his daughter. So it is only the females DNA that changes when passing to either of her children due to recombination. Is that the correct assumption?

  40. Hi! I realize I’m a little late to comment, but I could use some help. I found this blog while doing some research on idiopathic short stature. I am tall and my husband is tall, as is our son, but our 11-year-old daughter is short – way below her mid-parental height. Her pediatrician predicts that she’ll be below average in height when she’s full grown. The only person on either side of our family who was below average in height is my paternal grandfather (my father’s father) who was 5’2 (this makes sense – he was Lebanese, the oldest child of recent immigrants from Middle East, and the family was dirt poor). He would have been my daughter’s maternal great grandfather. His wife (my grandmother, my father’s mother) was German American and 5’4. Her siblings were all taller. My dad is 5’10, so he likely inherited tall genes from his mother, and I likely got those tall genes, too. My mother is tall (5’8) as is her entire family (women 5’8-6′, men 6’0-6’6) My husband’s mother and father are both tall (5’7 and 6’2). My husband’s mother’s father (my husband’s grandfather) was 5’9 and his mother’s mother was 5’6. My question is: How much genetic bearing would one below average height ancestor (my paternal grandfather) and three average height ancestors (my dad, my paternal grandmother plus my husband’s maternal grandfather) have on our daughter’s final height? When I look at this article, it seems like not much, as she would have likely inherited more of my father’s mother’s genetics through my father’s “x” and more of my husband’s mother’s genetics through my husband’s “x”, plus all the average to tall genes from my mother’s and my “x”.Thus, my daughter should be at least average in height since every other person on both sides of the family are way above average in height. Does that make sense? It makes sense in my family, as my sisters and I are all between 5’5 and 5’7. (I am the tallest, and my mom always wondered how us girls got to be relatively tall since my dad’s father was so short.) But we have no brothers, which would really help test my logic, since a brother would have inherited shortness from my dad’s “y”. We have an appointment with a specialist coming up, but I’m too fascinated by the whole subject that I can’t resist the research.

  41. Pingback: Joseph Preston Bolton (1816-1887), Twice Excluded Baptist Deacon, 52 Ancestors #38 | DNAeXplained – Genetic Genealogy

  42. Pingback: Utilizing Ancient DNA at GedMatch | DNAeXplained – Genetic Genealogy

  43. Pingback: Anzick (12,707-12,556), Ancient One, 52 Ancestors #42 | DNAeXplained – Genetic Genealogy

Leave a Reply