View Full Version : Chromosome 23 Admixture Comparisons - Possible?

09-10-2016, 06:39 PM
Ok, this is the first question I've posed to the forum since joining.

I've recently connected with four autosomal tested relatives of the 3C/4C range that helps me connect with a suspected Cherokee Indian 3GGM of mine. Family lore and much evidence strongly suggests she was either half or full blood Cherokee Indian. GEDMatch has identified a nice 15cM overlap on Chromosome 23 on nine relatives that could only have come from her due to the nature of the descent.

I wanted to compare this overlapping segment in the Admixture analysis to see how it might show up in Chromosome painting or others. Does this common segment correlate with other American Indian segments?

However, when I went to perform the check, Chromosome 23 isn't available. What am I missing? I think I'm finally close to another data point to support or not the heritage. Options? Suggestions?


C J Wyatt III
09-11-2016, 05:17 AM
The simple answer is that due to the quirky nature of X-DNA matching, the X-chromosome is really not suitable for admixture calculation. There is a bigger problem, but I really don't want to go into that now.


09-12-2017, 04:11 AM
The "the quirky nature of X-DNA matching", as C.J. Wyatt III puts it, lies simply in the different inheritance pattern of the X-chromosome. Otherwise there is no difference between matching on an autosome and matching on an X chromosome.

And just how does the inheritance pattern differ? The X chromosome can only be transmitted from father to daughter or from mother to a child or either sex; it can never be transmitted from father to son. This is obvious, perhaps, but what may not be so obvious is how drastically this manner of inheritance reduces the number of ancestors who may have contributed to the X chromosome.

For example, since I'm male I can instantly rule out my father and all of his ancestors of both sexes as being a contributor to my X chromosome*. I can also instantly rule out my maternal grandfather's father and all of his ancestors of both sexes. And I can do this for my maternal grandmother's maternal grandfather and his ancestors; and for my maternal grandmother's maternal grandmother's maternal grandfather and his ancestors; and so on.

Consider the effect that this has on, say, my 6th great grandparents. Potentially, I could have 256 of these. I have somewhat fewer in actuality, due to a certain amount of pedigree collapse. But what's the maximum number of these ancestors who could have made any contribution to my X chromosome? Believe it or not, it's only 34.

In addition, I can eliminate some of these as possible contributors to specific segments on my X chromosome. How? Well, it turns out that only about 20 cM of my X chromosome surrounding the centromere came from my mother's mother. The rest was from my mother's father.

This means that unless a segment is within that 20 cM surrounding the centromere, I can eliminate 21 of the 34 as potential contributors to my X chromosome, leaving just 13. So out of the 182 cM length of the X chromosome (as calculated for 23andMe), 13 6th great grandparents are all who could have contributed to 162 cM of it.

Potentially, 21 could have contributed to the 20 cM, but in reality it's likely that most of the 21 have contributed no X chromosome DNA to me at all.

But let me address the question of your thread, and show why it just isn't necessarily so that "the X-chromosome is really not suitable for admixture calculation". Since you're interested in Native American ancestry, let me tell you about mine.

I have two Native American ancestry through both of my mother's parents. The most recent fully-Native American ancestor of my grandfather is not documented. However, some of his descendants applied for inclusion on the Dawes Roll. They were rejected, but on technical grounds; and it seems that many of the ancestor's Y-line descendants actually do have a Native American haplogroup. I don't, but I'm not in the all-male line of descent.

My grandmother's ancestor, however, is documented -- though only inferentially. An ancestor named Magdalaine Pany Baudreau, who lived in the 18th century, was identified in her marriage record as "the natural daughter of Jean Baptiste Baudreau and 'an Indian'". Magdalaine's mother is not named in the document, but Jean Baptiste also had a son named Jean Baptiste Baudreau II. JBB I was not married to JBB II's mother when his son was born, but he did marry her later "to legitimize" his son. We know the name of JBB II's mother: Suzanne. If the two siblings had the same mother, then, my most recent Native American ancestor was Suzanne; otherwise, it was simply an unnamed Native American woman with whom JBB I had Magdalaine.

So what does all of this have to do with the use of the X chromosome for ethnicity calculations? Well, Native American DNA -- or at least the "Ancestry Informative Markers" that are used -- is fairly distinctive. It can sometimes be confused with Asian ancestry, but it is less likely to be confused with European ancestry.

As it happens, my five siblings and I all have a small amount of Native American DNA. 23andMe estimates around 2% for each of us, in the form of several segments. Individually, the siblings each have about 6-8 or so segments, but collectively we inherited perhaps 12 different segments from our mother -- none are from our father. And in some cases, I've been able to identify which maternal grandparent contributed the segment.

But since we're concerned with the X chromosome here, only one of us inherited a Native American segment on the X chromosome: my youngest sister. It's actually easy to see why she could have inherited such a segment when none of the rest of us did, because it's located on a part of her maternal X chromosome that she shares with no other sibling.

It's also clear that she inherited it from our maternal grandmother. Why? Because two siblings inherited a maternal X chromosome containing DNA only from our maternal grandfather. That means that they inherited an unrecombined copy of our mother's paternal X chromosome. I match both of these siblings except in an approximately 20 cM region surrounding the centromere. That, I inherited from our grandmother.

Now, recall that I said that no more than 34 of my siblings' and my 6th great grandparents could have contributed anything to our X chromosomes. 21 of these are ancestral to our grandmother, and just 13 are ancestral to our grandfather. My sister's Native American segment, if it's real at all, obviously must have come from among the 21 possible X-chromosome ancestors of our grandmother.

So, let's take a closer look at these. My grandmother's father was the son of two immigrants. His mother, though, is the only one from whom he could inherit his X chromosome. She was from Alsace-Lorraine. I think it's fairly unlikely that 23andMe's software would confuse Alsatian and Native American. However, it also happens that my sister shares a large segment that includes her Native American segment with a half 2nd cousin of ours.

This half 2nd cousin is descended from a maternal half sibling of our grandmother, meaning that the segment must have come from our grandmother's mother and not her father. (The two are not related.) So, we can trace it to a specific great grandmother. That great grandmother's father was from the Mediterranean island of Minorca, and his ancestors lived on the same island for many generations. I would also not expect this ancestry to be confused with Native American, but in any case GEDmatch shows that my sister shares the larger segment that includes her NA segment with relatives who are not descendants of my grandmother's maternal grandfather.

So this leaves our grandmother's maternal grandmother, Marie Eulalie Ryan. As it happens, this 2nd great grandmother's parents were 2nd cousins to each other. Both are descended -- by X chromosome paths -- from Magdalaine Baudreau. Eulalie's father Pierre Ryan would have had only 3 ancestors in Magdalaine's generation (including Magdalaine) who could have contributed anything at all to his X chromosome. Eulalie's mother Marie-Josephe Ladner would have had only 5 (again including Magdalaine).

That means the odds of a Native American segment being passed on the X chromosome to my sister from Magdalaine are actually higher than for any other 6th great grandparent. Of course, Magdalaine herself would actually have had two X chromosomes, one of which would have been fully Native American, and the other of which would have presumably been French.

But the bottom line is this: however much someone may want to disparage the usefulness of the X-chromosome in genealogy, they're misguided. There are no other chromosomes to which a maximum of 34 6th great grandparents might contribute if you're male, or 55 of you're female.

What's more, it can be fairly easy to eliminate some of these for specific segments, based on the relatives with whom you share these segments. Pedigree collapse might have an effect, but only to the extent that it involves a ancestor or ancestors who have multiple paths to you which are all X chromosome paths.

Just because an ancestor has more than one path to you and one is an X path does not mean that the other path (or paths) are also X paths. Any path which includes two males in a row which be one by which no X chromosome DNA can be transmitted.


*As many people know, there are actually two regions in which the X and Y chromosomes are known to recombine. These regions are known as PAR1 and PAR2 (where "PAR" stands for "pseudoautosomal recombining region), and they're the reason why men's DNA data files show some SNPs at the beginning and end of the X chromosome which contain two values. At these positions, some of the SNPs are actually on the Y chromosome, but since it's generally impossible to know which value belongs with the X and which belongs with the Y, both are reported on the X. But these regions are very short. They would never appear as shared segments at any of the testing companies, unless as part of a much longer segment.

Here is a very helpful reference on the X chromosome inheritance pattern: http://thegeneticgenealogist.com/2008/12/21/unlocking-the-genealogical-secrets-of-the-x-chromosome/