Homozygous blocks in women in Ben Moscia's spreadsheet

[JimT]

If a woman's parents have recent X chromosome ancestors in common, I would expect her X chromosomes to have more and larger homozygous blocks than a woman who had one parent who is, for example, 100% western European, and the other who is 100% east Asian.  Consider these two women in Ben Moscia's spreadsheet of 23andMe results:

ID021
- 6.5% of her SNPs are in homozygous blocks of 50 or more SNPs.
- She has 11,554,374 base pairs in homozygous blocks of 50 or more SNPs.
- She has no homozygous blocks of 100 or more SNPs.


ID030
- 11.9% of her SNPs are in homozygous blocks of 50 or more SNPs.
- She has 20,462,262 base pairs in homozygous blocks of 50 or more SNPs.
- She has 4 homozygous blocks of 100 or more SNPs including 1 block of 233 SNPs.

This suggests that ID030's parents have more recent common X chromosome ancestors than ID021's parents.

Jim Turner

[Seán MacGorman Powell]

If a woman's parents have recent X chromosome ancestors in common, I would expect her X chromosomes to have more and larger homozygous blocks than a woman who had one parent who is, for example, 100% western European, and the other who is 100% east Asian.  Consider these two women in Ben Moscia's spreadsheet of 23andMe results:

ID021
- 6.5% of her SNPs are in homozygous blocks of 50 or more SNPs.
- She has 11,554,374 base pairs in homozygous blocks of 50 or more SNPs.
- She has no homozygous blocks of 100 or more SNPs.


ID030
- 11.9% of her SNPs are in homozygous blocks of 50 or more SNPs.
- She has 20,462,262 base pairs in homozygous blocks of 50 or more SNPs.
- She has 4 homozygous blocks of 100 or more SNPs including 1 block of 233 SNPs.

This suggests that ID030's parents have more recent common X chromosome ancestors than ID021's parents.

Jim Turner

I've never thought of it that way, Jim, but it makes sense. 

One other possibility is that perhaps some populations simply have inherently less X-DNA variability than others, for whatever reason, and therefore people from certain populations are more likely to have large (and numerous) homozygous blocks.  I could be way off base on that idea, but does that sound feasible?

[JimT]

One other possibility is that perhaps some populations simply have inherently less X-DNA variability than others, for whatever reason, and therefore people from certain populations are more likely to have large (and numerous) homozygous blocks.  I could be way off base on that idea, but does that sound feasible?

I think that it is likely that some populations have had founder effects that reduced their X-DNA variability.  Within such a population, I think that a women with more large homozygous blocks probably has more recent common X ancestors than a woman with fewer large homozygous blocks.

[Seán MacGorman Powell]

I think that it is likely that some populations have had founder effects that reduced their X-DNA variability.  Within such a population, I think that a women with more large homozygous blocks probably has more recent common X ancestors than a woman with fewer large homozygous blocks.

That seems reasonable to me. 

The problem then becomes specifying the population from which any two given testees are being compared, to differentiate between population-based causes for large/numerous homozygous blocks, versus time-based differences (and the two are not necessarily mutually exclusive).

In other words, let's say that two people have a higher-than-average number and size of homozygous blocks.  How do we differentiate between the cause being (1) that they both came from a population with relatively low X-DNA variability, versus (2) a recent common ancestor? 

If an East Asian and a European have a lot of shared homozygous blocks, does that mean the same thing (as far as time to their most recent shared X-DNA ancestor is concerned) as if a European compared with an African have the same number/size of homozygous blocks...or on a finer scale, if a Portugese person is being compared with an Irishman, compared to two other kinds of Europeans?

The other thing that occurs to me is that it could be important to consider is not just how many (or how large) the shared homozygous blocks are, but where they are located on the chromosome.

I don't know the answer to these questions, they're just food for thought.  You could well be onto something, Jim, and it certainly deserves further consideration.

[JimT]

In other words, let's say that two people have a higher-than-average number and size of homozygous blocks.  How do we differentiate between the cause being (1) that they both came from a population with relatively low X-DNA variability, versus (2) a recent common ancestor? 

By context.  One should always look at DNA test results in the context of geographic origin, ethnicity, and genealogy.  We currently have little information about whether some populations have lower X-DNA variability than others, but we will someday.  I know of one published study that looked at one X haploblock, and found that it was more variable in Africa than elsewhere.  They did not report substantially different variability in that haploblock within the populations of other regions, but we really need a lot more data on that subject to get a good picture of X variability within populations.

Because of the pattern of X chromosome inheritance, X chromosome haploblocks are less likely to suffer from founder effects and bottle necks than Y chromosomes.  I would expect to see fewer differences in X-DNA variability between populations than we have seen in Y-DNA variability, but time will tell.

If an East Asian and a European have a lot of shared homozygous blocks, does that mean the same thing (as far as time to their most recent shared X-DNA ancestor is concerned) as if a European compared with an African have the same number/size of homozygous blocks...or on a finer scale, if a Portugese person is being compared with an Irishman, compared to two other kinds of Europeans?

The concept of "time to their most recent shared X-DNA ancestor" seems problematic to me.  When comparing two people's shared haplobocks, different haploblocks will have different most recent common ancestors who lived at different times. 

When comparing a haploblock between two people, we might be able to get a general idea of when the most recent common ancestor lived by looking at STR differences within that haploblock, just as we do with Y chromosomes today.  If a European and an African had matching haploblocks in their 23andMe results, I would expect to see more STR differences in that haploblock than if two Europeans had the same matching haploblock.

The other thing that occurs to me is that it could be important to consider is not just how many (or how large) the shared homozygous blocks are, but where they are located on the chromosome.

Absolutely.  For example, a large haploblock that spans the centromere has different implications than another haploblock of the same size that doesn't span the centromere.

[Seán MacGorman Powell]

The concept of "time to their most recent shared X-DNA ancestor" seems problematic to me.  When comparing two people's shared haplobocks, different haploblocks will have different most recent common ancestors who lived at different times. 

Yes, you're of course quite correct, and that was a poor choice of words on my part (in fact, it didn't quite sound right to me when I wrote it!).  Perhaps a better phrasing would be "time to most recent X-haploblock ancestor" (TMRHA?).

One thing is apparent... this is a hell of a lot more complicated than Y-DNA and mtDNA ancestry time estimates.