Do x

[tomcat]

Integrate X chromosome SNP's into FF analyses. Have been told the X SNP's have been seq'd but are not analyzed. Seems wasteful.

And/Or do a stand-alone X chip with an even larger number of X SNP's and its own matching database and forum section.

[Colin]

I agree, it would be good to have both of these, as they narrow down the possible lines to research

[tomcat]

It is a perfect addition given FTDNA's leadership in Y and Mt and considering the perennial difficulties researching female lines in our patriarchal society it would be a killer app.

[rdegnen]

Using the X Chromosome

Making use of the X Chromosome for the FamilyFinder and PopulationFinder would be great. DeCodeMe uses it for the ancestry analysis.

[nathanm]

According to Cece Moore's account of the administrators' conference, "Bennett Greenspan mentioned that FTDNA is working on developing an X chromosome browser." So it will happen, but I wouldn't get my expectations up about its capabilities. Matches on the X chromosome should limit the lines you can be related on, but only if you're related on a single line.

Plus, estimating the degree of relationship for X chromosome matches will be much more difficult. Autosomal DNA is fairly straightforward--the amount we inherit from ancestors essentially halves each generation. It varies somewhat, but actual inherited amounts won't be too far above or below the average. However, X chromosomes inherited from a long line of maternal ancestors have significantly different shared amounts than those inherited from alternating generations of male and female ancestors. Even for autosomal matches, the more distant the estimate of our relationship, the larger the margin of error. It's quite possible estimated 5th cousins are really 10th or greater cousins. Estimates for X chromosome matches will be even less reliable, if they even make estimates.

Regarding a stand-alone chip just for X chromosomes, forget about it. FTDNA is a business after all.

[mkdexter]

X can not be predicted accurately because it deals with only one chromosome and not an average of 22 and includes a non-uniform inheritance pattern from the parents.

A good example of this is where-

my son shares:
father X, 0% and 1-22, 100%
mother X, 100% and 1-22, 100%
paternal grandfather X, 0% and 1-22, 45%
paternal grandmother X, 0% and 1-22, 55%
maternal grandfather X, 0% and 1-22, 54%
maternal grandmother X, 100% and 1-22, 46%

my daughter shares:
father X, 100% and 1-22, 100%
mother X, 100% and 1-22, 100%
paternal grandfather X, 0% and 1-22, 42%
paternal grandmother X, 100% and 1-22, 58%
maternal grandfather X, 85% and 1-22, 52%
maternal grandmother X, 15% and 1-22, 48%

Notice the X being a chromosome by itself and having a non-standard inheritance is no where close to 50% per grandparent.

The chromosome 1-22 shares all hover much nearer a 50% average mark give or take a few % per grandparent. Remember this is not 50% inherited (because that would really be 25%) but it is 50% shared because 50% of the tested SNPs match test to test.

About the only use an X comparison would do in FF is to tell people when they have a maternal connection. In the X the amount actually shared could be any amount. Carry this down a few more generations and see what a mess it would be to figure out. The 100% for my son to my wife's mother happens to be 100% in his daughter to my wife's mother. Not one ounce of a change after 3 generations.

In the autosomes at least you have an idea that 100% of the SNPs shared means parent/child and 50% of the SNPs shared means something like grandparents and 22% of the SNPs shared pretty much means something like cousins.

Matt.

Matt.

[dwight]

this is very interesting, and helpful. one comment though: when you write that "[a]bout the only use an X comparison would do in FF is to tell people when they have a maternal connection" it almost sounds like a bad thing! at most, a glass half-empty. but i would counter that it is most definitely a glass half-full! to know that a match is one your maternal side is a lot. it's a certainty often hard to come by. so at the very least, how 'bout a little column for x-ch match; without regard to the size of the match, just knowing there is a match on the x would give us a very valuable piece of information.

X can not be predicted accurately because it deals with only one chromosome and not an average of 22 and includes a non-uniform inheritance pattern from the parents.

A good example of this is where-

my son shares:
father X, 0% and 1-22, 100%
mother X, 100% and 1-22, 100%
paternal grandfather X, 0% and 1-22, 45%
paternal grandmother X, 0% and 1-22, 55%
maternal grandfather X, 0% and 1-22, 54%
maternal grandmother X, 100% and 1-22, 46%

my daughter shares:
father X, 100% and 1-22, 100%
mother X, 100% and 1-22, 100%
paternal grandfather X, 0% and 1-22, 42%
paternal grandmother X, 100% and 1-22, 58%
maternal grandfather X, 85% and 1-22, 52%
maternal grandmother X, 15% and 1-22, 48%

Notice the X being a chromosome by itself and having a non-standard inheritance is no where close to 50% per grandparent.

The chromosome 1-22 shares all hover much nearer a 50% average mark give or take a few % per grandparent. Remember this is not 50% inherited (because that would really be 25%) but it is 50% shared because 50% of the tested SNPs match test to test.

About the only use an X comparison would do in FF is to tell people when they have a maternal connection. In the X the amount actually shared could be any amount. Carry this down a few more generations and see what a mess it would be to figure out. The 100% for my son to my wife's mother happens to be 100% in his daughter to my wife's mother. Not one ounce of a change after 3 generations.

In the autosomes at least you have an idea that 100% of the SNPs shared means parent/child and 50% of the SNPs shared means something like grandparents and 22% of the SNPs shared pretty much means something like cousins.

Matt.

Matt.

[vinnie]

Thanks, Matt for this visual and the ones you posted in the FF thread. But I agree with Dwight and others that knowing whether or not there's an X match is helpful. My father's maternal grandmother was adopted, so identifying his X matches helps to narrow down which persons I plan to contact in this regard.

[mkdexter]

Vinnie and Dwight. Yes I agree with you that data showing how much is shared (such as in the chromosome browser) would be nice. I do not think a prediction based on X would work - but some type of indication that X is shared would be nice yes. Maybe it could be part of the new Advanced Matching filter.

[dwight]

Advanced Matching is really cool - and i agree, the 'x' factor would be a nice (simple) addition to that report.

Vinnie and Dwight. Yes I agree with you that data showing how much is shared (such as in the chromosome browser) would be nice. I do not think a prediction based on X would work - but some type of indication that X is shared would be nice yes. Maybe it could be part of the new Advanced Matching filter.

[k.o.gran]

I too think that X matching could be another good clue when trying to find the connection with existing FF matches.

-Kai

[thetick]

I agree with everything Matt stated, but an X match could just be listed as possible cousin. Those of us familiar with the details would be thrilled to have that information!! A simple link to a FAQ for those not as knowledgeable could explain why X matches are listed without a range and are only possible cousins.