Retesting or resampling requested - sisters in two haplos

[clintonslayton76]

A result for one woman #1 who tested her sister came back in D haplo for full sequenceMT, and now her sister comes back H. Each have full sister atDNA share and all uncles and nephews show ranges that strengthen that they are full sisters within the nexus of the same parents. Mother is deceased, but each tested woman has daughters
This has generated a request from #1 to review or retest, before spending any more money with FTDNA (such as to test the daughters).
Of course it is not impossible that they had different mothers, but it is unthinkable for the two of them.

[MoberlyDrake]

If their atDNA results prove they are full sisters, but their mtDNA results show different haplogroups, contact FTDNA and ask them why. I don't know what their current policy is, but in the past, if the error was their fault, the retesting was free,

[clintonslayton76]

They were contacted, and are looking into it. I am so happy that #1 & 2 did atDNA because this sort of problem with only yDNA or mtDNA results can be troubling for the family. I will keep this posted with whatever the subject feels comfortable with sharing after (I hope) FTDNA retests the sample. Since there are two people involved, it will be interesting to see how many samples have to be retested and if the daughters will have to be tested.

[clintonslayton76]

Because there is no Status report for resampling, we are in the dark about what is going on. The Customer Service representative stated that they must have used "one of my wife's samples" which had a problem which means nothing, since my wife's first sampling from atDNA (years ago) is the only sample FTDNA had ever rec'd. A test kit was provided, but without any response from FTDNA since this was sent in early February. This is in the nature of a Progress Report that can indicate no progress, since the normal Pending status tool does not appear to apply to resampling. We are not happy campers, since there is NO feedback of any kind.

[dna]

@clintonslayton76, if you consider how many avenues FTDNA have to test, retest, investigate, etc. it is clear that the final results might take quite some time to arrive.

On the other hand, essentially zero feedback is quite unnerving. I can only assure you that, like with a details of a major surgery, almost nobody wants to know all the details, since that makes us, mere mortals, even more unsettled...


Mr. W.


P.S.
Some feedback, without revealing internal procedures or know-how, could surely be arranged by FTDNA. For almost any issue that requires an investigation, I can envision a possible communication process like

• an initial e-mail to the customer explaining that such an investigation has say three stages, and that each stage can take weeks, months, years (whatever appropriate) to complete, and reassuring that the investigation has already started;
• a followup e-mail that the first stage is ended, and the second one has started;
• another followup e-mail that the second stage has ended, and the third one has started;
• the final e-mail that the last stage has completed.

If any of the stages has a clear progress line, another e-mail could be sent in the middle, etc.

[dna]

@clintonslayton76, when you look at their mtDNA results, how many differences can you see?


Mr. W.

[clintonslayton76]

A different haplotype! One sister is D (we think WRONG!) and one sister is H (we think accurate). It is highly improbable for full sisters (sharing over 2200cM of atDNA with one another) to have different mtDNA haplos.

[KATM]

It does seem that FTDNA made some type of error in the mtDNA testing, since these two sisters match as such autosomally.

But, just in case there could possibly be something else (unthinkable to them) going on:
The DNA Detectives Autosomal Statistics Chart shows, for Full Sibling (Group A), a range of 2300-3900 cM shared, with the average at FTDNA being 2650. It shows 1300-2300 cM shared (Group B on the chart) for Half Siblings, Aunt/Uncle/Niece/Nephew, Double First Cousin, or Grandparent/Grandchild. Certainly there can be outliers who are still related as full siblings, who share just under the minimum.

For another source of autosomal statistics and relationship ranges, Blaine Bettinger's Shared cM Project page has some charts as well. You can also use an interactive version of it at the DNA Painter site, to view possible relationships.

While waiting for more information from FTDNA, have you tried uploading the sisters' autosomal files to GEDmatch, to compare them One-to-one and check the Fully Identical Regions (FIR) they share? The ISOGG page for Fully Identical Regions explains what amount of FIR to expect between full siblings and other relationships:

The expected percentages for full siblings are 50% half-identical, 25% completely identical, and 25% of regions with no shared autosomal DNA for an overall average of 50%. Half-siblings do not in general share any fully identical regions. Double cousins who are related through both their paternal and maternal lines also share some fully identical regions. It is also possible that small FIRs will be shared by more distantly related individuals whose parents are all from the same highly endogamous (inbred) population.

[clintonslayton76]

Only one sister is on GEDMatch and the other has not given permission to do so, only one sister is a genealogist, the other has no doubts regardless of FTDNA findings that they are from the same mother. I am an onlooker, but the autosomal results are quite clear without going to the effort: 2724 shared cM with a block of 249 (and an X match). One shares nearly ~1700 cM and the other 1975 cM with an uncle who was brother to their mother.

[clintonslayton76]

Non-progress report:
Still waiting for FTDNA to give result, no feedback of any kind in the meantime. 2 months since resampling.
At least two other Projects are confused by two sisters with unmatched mtDNA, and one admin removed it from view.

[clintonslayton76]

FTDNA announces May 13 that these sisters do match mtDNA haplos, so all of the projects should now show them both as H, instead of one sister as D and one as H. I would prefer to see some contrition for this mistake in the form of some refunds, because now we have to wonder how many customers have had samples switched. Okay, it is not as bad as going home from the hospital with the wrong baby, but such a fundamental failure could affect research where not-so-attentive customers have wrong results for this narrow, deep form of DNA (as opposed to the broad but shallow autosomal).

This was really upsetting for the months before the second sister was tested, because all of the mito matches led sister #1 to the totally wrong continental migration theory, admittedly not genealogical level information but very confusing looking at the list of haplomates. How would they have known had they not both also done autosomal? This leads to the possibility that any siblings testing only for Y or mito could have their world turned upside down by something like this, and these are not inexpensive tests.

[clintonslayton76]

FTDNA announces May 13 that these sisters do match mtDNA haplos...

Wrong!
The last posting was premature. Having now seen the communication, it seems that FTDNA agrees that they should match mtDNA haplos, and a batch number has just been assigned. Sorry about the premature announcement, but most of what I posted holds true, except that the posting of an amended haplo (assuming that there is one) will not be until at least June. This is the first feedback these sisters have received since mid-February.

[clintonslayton76]

Finally, the mtDNA is posting showing these full sisters as mitochondrial matches. Had one not been a genealogist who understands something about mtDNA, and had they not done autosomal DNA supporting their full-sibling relationship, the initially reported mito mismatch last year could have remained an upsetting mystery.
Let this be a warning to all: lab mistakes happen!

[Cats]

Thanks for sharing this and the final update.

[chrisg]

I had asked them about testing my mother, to look into her paternal grandmother's side. This might also apply here. They told me, her XX would be indistinguishable. My impression based on that, is a higher margin of error when testing a female's lineage.