Y-STR Value Changed Without Notice

[Charles]

This is actually part of a larger problem which I have written about in a separate thread about my Y-DNA67 test.

In summary, in January (one month after the results for markers 1-60 were posted) the results for markers 61-67 were finally posted and my test was reflected as being completed. However, I noticed that marker 63 (DYS487) was blank. Since then, I have been trying to resolve that problem. Yesterday, the panel for markers 61-67 was removed and replaced with “Awaiting Lab Results.” The status page indicates an estimated completion date of next week. Once I receive the final report, I will update the other thread and outline what happened.

While reviewing my records, I have now been able to confirm my suspicion that another Y-STR marker value had been changed without my receiving any notice or warning.

In January a few days after the last panel (61-67) had been posted, I noticed that my match list at 37 markers had changed. Presumably because of the blank value the test at the 67 marker level would fail (even though I have three matches who have taken the 67 test) and fall back to the 37 marker level. Since I had saved a copy of the December match list at 37 markers, I was able to compare the lists. One match had been moved up from the 25 marker match and the relationships had been recalculated and reshuffled. Except for one match, everyone was now being shown as more closely related to me.

Based on my understanding, the only way that could happen was if something had changed in my 1-37 markers. In January, I was not able to find a copy of my original Y-STR values, but yesterday I found a file where I had saved the results. Marker 27 (GATA-H4) had been originally shown with a value of 10 but is now reflecting a value of 11.

On the one hand this is good news (and explains the change in matches) because I now match on that marker everyone in my surname project subgroup.

However, it also points out a quality control issue with FTDNA. Users should be aware that Y-STR values are changed without notice being given to the user.

[rivergirl]

This happened to one of my project members back in January.
GATA H4 changed from 10 to 11.

He rang FTDNA was told that there were some IT issues with a few markers, one being GATA H4, and that the current numbers were correct.
This also affected his GD with others at 67 markers, as in the GD didn't change for his matches at 67 to reflect his new result.

No other explanation was given, no notification was given.
If the project member hadn't picked it up, we may have missed it altogether.

[Charles]

Thank you for your response.

Does this mean that changes made to Y-STR values after a user has joined a project are not automatically communicated to the project?

[RandyZie]

This really bothers me, I'm not sure If I can realistically trust my matches. I have one panel left on my test (48-60), but can I trust the previous panels? I show zero matches for 37, and twenty two for 25 markers. With DYS578=6, which after googling and looking at some projects seems wrong.

[dna]

That should not be happening.

However, I will offer my guess on how possibly one could deal with STR results and present an additional option I am using.

After user rivergirl reported the issue with GATA H4, Darren from FTDNA announced that results for all the panels would be posted together = no partial results. That made me think that Quality Control is being done after all the testing has completed. And it is then when the value for GATA H4 is possibly adjusted. So, if anybody is asking, I am now suggesting waiting for results from all the panels to be returned and looking like a correct one haplogroup prediction available. That in my eyes means QC was completed (at least I hope so).

GATA H4 is a marker that might need to be adjusted if the lab equipment is reporting it in a different standard than the one used by FTDNA. I have no idea about FTDNA lab, but I can easily imagine that different models were bought over the years as the lab capacity kept expanding. And that not all of them report in the same standard. (Yes, it appears that there is more than one standard...)

I have somewhat semi-scientific approach to my genealogy research, so if possible I am trying to test in pairs: siblings, cousins, parents/children etc. That doubles the cost, but makes us (the family genealogists) really sure that the results we are looking at are the real ones. Thus we do not ask ourselves or the forum whether such results are possible, but devote time to inventing explanations we can present to the family .

W. (Mr.)

[Ann Turner]

GATA H4 is a marker that might need to be adjusted if the lab equipment is reporting it in a different standard than the one used by FTDNA.

That sounds like a plausible explanation. GATA H4 is one marker where FTDNA doesn't follow the NIST standard nomenclature. This table from SMGF shows you need to add one to the NIST value (it's the very last marker on the list):

http://www.smgf.org/ychromosome/marker_standards.jspx

[Darren]

If anyone has any concerns if they don't think there matches are correct please send me a private message. I had seen with the GATA H4 issue but that looks to be corrected. I agree though, when anything changes with results an email notification needs to be sent. I am sorry that it was not in this case.

-Darren
Family Tree DNA

[rivergirl]

That should not be happening.

However, I will offer my guess on how possibly one could deal with STR results and present an additional option I am using.

After user rivergirl reported the issue with GATA H4, Darren from FTDNA announced that results for all the panels would be posted together = no partial results. That made me think that Quality Control is being done after all the testing has completed. And it is then when the value for GATA H4 is possibly adjusted. So, if anybody is asking, I am now suggesting waiting for results from all the panels to be returned and looking like a correct one haplogroup prediction available. That in my eyes means QC was completed (at least I hope so).

GATA H4 is a marker that might need to be adjusted if the lab equipment is reporting it in a different standard than the one used by FTDNA. I have no idea about FTDNA lab, but I can easily imagine that different models were bought over the years as the lab capacity kept expanding. And that not all of them report in the same standard. (Yes, it appears that there is more than one standard...)

I have somewhat semi-scientific approach to my genealogy research, so if possible I am trying to test in pairs: siblings, cousins, parents/children etc. That doubles the cost, but makes us (the family genealogists) really sure that the results we are looking at are the real ones. Thus we do not ask ourselves or the forum whether such results are possible, but devote time to inventing explanations we can present to the family .

W. (Mr.)

This particular project member ordered a 37 marker test and received his results in mid Dec 2014.
We looked at his results, in particular his GATA H4 result at this time, as it was a result that separated him from a close match.
In Jan 2015 he ordered an upgrade to 67 markers.
It was after the 38-67 results came in that he noticed his GATA H4 result had changed.
How long after a test result is in do you think we should wait until we check the results? One or two months maybe?
At what stage can we assume that FTDNA have done their quality control?
Why do FTDNA think it is not necessary to inform the tester and/ or Project Admins that results have changed?

How lucky you are W (Mr.) to be able to test in pairs, not all of us are so fortunate.

[dna]

That should not be happening.

Like in another today thread wrong data posted in Family Finder results.

QC rules are simple, product is not released until QC is completed.

In my opinion, waiting for the results is stressful by itself. Getting wrong results, is like getting hurt.

W. (Mr.)

P.S.
If somebody says that Quality Control needed Quality Control...

[QuantumBiker]

I submitted a bug for y-str dys437, dys438 which appear to have been swapped. Have known y-dna match that is not “matching” because of this. Yfull showed “correct” values. Found 7 others that did not maych y-full, so having real doubts over accuracy of STR vs raw data...