6

Using Gedmatch's file diagnostics, an autosomal DNA test kit I manage shows that there are 31200 (4.4%) no-calls. In most matching algorithms, a no-call is treated equivalently to a match. Therefore, as Gedmatch states:

This kit has an unusually high number of no-calls, which usually results in a larger number of false matches.

Which would explain why there are so many matches for this kit with (likely) no actual genetic relationship.

This kit was tested through FamilyTreeDNA.

Is this truly an unusual margin of error? Is it unreasonable to ask FamilyTreeDNA to retest on a new DNA sample?

Improve this question
5

I have 27 Family Tree DNA raw data files.

They have between 624 (0.1%) and 23,488 (3.4%) no calls. Their average is 8,537 (1.2%).

So yes, your 31,200 (4.4%) no calls likely can be considered to be on the high side.

According to a Family Tree DNA forum post from 2015, FTDNA may re-test you if the no call rate is 3% or more. So it may be worthwhile for you to contact them.

Improve this answer
5
  • 1
    A quick update: I contacted FamilyTreeDNA and they stated that this kit "is well within our quality thresholds". They also add that "If your raw data analysis is being performed by Gedmatch, it is likely including autosomal SNPs not tested by Family Tree DNA." Not sure what to think of this.
    – Harry V. ♦
    Jan 21 '18 at 19:14
  • @Harry - I'm very surprised they said that your 4.4% no calls were within their quality thresholds. GEDmatch doesn't say what they do to compare results between companies, but it's no different than what Family Tree DNA itself does when they accept raw data from other companies. FTDNA thinks their comparisons with other company's raw data is okay. So I don't know why they would have reason to believe that GEDmatch's comparisons are not okay. You might yourself want to try comparing the segment matches that FTDNA and GEDmatch give you for the same 2 people to see what might be different.
    – lkessler
    Jan 21 '18 at 22:04
  • 1
    I've now checked the raw data downloaded myself and come up with a very similar number to Gedmatch (4.45%), so I don't think untested SNPs play into this at all. Checking chromosome by chromosome, some are worse than others (X worst at 5.6%, chr 19 and 22 at 5.4%). I've checked one of my matches (single segment match of 16cM on chr 22), and in the matching segment over 15% of "matching" SNPs in my kit were no-calls. I'm going to follow up with FTDNA to try to find out what their quality thresholds are.
    – Harry V. ♦
    Jan 21 '18 at 22:22
  • FamilyTreeDNA: "The specifics of our call rates are proprietary." I don't think they're going to be more forthcoming than that.
    – Harry V. ♦
    Jan 25 '18 at 18:07
  • @Harry - That's too bad. Then, if I were you, I'd pay for a retest at FTDNA, and likely it will result in fewer no calls. You'd could then create a new raw data file that merges the two. Any reads the same are good. Any reads where one file has a no call can use the read from the other file. Any reads with both different should be rare (maybe 50 per file) and are misreads and you don't know which is correct, so you could make that a no call. Then take the new file and upload that to GEDmatch for better results.
    – lkessler
    Jan 25 '18 at 19:13
1

I recently contacted FTDNA about this:

"No-calls = 30053 = 4.2313865735526 percent.

This kit has an unusually high number of no-calls, which usually results in a larger number of false matches."

Their response:

Our internal autosomal call rate thresholds are proprietary. However, the autosomal raw data for kit MKxxxxx has been confirmed to be well within our quality control thresholds.

Unfortunately, we do not know what criteria Gedmatch is using for their analysis. Gedmatch may be able provide a more detailed explanation of their analysis.

Improve this answer
1
  • Sean, thanks for your answer. This is an almost verbatim reply that I received to my query when I asked FTDNA directly, and is not, I think, a very satisfactory reply.
    – Harry V. ♦
    Jan 27 '18 at 3:13
1

I agree that the FTDNA response is surprising. I have 12 FTDNA kits uploaded to Gedmatch that vary between 1.5-2.3% no calls, but there is one kit that is 4.4% and gets the Gedmatch caution message of an unusually high rate of no calls. If the same SNPs are being tested by FTDNA and there is no problem with the other FTDNA kits, why should just the one kit be so outstandingly different? I also will follow this up with FTDNA.

Improve this answer
0

I believe everyone who uploads their FTDNA raw data will have similar No-call results. I spoke to FTDNA; she checked my results in their end & everything looked good. Plenty of matches.

Gedmatch says in their eye predictor they cannot accurately predict eye color from FTDNA because of a lack of SNPs that FTDNA scrubs due to not giving out health portions of your DNA. I have to also assume that the admixture with Gedmatch is not giving a complete picture if they don't have all the SNPs.

Gedmatch shows my no-calls at 31616 and 4.4% seems to be very similar to what others are reporting.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.