When is DNA testing too much of a good thing?

Question

The Winter holiday is a prime time for feasting and over indulgence.  A little chocolate is nice but....

It is great when a DNA test breaks down a brick wall for one person but more DNA testing can be like too much chocolate.  Take for instance when the same person tests with AncestryDNA, 23andMe, and Family Tree DNA's Family Finder and uploads all three results to GEDmatch.  It is great to be able to compare with relatives in each lab's customer database, but it causes indigestion in GEDmatch for the tester and any cousins wanting to do triangulation.  

Too much of a good thing also happens when a mother orders mtDNA tests for both herself and her child.  Unless she suspects her baby was switched at birth, the outcome is going to undoubtably be a waste of money.  Those funds would have been much better spent testing a cousin who would have revealed a different ancestral branch.

Y-DNA testing both fathers and sons is a similar story and any surprises are bound to wreck the holiday spirit.

Comments

You made me smile on the last line. I'm waiting for my Dad's Y-DNA results as we speak. As a female, it's a bit difficult to produce that particular DNA myself!

I agree with saving it and spending it on cousin testing. I'm considering getting my second or third cousin (not sure where he falls) tested next year as I can afford it, because between him, my Dad and my 5th cousin in England, we'd have more area covered.

I got some good tips from one of the other members here though regarding genetic testing and what tests to avoid if you're actually looking at tracing the paternal line. Maternal line would be next to impossible, I suspect, since last names often change and the male surname has traditionally been carried on. It's changing now, but then we're getting hyphenations and a mash-up of names.

Answer 1

Totally agree Peter that there is no point in testing mother/daughter, father/son  (or even close cousins) for mtDNA or Y-DNA respectiveiy, unless there is indeed some doubt.
(The same comment does not apply to autosomal DNA though, the more the merrier there as they all add pices to the jigsaw)

Re testing all three companies.
I find it invaluable to catch those who have only tested at eg ancestry, and will  probably never upload their file to GEDMatch where actual comparisons can be made to get the details we've paid for but ancestry simply does not provide.
For all the "innovative" trees, they are all still great hints, helped a little more recently by ancestry now telling us how much dna is shared.

Re uploading more than one version of your file to GEDMatch from multiple companies, there is a facility to still do that if you wish to find the very small, differences between them and NOT clog up the works at GEDMatch.

Go into the link Edit or Delete your DNA Resource Profiles
Select the duplicate and click Edit
Change from Public to Research and save

But for the majority of testers, if you have uploaded kits from more than one company, or ended up with more than one copy of the same kit because you think your upload didn't work, simply pick the kit number you wish to continue to use and select delete for all the others using the above Edit/Delete process.

Comments

Thanks for that tip, Lorna. I uploaded first my FTDNA, later my 23andMe results, to Gedmatch, thinking that i was doing myself and others a favour. Now, afer reading Peter's - and others' - posts, it seems not. So i intend to do as you suggest and turn my 23andMe autosomal results at Gedmatch into ''research''. However, i am very glad that tested with both FTDNA and 23andME. I would gladly test with Ancestry as well - even though i consider Ancestry to be the least useful of the three - for the sake of the matches that i might find there. Sadly, Ancestry does not send its kits to Europe.

---

My pleasure Albertus, glad it helped.

If only those who have uploaded the same kit from the same company multiple times and really really do clog up the works would realise it is really easy to get rid of the duplicates....

Re Ancestry not testing in Europe - when I tested, they didn't ship outside America - so I used one of the Freight Forwarders listed on the ISOGG Wiki. Worked fine.

---

Thank you Lorna! This is exactly what I had in mind. I feel silly for having missed it.

I've just gone through and cleaned up all the accounts I manage.

Answer 2

Thank you Peter. Given the prices for various DNA tests, this is good info to know.

Answer 3

Thanks for pointing out that people shouldn't upload their duplicate tests to Gedmatch! There seems to be a widespread misconception that different companies will give us different results, but that's not true for the autosomal DNA that Gedmatch works with.

Your comments about not testing children for mtDNA cause me to observe that people of have different reasons for doing DNA testing, some of which justify testing several generations of the family. Many 23andMe customers of my acquaintance have bought tests (which include mtDNA) for their children as well as themselves because they are curious about inheritance patterns in the family (for example, how much genetic material do the siblings share?) or they are interested in finding out whether a parent's carrier status for certain health conditions was passed on to the kids.

Comments

Thanks Ellen,

It makes sense to have all relatives take an autosomal test.  For mtDNA I was trying to discourage repeatedly testing close relatives of the same direct maternal line (mother/child, siblings, etc).

---

For the reason that Peter states, i bought the mtDNA test only for myself, and not for my maternal brothers, two nephews, a niece, and maternal cousin. The nephew's and niece's mtDNA does not interest me, whereas my brothers and cousin have the exact same mtDNA as I do. The mtDNA test is quite costly, and my results will have to do duty for all of us! By saving here, i have been able to buy many autosomal test kits.

---

I did the mtDNA Full Sequence and when asked advised my cousins and aunt that their test would be redundant.  When pushed for what else we could do with my aunt (then 104 years old), I decided to test her just in case I had a new mutation.  I did!  Actually I had two mutations that were heteroplasmic.  That caused me to be a distance of 2 from people who were perfect match with my aunt.  This condition is not rare, but it is not frequently encountered.

An explanation from the esteemed Ann Turner:  Your additional mutations have special codes to show they are "heteroplasmic." When a mutation occurs in one mtDNA molecule, other mtDNA molecules in the cell retain the original version. This mixture is called heteroplasmy. When the mtDNA molecule with the mutation is copied, the levels in the cell go up. If the mutation occurs in an egg cell, it can be passed on to the next generation.  It can only be detected when it reaches a certain level (about 20% or so), so it's even possible that your aunt has some heteroplasmy at a lower level. In that case, the mutations would have originated with your grandmother or even earlier.

So it may not be a total waste to test another member of your female line.

---

Jane, Thanks for the informative post dealing with heteroplasmy.

I have found a similar occurrence within my maternal line except only with one heteroplasmic mutation.

I tested my mtDNA Full sequence in 2007 at which time I found I had two apparently unusual differences.  One diff was heteroplasmic  in HVR1 and the other diff was found in HVR2  for haplogroup H1a (maternal Norwegian origin).

With these two unusual changes on board I have never seen anything closer than a difference of 3 even though H1a is a pretty common Haplogroup

Fast forward to 2016 when my my male 2nd cousin tested his FMS mtDNA and we found he did not carry the heteroplasmy marker but was identical for the other unusual HVR2 difference.  Consequently he shows as a one diff to me -  so the mutation appears to have taken place when my gr-grandmother was conceived as his gr-grandmother (next older sister) evidently did not get the mutation to pass on.  For all practical purposes we are exact mtDNA matches.

 Sort of makes the expenditure of testing $ feel worthwhile.  And yes, I also used the good services of Ann Turner.

Bill Harvey

---

Very interesting to hear about significant mtDNA changes between generations. I guess this is a valid reason to supplement basic haplogroup tests with more refined tests.

Answer 4

When it comes to GEDMatch, you can, and in fact likely will, see a few differences between your results from the "big 3" companies. For example I have uploaded all 3 and see around 20 or so differences between the 3. These are the result of errors in the testing. Even though the error rate is extremely low, when you're testing for half a million SNPs, a few errors creep in. Running 1 to 1 comparisons between your results from two different companies can help reveal which of your SNPs is potentially in error.

There are even tools available that can help you narrow down those bad SNPs and create a hybrid file that is likely to be more accurate than any of the individual results.

I haven't tried it yet, as I'm on Linux only, but for the Windows folks here there's a program called "DNA Error Fix" which claims to do just what I was describing. I'd be interested to know if anyone tries this out and gets it to work. :)

Ideally for GEDMatch there would be a setting to choose one default company result to use for matches and triangulations etc, but still have your other two listed for you personally, or for direct one to one matches etc. Invisible unless specifically selected, like your phased results currently are on GEDmatch. That way you could avoid having it look like you had several clones on the site. ;)

As for testing siblings, I had my sister tested for the reason some others here have mentioned. I had both my parents tested to help with phasing; determining which DNA came from which parent to help determine both which chunks of DNA came from which parent, and to show which side of the family previously unknown DNA relatives are related to you on. And also simply because I wanted to see how different our patterns of inheritance were.

Beyond that I plan to be a bit more careful with my testing. The only additional test I will likely run on my father will be the full sequence mitochondrial test so that I have a record of his mother's line, which is shared with one of his sisters, her daughter, her daughter, and her daughter now. And of course other earlier female ancestors from that line.

I will also likely buy the basic autosomal tests on the other two of the "big 3" for the aforementioned purpose of phasing and allowing me to better see which side of the family new DNA relatives fall on.

Beyond that I'll be on the hunt for somewhat distant Y DNA cousins next I think, to help flesh out our particular branch(es) of the haplogroup tree. :)

So, in general; good advice Peter. :) Testing mother and daughter usually is a waste of money, unless you know what you're doing and have a good reason.

For fathers and sons, generally the same thing. Don't do it unless you really know what you're doing. And the current wisdom is to test the son, not the father, as there's a chance the son might have a SNP or two the father doesn't have, and thus represents the most up to date tip of the branch going forward, which ideally you'll be able to compare against more ancestral parts of the tree by finding more distant genetic cousins through Y DNA tests like the STR tests or SNP tests like the "Big Y" at FTDNA, or the "Y Elite" at FGC.

Answer 5

Maybe advices to be displayed in Wikitree lists to help people select the next candidate for testing...

http://www.wikitree.com/treewidget/S%C3%A4lg%C3%B6-1/899 



Red text should give an advice of good or less good test takers...
 

Comments

Good idea.

Answer 6

From personal experience I can say that my sister's auDNA has matches to people who I do not. Moreover, she has some of my father's X-chromosome, where I do not. So in the case of a brother and sister, I can say there are distinct advantages to testing both. Even with siblings of the same sex, they may inherit different sections of auDNA and match different people.

Ditto for my first and second cousins. We share many matches, but some much more than others. Triangulating on our matches allows us to indentify ancestral origins for specific sections of our DNA. I would always recommend testing older generations if they are available and willing. I would not test a male cousin if my uncle is available.

X-Chromosome matches and mtDNA data can also be important for cousins. The way these elements are inherited should be considered when evaluating subjects for testing. If my father is dead, but my mother is alive, I would test both my mother and sister for X-chromosome purposes. (The same cannot be said for testing father and son.)

I tested my daughter mainly for the purpose of Health Markers, not genealogy. However, from a standpoint of curiosity, we have enjoyed examining our chromosome matches on Gedmatch.com where we phased our results. We can see exactly which portions of her auDNA (more than half) inherited from me.

auDNA is not as predictable as many people imagine. I have learned to make no assumptions about who I or other family members might match. For those with more a passing interest in DNA matching, more testing can be helpful, or at least it has been for me.