Can Autosomal DNA solve this puzzle?

[AndyMick]

A friend of mine has asked me for help/advice. I've had YDNA and mtDMA tests but I've no experience of Autosomal atDNA. He is aware of family rumours that his father may not be his biological father and seeks to clarify the rumours. He has not been able to find any male paternal cousins who would be able to do a YDNA test. So he is considering atDNA testing with one of his niblings (his sister's children). They would share any DNA from his mother (their grandmother) and would either share or not share DNA from his father. Now on average we inherit 25% of our DNA from our grandparents, so he should see either 50% match or 25% match depending whether his biological father is the same as his sisters. But 25% is only an average. Would it be a clear-cut result, or is he likely to get a result of something like a 37% match, which would be very confusing? Any help/advice welcome.

 

[peter]

I'm afraid there is no simple answer - there are some statistics on the range of results but to me they look highly suspect (see this ISOGG page for more).

I would suggest taking the test and worrying about the significance of the answers later - Family Finder tests are a lot cheaper than Y-DNA tests anyway, and there's a good chance that the answer will be obvious; if not testing another niece/nephew should help to clarify things.

Please note that even if his father is biological ancestor on average he'll only share 25% of his DNA with his niece or nephew, or 12.5% if the rumours turn out to be true. Although on average we inherit 25% of our DNA from each of our grandparents, precisely which 25% it is will differ.

 

[Alexander Bisset]

It's important to bear in mind that whilst we get roughly 50% of DNA from father and 50% from mother. It's essentially millions of coin tosses. A brother or sister with EXACTLY the same parents could theoretically share zero DNA with a sibling, or theoretically they could share 100%. However on average it will be around 50% of the atDNA will be shared.

So it is NOT the case he should see 50% match or 25% match it is "up-to" 50% and "up-to" 25% with the averages being a lot less than that. This is simply because it is possible for sibling A to have inherited a strand of DNA from father and for same strand for sibling B to have inherited from their mother. Thus for that strand you'd get a zero % match even though they have the same parents. Similarly they could have different fathers and same mother but sibling A & sibling B both inherited a strand of DNA from the mother therefore for that strand there is a 100% match.

It is only because there are millions of these strands that the chances tend to even out and the averages settle down. This is where I was slightly uncomfortable with the way Peter had explained things in his newsletter as whilst a lot of what was said was right the maths was wrong as those were absolute upper limits and the averages would be considerably less in most cases. Especially the more generations back you go.

 

[AndyMick]

Thanks both for your replies. Food for thought!

 

[Tim]

It's definitely worth taking the Family Finder test. This is where I've had the best results from.

 

[peter]

This is where I was slightly uncomfortable with the way Peter had explained things in his newsletter as whilst a lot of what was said was right the maths was wrong as those were absolute upper limits and the averages would be considerably less in most cases.

Any DNA statistics I quote are averages (means), not upper limits. Whilst the actual figures will vary, they'll tend to be clustered around the mean.

An autosomal test looks at markers across (almost) the entire genome. The chance that two full siblings will share 100% of their DNA is infinitesimally small (unless they are identical twins, i.e. from the same egg). Similarly the chance that they share 0% is negligible. Put it another way, you're more likely to be struck by lightning on the way to post your DNA sample than you are to get such an extreme result.

One proviso: if your parents are related to each other, the shared percentages will come out higher than expected. Similarly higher than expected shared DNA between distant cousins tends to indicate that there are hidden connections.

Autosomal DNA tests have incredible potential - it was only after testing myself, investigating resources such as GEDmatch, and reading books and blogs on the topic that I finally realised just how great the potential is.