Friday, July 7, 2017

Why Siblings Are Not The Same Genetic Ethnicity - Simplified

My brother in law took an ancestry dna test after Christmas, and then my husband took one later this year.  My nephew (very reasonably) asked if he could add his dad's percentages to his uncles percentages, and have the result tell him HIS ethnicity.

You really would assume it would work that way.  But the answer is no.  Not only is it no, but I explained to him that if him and his sisters took the test, they could get very different results.  (In this particular case, he does have an identical twin. Their results should be the same.  Our twins, his cousins, are NOT identical, and would likely not have the same results)

That's because although we inherit 50% of our dna from each of our parents, we do not necesarily inherit the SAME 50%  that our siblings inherited.  

Think of it as an estate.  If our farm is split 50/50 between two of our children, one child may take a goat, a cow, and the china.  Another child may take 5 goats, a chicken, and a bedroom set.  Equal value, but different items.  

This is a great article from Ancestry, where 4 sisters took a DNA test, to see "who is more Irish".  It helps explain this concept a bit better than my farm inheritance analogy, but without being too in depth or overly technical.  :-)

There's another great article here, with visuals (colored beads), explaining how DNA can be inherited differently by siblings who have the same parents:

Oh, one more thing to keep in mind.  The estimates on ancestry are just that - estimates.  They are based on the averages, and those estimates keep evolving.   


  1. Great article. I appreciated your "50/50" explanation. I like to use card deck analogies, where each parent generates their own 52-card deck for each child they produce ---but the two decks get "drawn" one card each at a time but only one card is kept and the other one is discarded. Eventually that produces a new 52-card deck, where the number of contributed cards from each parent may be 26-26 or 24-28 or even 20-32 for Dad/Mom respectively. (As a result, the 20/32 child may look more like the relatives on Mom's side of the family.)

    As to the "more Irish" idea, I've been wondering just how accurate are the presuppositions behind these types of commercial DNA tests. For example, is it really possible to know from such a test that particularly DNA markers are "Irish" rather than "Scottish" or even "Norse/Viking"? Yes, some markers are more common among people today in particular geographic areas. Yet, there have been so many centuries of migration, invasion, intermarriage, and a general mixing of populations to where I can't see how such evaluations through DNA tests are anything more than very general tendencies among modern populations. Are descendents of the Viking communities on the Irish or English coasts significantly less "Norse" than the Scottish mercenaries which were recruited to Sweden many centuries ago?

    As a result, I find it very questionable that one person being 22% Irish (allegedly) and another being 38% Irish are at genuinely statistically significant differences. Moreover, what does it mean to be "Irish"? Various ethnic groups lived on that island at various times. Are we talking about the Gaelic peoples who migrated from Ireland to Scotland in the 5th century? Or are we talking about Picts who migrated to Ireland after the last Ice Age? What about the Scots-Irish, a term usually applied to Scots who were recruited to the plantations of Northern Ireland not so many centuries ago. Some of them eventually drifted further south and intermarried with what some what insist upon calling "the true Irish." Are they "less Irish" even if their ancestors have lived there in Ireland for many centuries?

    DNA tests can be very useful for tracing one's heritage, especially when coupled with surname data from Y-DNA matches of short genetic distances. They can certainly distinguish Bantu from Brit and Middle Eastern from Central American aboriginal persons. But when distinguishing an "Anglo-Saxon Brit" from a German from Saxony, and a Scottish from an Irish person, things get extremely dicey.

    That's why the Ancestry DNA article about the sisters was interesting but raised my hesitations. I hope the company explains to potential clients what such percentage numbers mean and don't mean.

    From what I've heard, the same tests and comparisons run by different labs or even by the same lab at a later date, can produce radically different results. I wonder if they are being overhyped and oversold. I'd love to see what results the four sisters would get at another lab.

  2. After writing my post (above), I discovered another webpage on your website making many of the very same points. I appreciated the details and some very informative links. I'm glad to see that I'm not the only one who is a bit frustrated with some of the advertisements I see on TV.