Random (another DNA PSA)

One of the crazy things about DNA is the truly random nature of genetic inheritance. A person inherits roughly 50% of their DNA from each parent, but only for identical siblings is it the same 50% from each. By the time this distribution reaches the next generation or the one after that, differences are multiplied even more. This confuses the heck out of people -- how can Samantha, which whom I share 60 centiMorgans of DNA, and Belinda, with whom I share 320 centiMorgans of DNA, both be second cousins to me and to each other, when our "numbers" vary so widely? The answer, of course, is that Samantha and Belinda and I each got a varying amount of DNA from our shared great-grandparents. Lots of people use the very scientific Gummy Bear diagram to promote a more intuitive understanding of this concept, but I think Vegetable Soup makes more sense.

You'll notice that the corn has gotten mighty scarce in our bottom mixture. Yet another mix from same two "parent" ingredients might have more corn and less peas. Nobody is ALL corn or peas or carrots or potatoes to start with, either. If you're picky about your vegetable soup and carefully remove all the lima beans before consuming (who, me?), you know that some bowls will have a bunch while another bowl has only one or two. The random nature of just which bits of your parent's DNA make up the 50% you get is similarly random.

This explains why my brother and I have matches to different people, and even different "ethnicity" estimates. This table shows our estimated (i.e. "we haven't totally figured out how this ethnicity thing works yet, but here's a general idea") origins according to Ancestry.com. His ladle of DNA has different ingredients than mine; apparently, his is loaded with extra French.

Keep in mind that there are several other DNA database sites -- 23andme, FamilyTreeDNA, MyHeritage, Geneanet, LivingDNA, GEDMatch and I'm probably missing some. You can only match someone if they either test with or upload to the same database. Most sites offer their own ethnicity analysis, too.

In terms of the total number of matches, everybody's numbers will be very different. It's not uncommon for someone to have hundreds of thousands of matches, or to have tens of thousands of close matches. And it's equally common for someone to have far fewer in each category. It's all just a matter of who exists and who has tested. Ancestry.com has the biggest database, for sure, and the most productive for those of us whose ancestors have been in North America for a bunch of generations. For someone with more recent European ancestors, one of the other databases may have better results. And for those with South American or Asian ancestry... there aren't very big databases out there just yet.

Here are my numbers on Ancestry:

And here's the breakdown for Patrick's:

So out of the exact same database of people, Patrick matches roughly seven thousand more! His random bits match their random bits more than their random bits match my random bits.

We can look at those differences a bit more closely. The next table shows Patrick's and my top 15 matches in order. There are four people who make both lists, albeit in different places on each. None of our matches are closer than 2nd cousins one removed -- we share the same great-grandparents with them. Since our parents were only children and our grandparents had few siblings who also had few children, it's not at all unexpected. I've never actually met any of the people on the lists.

I know the colors in that table don't exactly harmonize; they are the colors I use to code my matches. The idea is that all the matches who are also descended from ancestors that they shared with a particular set of my great-grandparents will have the same color applied. Warm colors are Mom's side, cool colors are Dad's side.

Now, of Pat's 44k and my 37k matches, I haven't divided all of them into match groups. Without a shared ancestor in their family tree or shared matches already in a group, they just go into an "I don't know yet" bucket. These are usually 20 cm or lower matches, indicating our shared ancestors are likely to be 6 generations back or more. But, you never know, sometimes a 6 or 7 cm match will obviously fit into one of the groups, and they may be the key to solving a puzzle!

Based on the total number of matches that I have divvied up for each of us, this table reflects the proportion that fit into each group. Severin St. Peter and Maria Robert (navy blue) were French Canadian, so there's Patrick's bumped up French ethnicity.

Why so many of some and so few of others? Well, it could be that descendants of those "few" lines haven't tested, or they haven't tested with Ancestry. However, there are certainly descendants of some of the "many" who haven't tested, either. In general, it seems to be about fertility, especially four, five and six generations ago.

On the other hand, among those 2% of Barker and Moffatt matches, and the 3% of the Thomas and Strowbridge matches, there were people who were the key to finding mystery ancestors. If those one or two people hadn't tested with the same company as me... it would all still be a mystery. It's kind of miraculous, in a way.

Do some ancestors want to be found, or the Universe want us to find them? Or it is just random? I guess I'll take luck wherever I can find it.

No lima beans, though!