Are we really related to all those DNA matches?

Updated: Jan 19

So, let's get down to it. How does this DNA matching work?

First, a centiMorgan is a unit of measurement used to determine the distance from scientifically indentified positions, or markers, on each of our 23 chromosomes. It is not a physical measurement, but a tool to show closely we are related to the matches we receive in our DNA test results. It is used by all the testing companies.

DNA Inheritance Patterns

Each of us receives exactly 50 percent of our DNA from each parent. A mother’s DNA is passed through the ovum, of course, and dad’s is passed through sperm. Each ovum and each sperm has a different mix of that parent’s DNA, which is called “recombination.” Recombination is totally random.

That means that while I share both parents with my two sisters, and we each get the same amount of DNA from both of them, we will each inherit different combinations of Mom's and Dad's chromosome segments. The only time this doesn’t happen is with identical siblings – twins, triplets, etc. They share 100 percent of their DNA because each of those individuals grew from the same egg/sperm combination.

Fraternal twins or triplets come from multiple eggs fertilized by different sperm. Therefore, though they were conceived and born at the same time, they would still share the same amounts of DNA as siblings born years apart.

To illustrate this concept, below is a graph that shows the different amounts of DNA that my sisters and I share with known matches at AncestryDNA. (Click the image to see a larger version.) It also shows how AncestryDNA's predictions fare against the actual known relationships of these matches to my sisters and me:

As you can see, my sisters and I share varying amounts of DNA because of the differences in how our parents' DNA was recombined in each of the eggs and sperm that created us.

It isn't a huge difference, but it's significant to see that there will be differences.

The next person, Lisa, is Patti's daughter. She matches her mother at roughly twice the amount of DNA as Paula and I share with her, as her aunts.

As you look down the "Matches" column, you'll see relationships of increasing distance.

First is Melissa, who shares with us about the same number of cMs of DNA as Larry. Both are predicted to be first cousins. That's true in Larry's case, but it is not true for Melissa. Because we also have used traditional genealogical research methods, we know she is actually our paternal half-niece. Her mother, now deceased, was a half-sister my sisters and I didn't know about until DNA came along. This proves that DNA testing is just one more tool and is not going to miraculously give you all the answers you might be hoping for.

As relationships become farther apart, note how the shared cMs and the number of matching segments decrease. Also, note how AncestryDNA's predictions become a bit wobbly. Why?

Again, that's because of recombination. I inherited more DNA segments (or longer segments) to Steven and Janet than my sisters inherited.

So, I match Steven at 218 cMs across 10 segments. Paula matches him at 172 cMs over 12 segments, which shows that I share longer segments with Steven than she does. Patti, however, only shares 61 cMs over 4 segments, which makes it appear that they are 4th cousins, which we know is not the case.

The same with Janet: I share 102 cMs with her, but my sisters share about half of that with her. I know from my research that she is our 2nd cousin once removed, but Ancestry's prediction is that we are either 3rd or 4th cousins.

So, are we really related to all of our matches?

The answer is – it depends. We can count on those matches that fall into the 4th-cousin category as being related to us. When you get to Distant Cousin matches – 5th cousins or more – shared segments decrease to as little as 7 cMs. At this point, you might be looking at "false positives," which is why when I look at DNA, I disregard matches that are less than that amount because they aren't reliable.

Plus, they are difficult to prove on paper. Keep in mind:

  • 1st cousins share one set of grandparents

  • 2nd cousins share one of four sets of great-grandparents

  • 3rd cousins share one of eight sets of great-great-grandparents

  • 4th cousins share one of 16 sets of great-great-great-grandparents

  • 5th cousins share one of 32 sets of great-great-great-great-grandparents.

Even if you have a super-accurate tree that goes back to all 64 of your 4th great-grandparents, which would be very rare, it is still difficult to find those connections. To do so would require building forward the trees of every descendant of every one of those 4th-great-grandparent couples. (Just thinking about that makes my head hurt!)

Happy testing!

Coming up next: A Case Study – Using DNA to determine unknown parentage.