Strand of DNA with title Are small DNA segments real matches

How many small DNA segments are legitimate?

If you’ve looked at Family Tree DNA data, then you’ve seen that some of your matches have some small segments that are less than seven centimorgans. Ever wonder how many of these small segments could actually be matches?

The Small DNA Segment Debate

There’s some debate on them.  Some people see them and they think that well we can use these to help find others. However, a lot of the research on small segments shows that most of the small segments are not real. They are DNA that could have been passed down from 10 or 20 or 30 generations ago.

From a genetic genealogy standpoint, you are not going to be able to show from in family tree or with other records. In other words, it’s not really good as a piece of evidence.

There are also a lot of small segments, particularly the smaller they get, that are just common amongst big populations particularly in endogamous populations. You don’t have to have endogenous ancestors to have small segments that are nearly ubiquitous in that population.

When looking at Family Tree DNA data,  we’ll look at the segments. How many of those small segments should we be able to trust as maybe even being actual matches?

Analyzing Small Segment DNA for Reliability

To begin with, I have both of one set of my grandparents tested. In doing that, I know exactly which of these large segments of DNA that I have that are shared between me and that grandparent. 

Because of the way recombination works, these are only going to be great big segments, particularly because it was through my father. This means there’s even less recombination so these I will have great big large segments. I don’t see any small segments. 

In fact, in my DNA that I share with my grandmother, there’s only one segment that is less than 40 centimorgans. Its 36 centimorgans. There’s nothing in the range of five, six, ten, or even 20 centimorgans. 

This is perfect for looking at small segments. Why the first thing I need is a grandparent that I match with that. I have that data so then I can define what segments I know for sure are from that grandparent.

The second thing I need is some other relative, such as first, second, third, or fourth cousins, that have a common ancestor with that grandparent in some way. Then I can look at the segment data from each one of those matches and overlap it on my grandmother’s data.  If those small segments are outside of those big segments, then they couldn’t have been inherited through that ancestor. If they are within it, then they could be but it’s not definite.

How many of those small segments are actual matches? - A Segment of DNA
Watch this video on YouTube.

Analyzing the Small Segments of DNA of a First Cousin Twice Removed

This first set of data is from a first cousin twice removed. I can mark all of the shared segments that are bigger than seven centimorgans and they also fall within where I share DNA with my grandmother. All of the other segments are segments that are smaller than seven centimorgans. You can color code these small segments. Red if they fall outside of what you share with the common ancestor and yellow if they are small but you and this match share the segments with your common ancestor.

In this example, I share 14 segments with the first cousin twice removed, but four of them are impossible segments. My match and I couldn’t have inherited these segments through my grandmother because I don’t share them with my grandma. 

In short, 28% of the segments I share with this first cousin twice removed does not point to a common ancestor.

Analyzing the Small Segments of DNA of a Second Cousin Once Removed

With a second cousin once removed, I share 104 centimorgans with them. When I take out the small segments (less than 7 shared centimorgans) it’s only 74 centimorgans. This amount falls within the range of where would expect a second cousin once-removed to be you.  

We share 11 total segments that are small segments, with three being impossible segments. Of the amount of shared DNA, 27% would be from the common ancestor. Compare that to the first cousin twice removed, and you’ll see they are relatively the same.

If I stopped comparing matches here, then you might start to think that about a quarter of these segments are impossible segments (segments not from the same common ancestor). 

Analyzing the Small Segments of DNA of a Fourth Cousin Once Removed

I have a fourth cousin once-removed, which represents a distant relative. According to Family Tree DNA, overall I share 49 centimorgans with them. When I remove all those small segments, I only share 28 centimorgans with this individual. Those centimorgans only appear on one segment.

I share nine additional segments of DNA with them, but they are all small segments. Four of those nine segments are impossible segments which again means I don’t share DNA in those areas with my grandmother so I couldn’t have inherited that DNA from the same person. 

I should point out, that I have constructed extensive family trees for me and this match and there isn’t a second path of relationship between us. 

In short, 44% the small segments of fourth cousins once removed are impossible segments. 

Analyzing the Small Segments of DNA of a Fifth Cousin

For my fifth cousin match on FamilyTree DNA, I share 52 centimorgans initially but once you remove all those small segments it’s only 11 centimorgans. Other researchers would define many of these 11 cMs as small segments. 

I share 17 small segments with this person and 9 of them are impossible segments.

Thus, just over 52% of all the segments are impossible segments. 

How Useful are Small DNA Segments

This analysis helps me to draw a couple of conclusions. 

  1. The number of small segments doesn’t appear to line up with how distantly related you are. The number of small segments is really a random number.
  2. The percentage of impossible segments appears to be going up the more distantly related you are. You share less and less DNA the more distantly related you are and so you’d think that more of these small segments might end up being impossible segments. 

There might be a lot of different things that people can do with small segments but if all you share with a match is a  small segment, then they’re not going to be very useful for you.

Should You Research Small DNA Segments?

If you’re just beginning in genetic genealogy, then completely ignore the small segments.  For an experienced genealogist, you may want to start tackling some of these small segments to see whether they are real matches.

If you never get to researching your small small segments in your genetic genealogy research, then you’re probably gonna do fine because there’s plenty of matches out there that share these big segments with you that you can find matches for and you can find relationships.

Learn why all of your small segment DNA matches might not be actually real realtives. #genealogy #dna #dnamatches
Share this post to Pinterest if you found this article helpful.

Author, Engineer, Educator. Andy Lee is the DNA guru of Family History Fanatics. He is also your knowledgeable but casual genealogy researcher.

One Comment

  • moira wylie

    Hello Andy, I have had my DNA done through I have been to a DNA day organized through a genealogy branch. I’m afraid I am still lost and bewildered. Hard to involve people as it takes up time. Question is I want to connect to my father’s mother somehow. She has no birth date and was born in Tasmania but died in New Zealand. She had three children who are all dead. One child had two children. (me and my sister). The others did not have children. I know this for sure as I knew them. Most of the high segment names that I recognize are connected to my fathers family, which of course branches off in two directions. (his mother and his father). I don’t know how to connect to his mother (the one I’m interested in). Still with me? I have no first cousins but have one sister and two children. My DNA is mostly 3rd cousins up. only about 3 people are 1st and 2nd cousin and they seem to be related to my father and a couple I asked don’t know a connection to my father’s mother. Have you any ideas Andy. Thanks

Share your thoughts on this post

%d bloggers like this: