Pages may contain affiliate links.
See my disclosure page for more details.

Pitfalls Using Shared Matches or ICW Lists

Shared matches, or your "ICW" (in common with) list, are supposed to be an easy way to get clues, or even answers, from your DNA results. There's one pitfall everyone can have and several you may discover if your family tree doesn't branch like it should.

These problems won't put you in a pit, exactly, they'll lead you down the wrong path. That path can go on and on, wasting your time and leaving you frustrated and confused. Learn what the pitfalls are so you can spot them.

This post is the second part of a three-post series. The first post explains why you may be having confusion using your shared matches and what a "knotted" tree is. Read "Confused By Your Shared Matches?" to learn more.

These posts also provide more in-depth explanation of what I call a family "knot" (instead of a family tree). I have a free webinar at FamilyTreeWebinars you can check out. Watch "MyHeritageDNA for Your Family Knot."

Pitfalls Using Shared Matches


The Problem Using Shared Matches

This post is about the problems with your shared matches. I'll talk about the solutions in a different post because this topic takes a good bit of explanation.

I want to look at the pitfalls of using shared matches if your tree is overly interrelated (knotted). If you don't know what a knotted family tree is, you need to read the previous post because it's important to know if you have this situation.

Understanding what a knotted tree is and the pitfalls of using your shared matches will help you use and not misuse your shared match list.

Summary: For interrelated populations, you and a match will have more than one shared ancestor in your family tree. Often you only have one shared ancestor in your DNA tree. This may also be a situation where you have several shared ancestors in your DNA tree but it is fewer than the shared ancesotrs in your family tree.

Like anyone else, you have to figure out which ancestor the shared DNA came from. However, unlike some people, you will be misled because you share multiple ancestors. The more interrelated the population, you may even share the same unexpected mix of ancestors with multiple matches even though you do not have DNA from all of those ancestors.

Time for some definitions.

What's the difference between a family tree and a DNA tree?

Your family tree consists of all your ancestors.

Realize the issue is not that you've misidentified your biological ancestors in your family tree. That obviously makes a difference with using DNA but what I'm discussing applies for someone who has perfectly identified their biological ancestors.

Your DNA tree consists only of the biological ancestors from whom you have identifiable DNA. This has to do with how autosomal DNA works and I'm not going to discuss that in this post. If you don't understand how autosomal DNA is inherited, you need to learn, especially if you have cousin marriages or more intermarriage than a "normal" person. Not understanding how autosomal DNA is inherited will eventually cause you more confusion.

What's important right now is, you will not have autosomal DNA from every ancestor and you will have too little from some ancestors for it to be helpful.

This is also why it's so important to think about the test taker, not you. You might have no DNA or too little from a particular ancestor but your sibling or cousin might have some useful DNA from that ancestor (even though they are the same generation as you).

When you are talking about a different generation, it's even more likely there's a difference.

My great-aunt has a lot of matches I don't have. All her ancestors (her family tree) are my ancestors (I have ancestors she doesn't but in our shared branch, my family tree is identical to her whole family tree). Our DNA trees are very different.

Notice there's an inverse relationship. My FAMILY tree has more ancestors than hers due to the additional generations creating additional branches. My DNA tree has fewer ancestors for our shared branch (whether I numerically have fewer or more ancestors total I can't know and it doesn't matter). This is obvious if you understand how autosomal DNA works but when you're getting started, sometimes you're misled because you're only thinking about the family tree.

If you don't know how autosomal DNA is inherited you might think having yourself or your child/grandchild tested would be a better deal because your/their family tree includes all the branches you are interested in. Instead, you'd be limiting the usefulness of the DNA vs. having older generations (from each branch of interest) tested.

Now that you know what a family tree is vs. a DNA tree, let's look at the problem with shared matches.

Except for very close relatives, there will be a mismatch between the number of shared ancestors in your family tree vs. your DNA tree. Add to that, you will only be aware of a portion of your family tree. For some of you, it's a very small portion.

SPOILER! The #1 solution for dealing with this problem is building a robust tree for the test taker. If the test you are using is not your own, you need to know who is in those other branches in case they are related. If you happen to be reading this post and you have no confusion using shared matches, sure, you might be able to continue without building out the other branches. It's dangerous for anyone to make assumptions based on a very incomplete tree, though.

[If you are dealing with a recent unknown parentage situation, so the whole point is filling in those empty branches, most of this still applies to you. You might still be tipped-off to a not-normal situation because you have confusion between shared matches based off the trees of the matches. You'd just skip this first solution and move to #2 which I'll cover in the other post. Not knowing your own parentage or parentage in recent generations doesn't prevent this situation, it just makes it harder to deal with, like all aspects of DNA.]

So I've said the problem with shared matches, for a knotted tree, is you have multiple shared ancestors which will mislead you. Let's make sure you understand how a list of ICW matches or shared matches is created and why it's provided as a tool.

The Basics of Shared Matches

A shared match is someone who matches you (the test taker) and is a match to a selected match. I like to think of this as you are #1, the selected match is #2, the shared match is #3.

That's it.

This is NOT triangulation. Don't get confused by the "tri-" and three aspects.

All shared matches or "in common with" (ICW) matches are, are the matches (#3) you (#1) share with one particular match (#2), nothing more.

Someone matches to a test taker because they share at least one segment of DNA. ICW and shared matches are not created based on the shared DNA, that IS triangulation. Triangulation means all three people are shared matches with each other and share the same segment of DNA.

If you come from a population where people are related to each other in multiple ways, you will have confusion with shared matches.

Let's say you, #1 share a Smith segment with #2, you share a Jones segment with #3. #2 and #3 share a Miller segment. This is an ICW/shared match group.

You are a match to #2, you are a match to #3, #2 is a match to #3. ICW and shared match lists do not take into account that it's different segments that you share.

Here's why shared matches are provided.

You have fewer triangulation groups. Remember, when we're talking about the DNA you inherited from an ancestor, that DNA is getting recombined down the lineages for you and for the match. You obviously both retained one piece from that shared ancestor but you each may also have other pieces from that ancestor, but different ones.

You might share Smith segment X with match #2 and Smith segment Y with match #3 and #2 and #3 share Smith segment W (you didn't inherit W). Then your shared match group is accurate. If you only share the Smith ancestor with #2 and #3, this has a decent chance of happening (no mistakes due to Jones and Miller).

The idea with using a shared match list is it includes more matches likely to have a single common ancestor than a triangulation group. With segments X and Y, that would not form a triangulation group and you'd miss the clue that there is a shared ancestor between their trees if you didn't use shared matches. More matches increases your chance for success (it means more trees, more people to contact for info, etc.).

When #2, the selected match, is a pretty close relationship, the shared match list should also be like an umbrella list. In this example, #2 has multiple segments of shared DNA with you. The shared match list should include matches who share each of those segments. Breaking this down by more distant ancestor is where clustering comes in, but that's a topic for a different post.

If you look at the shared match list where you only have one shared segment, the list is shorter because there are fewer segments and probably generations (the closer match would share branches of ancestors and the shared matches could share any of those. The more distant match shares fewer relatives with you so the shared matches should only include those ancestors).

Everyone has some confusion or mix-up because you can randomly have the situation I described with the Smith, Jones, and Miller segments.

When you come from a population with lots of inter-relatedness, this situation is common.

This is all very nice theory of how genetic genealogy should work but here's what we sometimes forget...

The point of using DNA is to uncover new ancestors. We do this by comparing matches to ourselves and to each other.

That means you don't know about that Smith ancestor.

If #2 and #3 both have Smith in their tree, that's a clue that's your ancestor. That is how the shared match list is meant to be helpful.

With a knotted tree you can occasionally to frequently have a pitfall instead of a clue depending on the severity of the inter-relatedness.

Shared matches are meant to reveal the ancestor shared between the three matches. For a correct shared match group, you'd expect to swap out #3 for any match on the shared match/ICW list of #2 and you'd share the same ancestor. That's the theory.

When a population has a lot of intermarriages, you get lots of cases of the Smith/Jones/Miller example where not only do you not share one ancestor, the ancestor shared between #2 and #3 is not even yours.

If you saw you shared Smith with #2 and Jones with #3 it wouldn't matter much that Miller (who you aren't related to) was shared with #2 and #3.

You are more likely using a shared match list because you don't know who the shared ancestor is. If you see #2 and #3 share Miller (or determine this through your own research), you'd be wasting your time searching for your Miller ancestor. You don't have one.

This situation is common with interrelated populations meaning pitfalls are more common. This occasionally can happen to anyone but if you have a list of shared matches, you have many chances to discover there are different shared ancestors between #2 and the shared matches. In a "normal" population you would most likely just have an odd one or two shared matches who shared Miller/not your ancestor. In an interrelated population, it is more likely several or even all the shared matches share an ancestor with #2 who is not your ancestor.

But there's also an additional pitfall, as well.

When you are related to someone in multiple ways, you may only know of one or a few of those ways. The shared DNA may not even come from the known relationship(s). Remember, this is the issue of family tree vs. DNA tree. Even with a very complete family tree, you won't know what the DNA tree looks like without lots of analysis plus you have one DNA tree and a match has a different one.

Let's look at our shared match group of #1, #2, and #3 again for an example. We're going to add more relationships but we've still got the problem where 1 & 2 share a Smith segment, 1 & 3 share a Jones segment, and 2 & 3 share a Miller segment. You (#1) don't know of any of these relationships.

But you do see you are related to #2 and #3 through your shared Shelton line. Clearly that's the source of your shared DNA!

Nope.

In this example, there's no shared DNA I'm not mentioning.

In your family tree you would likely find you are related to #2 and #3 through the Shelton, Smith, and Jones lines. In your DNA tree you are only related to each in one way.

This is a made-up example because it's so hard to figure out where the DNA segment actually came from. It is a completely possible situation.

If it's not likely it's only because in an interrelated population you'd probably have even more shared ancestors in both the Family tree and DNA tree (but the two trees wouldn't be identical).

So you missed discovering the new ancestors Smith and Jones. Not the end of the world, right?

Did you notice the THIRD pitfall you run into in this situation?

If you assume the DNA comes from Shelton, you are now looking for every match that is shared with #2 or #3 to belong to that Shelton line. You would keep working on this because those more distant matches might lead you to undiscovered Shelton ancestors (the ancestors of the shared ancestor).

Once again, you'd be wasting your time.

The problems go beyond just being confused by your shared match list.

Knowing you are dealing with multiple intermarriages might reduce or eliminate your confusion. You still need to use your matches.

If you assume the known relationship is the DNA relationship, you can be spinning your wheels. When you do identify a known shared ancestor, you naturally use it to try and find the same ancestor in the tree of other shared matches (or find the known ancestor's ancestors in their tree).

Not assuming shared matches have a single shared ancestor is the first part. Recognizing that you can have multiple relationships, most of which might not be the DNA relationship, is the second.

This is more difficult than someone that is only related to a match in one way (when they identify a shared ancestor, they can follow the common advice and practices for genetic genealogy).

It's not hopeless, you just need to approach the problem differently. Just as the situation I'm describing falls between a "normal" population and an endogamous population, the solution is between them, too. Dealing with endogamy is very hard. You don't need to go to that extreme if your family is not endogamous.

The third post in this series will discuss the solutions to these issues.

No comments

Post a Comment