MyHeritage AutoClusters: Update
Earlier this week, MyHeritage raised the upper threshold for AutoClusters to 400 cM. Excitingly, this allows us to see some of our closer matches in our clusters! (By the way, this is the number I recommend when starting the Leeds Method.)
What Are AutoClusters?
According to MyHeritage, AutoClusters is an “automatic tool that organizes your DNA matches into clusters that likely descend from common ancestors.” This tool, which was developed in collaboration with Evert-Jan Blom, is based on the Leeds Method. The chart creates clusters for approximately 100 of your DNA matches, though it does not use your closest matches.
Below is a chart showing my dad’s MyHeritage AutoCluster chart. The names are blocked out for privacy.
Cluster Arrangement
These clusters are arranged in size from largest to smallest. For example, the two largest clusters in the upper left have 12 members each. The lowest clusters each have 3 members.
You are more likely to either know how you’re related to the matches you share more DNA with or to be able to figure out those relationships. However, your highest matches might be in some of the smaller clusters. For example, in this chart my father’s highest matches are as follows:
-
- 381 cM – cluster #11 with 4 people
- 375 cM – cluster #4 with 8 people
- 344 cM – cluster #11 with 4 people
- 308 cM – cluster #2 with
Working with Clusters
I would suggest working with clusters that have some of your higher matches first. For example, let’s take a quick look at cluster #11 which includes two people who share more than 300 cM of DNA with my dad.
If you scroll down from the chart you received, you will get information about the individual clusters. Below is the information about cluster #11:
Diagramming the Cluster Part I
I already know how three of the people fit into our tree. They, along with my dad, are all descendants of Andrew Stewart and Bessie Merrill. I’ll start by diagramming them.
Identifying Doni
However the fourth match, Doni, is someone I don’t recognize. Because of my diagram, I can hypothesize that she is related to my dad through the Stewart or Merrill family.
She has a tree with 59 people in it, but I don’t see either of those surnames or any others I recognize.
Although I don’t recognize anyone on this tree, Doni does have a “hole” in her tree with her grandmother, Annie. So, I went to Ancestry to see if I could find a public tree showing her father’s ancestors. Since I had a full name and dates of birth and death for her father, I searched the public member trees using that information. I quickly found a tree. And, it happened to be Doni’s tree!
Doni’s more complete tree on Ancestry.com shows that her grandmother, Annie’s, maiden name was Merrill! And, I know that Annie’s grandfather was the brother of my 3rd great grandfather, James Merrill. So, now I can add Doni to the diagram for cluster #11.
Diagramming the Cluster Part II
Below is the complete diagram of the 4 people in cluster 11. As you can see, everyone is a descendant of James Merrill. They clustered together because of the DNA they inherited from him and/or his wife, Nancy Waldron!
Clustering is a powerful tool that can help us more easily identify our matches. In this case, Doni was an unknown match who clustered with 3 other people in cluster #11. Because of her Shared DNA Matches on MyHeritage, I knew she should be related to my dad through either the Stewart or Merrill family. And, with a little bit of research, I was able to connect her to our tree!
Wow! Dana you’ve given me new insight. I haven’t yet diagramed my clusters. I’m very visual so this may help me a lot. Yes, I’m still working on that brick wall I shared with you at Rootstech, along with another brick wall. Off I go. 😍
They must have changed something else as well because until now MyHeritage has just given me one big pink blob because of endogamy, but when I ran it a few days ago I got 14 need clusters. A lot of gray dots, but usable clusters. Even my mom got clusters and she has a lot of pedigree collapse.
Now I’m working on verifying that the clusters are ‘correct’ since on a first pass, the matches in the clusters are from different branches of the family. So I’m looking to see if there is an other connection from the correct branch.
That’s awesome!!! I’m not sure what else they changed, but if turned your blob into clusters – even with the gray dots – that’s a lot more helpful!
And, that’s strange that they appear to be from different branches of your family. I hope you do find another connection!
I love diagramming clusters! It really helps me see everything a lot better than just putting it onto trees. 🙂
This is a huge help to me as I try to envision a practical way to sort through matches in my clusters. TY for explaining in such detail!
I’m glad it helped!
Hi, Dana,
What approach would you suggest for those of us who have no idea who these clustered matches are or how they relate to us (not even one recognizable person in the cluster for almost all of the clusters), especially if they don’t have trees or if the trees don’t share any common surnames?
Thanks!
Hi, Kat. You often have to build those trees back even farther and/or build trees for those who don’t have a tree or have very small trees. Unfortunately, on MyHeritage you usually have only a few matches in a cluster and not many trees. If you have tested at Ancestry, it is often a better site to work with.
Hi there,
I just used the cluster tool with My Heritage.
My original test was Ancestry.
It has a mother and son that I match with, in different cluster groups.
It also left out a new match of over 100cM, from any clusters, even though I know we both match with another person who was in a cluster. The person left out actually has a much closer match to the person in the cluster, than to me. They are about the same to me.
The explanation given was that the left out person (one of over 80 other people) met the cluster criteria, but did not fit into a cluster with anyone else.
I don’t understand why that person was left out and did not fit in with at least one person I know is a match to the same family, or why the mother and son (not the same family as the other people) aren’t in the same cluster as each other.
Thanks in advance for your help/answer.
Hi, Stephanie. A mother and son could show up in different clusters based on who they are matching. And, a person can be left out even if they match matches IF they don’t fit the matching criteria. I’m not sure whether the matching criteria on MyHeritage is publicly available or not. If you are on Facebook, I recommend joining the group “Genetic Affairs – User Group” where the programmer will likely give you a better answer.
Dana
Hi
Why do some people not appear in Autocluster when people with both higher and lower DNA matches do?
The person I am most interested in is linked thus:
1.5% (107.9 cM) Shared DNA
3 Shared segments
79.9 cM Largest segment
Is it because they share no DNA with anyone else other than me? If so is that common for someone whose heritage is 60% Ashkenazi Jewish, 22% Irish, Scottish and Welsh and 16% Scandinavian?
Thanks
Hi. If a match does not share DNA with anyone else they will not show up in an Autocluster report. And this wouldn’t be common for someone who was 100% Ashkenazi Jewish. But it could happen in almost any population. It is pretty common in countries outside of the U.S. or in recent immigrants to the U.S. It could also happen in a family with few children.
Hope this helps!
Dana