Mad about clusters

One of the great things about genetic genealogy is how fast the field is growing. It’s changing and improving all the time. We can do all kinds of things now that 5 years ago were impossible. There’s so much to learn.

I’ve written previously about clusters on my blog and I’m still periodically reviewing my Johnston matches and finding additional people. It’s now a tree of 270 people with 16 DNA matches across all of the major DNA databases. I also have a huge mystery paternal cluster – a tree of 184 people with 11 DNA matches. On that one, I’ve no idea how we’re connected yet.

Recently, I decided to play with the auto-clustering tool on MyHeritage to see if it could help me. If you’re not familiar with the concept, MyHeritage runs all your matches through their magic system and it produces a visual chart of matches clumped together. In an ideal world, you’d have four neat squares linking to your grandparents, but of course things are not always that simple. Here’s what mine looks like currently.

Down each axis you get a list of names and each square represents a confluence between 2 of your matches. I’ve removed the names for privacy here. The grey squares also represent a connection but it may be that one of the two matches is too closely related to you to be properly sorted. Below the image, you also get a text version of each cluster, and it comes with a download openable in a spreadsheet program to allow for better manipulation of the data.

Cluster 1 in the picture is in red and it’s the largest cluster. Of the people in it, I have identified one individual as part of my Johnston mystery cluster. A second person matches my 3R1C and I had previously presumed she connected to that branch, but I now realise that she must connect instead to the other half of that couple. The 7 other people in that cluster have either not responded to messages or do not have enough of a tree for me to play with. Therefore, running this auto-cluster report has already been of great value because I now know all of those people link to my Johnston cluster.

But even better, I also have access to that 3C1R cousin’s kit on MyHeritage and I also ran the AutoCluster report on him. By using our combined reports, there is more information to be gleaned.

One area where the clusters are less useful is if you have endogamy or pedigree collapse in your tree, which is common in more isolated rural parts of Ireland like the islands, Kerry and Donegal. Here’s a cluster image for a client who came from one such area.

If you have a kit on MyHeritage, it is well worth the effort to generate an AutoCluster report and have a play with it. You can learn more about how it works here. If you subscribe to Legacy Family Tree webinars, they also have a couple of videos on how to get the best from this tool.

Leave a Reply

Your email address will not be published. Required fields are marked *