Categories
Unique Visitors
33,349
Total Page Views
505,827
View Teri Hjelmstad's profile on LinkedIn
 

 
Archives

I Discover a DNA Cluster Tool

This week I had the opportunity to hear a genealogy presentation by a local DNA expert. Dr. Greg Liverman of Pinewood Genealogy addressed the first meeting of the Highlands Ranch Genealogical Society’s 2020-2021 season via Zoom.

He showed us how to use DNA clusters to identify unknown ancestors in recent generations. I have such an ancestor—my dad’s maternal grandfather. According to Greg, this is just the sort of research problem that cluster analysis can solve.

The task of sorting all of Dad’s DNA matches by hand was overwhelming. I learned from Greg that I do not have to do that. There are tools available that will do the sorting for you. He mentioned several.

I decided to try the Auto Clustering feature on the My Heritage (www.myheritage.com) online genealogy platform. It is free to use when you have an account, and my husband/tech advisor had already uploaded my dad’s DNA data to the site. All I had to do was run a report.

To my surprise, what takes hours to do by hand required only a few minutes with the sorting tool. I soon had a color-coded report that put my dad’s DNA matches into relationship groups, or clusters.

To find relatives of an unknown ancestor, Greg told us to look for a cluster of matches who are not related to any family members whose family tree you already know. Locate the names of people in that cluster who are most closely related to you, who match you at the level of 50 cM or more. Then look at the family trees provided by these people for their common ancestor. This ancestor is probably yours, too.

This seems a simple enough process. I needed to search for people not related to anyone in Dad’s paternal line or his direct maternal line.

The report came back with twenty clusters of matches, some including only a couple of people who are quite distantly related to us. I set those small, remote clusters aside for now and began with the first, larger clusters to see if I could recognize anyone:

  1. Cluster 1 had three people who matched at over 50cM. Two I do not know, and they have no family trees on My Heritage. The third is a descendant of my Dad’s paternal cousin. I assume the others belong to the same paternal line.
  2. Cluster 2 included two people closely related to us. Both had family trees on the site, and both are descended from Dad’s paternal grandmother’s family.
  3. Cluster 3 had two close relatives, both descended from another of Dad’s paternal cousins.
  4. Cluster 5 included Dad’s double cousins, who descend from both his father’s family and his direct maternal line.
  5. Clusters 6-20 included either no close relatives or else relatives who barely met the 50cM threshold. Many of them also matched someone in Clusters 1, 2, 3, and 5.

That left cluster 4. It contains just 5 people, none of whom reach the 50cM match level. But these people do not match those in the other clusters, all of whom descend from the families of Dad’s three known grandparents.

If Greg is right, these people in Cluster 4 should belong to the family of my unknown great-grandfather. If I can identify their common ancestor, I can learn who his family was.

I was disappointed when I found that none of these people has placed a family tree on My Heritage. Learning about their families will take more legwork. That is the next step.

Perhaps with the tool of cluster analysis I have at last found the key to discovering the identity of my unknown great-grandfather. He has hidden himself from us for over 100 years. I hope his DNA trail can now force him to step out from behind the curtain.

Leave a Reply