The AutoCluster Craze

It seems all the cool genetic genealogists are using AutoCluster at Here is the welcome page for this new DNA analytical tool:

I have decided to try it. I have seen some screen shots. Autocluster appears to be a way of easily clustering your DNA matches to see which ones go with which.

I registered and first tried Ancestry where most of my matches are. I added Ancestry and it showed all the people that I am linked to through Ancestry. There is a blue  autocluster button to select:

The second button is for my profile and I chose that. Then there are three choices:

I chose A. I see now that if I was in doubt, I should have chosen A so that was good. In not too long a time, I got an email giving my 20 closest DNA matches. I knew this already. I also got a spreadsheet and the important graph:


On the top and sides of the graph are names of my matches and how they match each other. The Key above shows 28 matches. This is based on the default values:

This forces my matches into a fairly narrow range.

What Do the Clusters Mean?

The clusters are on the idea of “birds of a feather flock together”. These are matches who match each other. The first orange cluster would be people who descend from my Hartley great-grandparents. This couple had 13 children. That means that I have a lot of 2nd cousins and some remaining 1st cousins once removed, then 2nd cousins once removed.

The Snell Side

As I look at the Hartley cluster more closely, I see that there is also a Snell subcluster within it that is not Hartley:

That is an important distinction as I try to separate  Hartley and Snell DNA.

A Small Maternal Red Cluster: Lentz

Hartley is on my paternal side. The only cluster on my maternal side is the red one. Here is my tree up to my 2nd great-grandparents;

The orange box is around my Hartley/Snell ancestors. The red box is around my Nicholson ancestors. This corresponds to the red cluster in the chart.

The Purple Frazer Cluster

Gladys is the third person in the purple cluster. She is in my Frazer DNA project. Here is how we match:

What is also interesting is that Gladys does not match the first person in the purple cluster. However, Gladys matches possible Frazer relative #2 who matches possible Frazer relative #1. Now what is very interesting is that I had that the match [that I am calling Frazer relative #1] has a McMaster ancestor. I have tried to show in the past, that “Frazer relative #1” has this ancestry through Marriainne below:

Although the “Cluster” is not proof that I was right, it seems that it provides strong evidence that I was right. It appears that Match #1 has Frazer DNA even though she doesn’t know she has Frazer ancestry. Even though I did a simple cluster, it appears that the results are quite powerful.

The Green Clarke Cluster

I have written quite a bit on this line. The people I match with are aware of their McMaster ancestry. I match these people on their McMaster Line but more closely on their Clarke line.

Actually, the last person in the cluster isn’t sure how his Aunt fits into the picture. However, I have not seen the family tree.

The Last Cluster

This last cluster is a little harder to nail down with only two people. Note that there is a link to the Clarke Cluster above. I had originally thought that this might be a McMaster Cluster, but the last person in the cluster has Spratt ancestry. The reason that I thought that this might be a McMaster Cluster is because the person matching from the Clarke Cluster had McMaster and Clarke ancestry.

I’ll keep an open mind and put both names into the mix.

Cluster Summary

I am very happy with this new tool as are other genetic genealogists.

These are the lines above that I have identifed using a simple threshold cluster technique. I have Hartley/Snell, then the earlier Snell. I have Frazer and Clarke. Then I have either Spratt or McMaster for the very small cluster. For my maternal side, so far, I only have Nicholson.

Trying FTDNA

Next, I added the FTDNA website to the AutoCluster Program. This time I lowered the lower threshold to 25 cM. Perhaps this will take longer to run. This will be a good chance to look at my FTDNA matches as I haven’t checked them out in a while.

This time, instead of 5 clusters, I have 33. My first Cluster has 108 members instead of 16. I assume that Cluster 1 at Ancestry is the same as Cluster 1 at FTDNA, but there are not a lot of people that have tested at both places. I would need to lower the threshold at Ancestry and then see if I saw any common names.

Frazer Cluster

The Frazer Cluster below in purple is Cluster 4:

I recognize my 2nd cousin once removed who is the first match. These Frazers tended to intermarry. This clusters carries an overlapping look that I also saw at Ancestry – even though there are different Frazer ancestors in the Ancestry Cluster versus the FTDNA Cluster.

Rathfelder Cluster

Beneath the Chart is an analysis part:

The first cluster to come up there is Cluster 9. That is likely due to the large match with my 2nd cousin Catherine. She is on the Rathfelder side which was missing in the high threshold Ancestry Cluster Analysis.

Cluster 9 is the Blue Cluster on the lower right above. It would be worthwhile pursuing the other two in the cluster. According to the tabular analysis above, Pamely has a tree. The link brings me right to Pamela’s tree. It goes back to her grandparents. So if I were to expand Pamela’s tree, I might see where the match is. The Rathfelder’s were from Latvia, so that is an easy place to notice.

Cluster 6 – Clarke

I could recognize this Cluster by one person who had Clarke as his middle name.

FTDNA Clusters

With 33 Clusters, it would take to long to look at all of them in this Blog. However, I am curious as to the 33 Clusters that came up.

Summary and Next Steps

  • The AutoCluster Tool is very helpful with AncestryDNA. This is because AncestryDNA doesn’t have the Chromosome Browser to check matches.
  • I would like to figure out why I have one large cluster at FTDNA versus all the small clusters
  • I would like to try this at 23andMe to see how it works there.
  • If nothing else, this tool should help focus my DNA research.
  • I would like to be able to cross-reference the clusters. For example, between AncestryDNA and FTDNA.
  • I would also like it if there would be a way to combine the clusters from the three companies.
  • Further, it would be great to add MyHeritage to the mix.







Leave a Reply

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