Part 1 of Walking Back My Clusters was long and rambling. I learned a few things, looked at a few family trees and reached out to a few DNA matches at AncestryDNA. While writing my previous Blog, I came up with a better way of presenting the results of walking back my clusters. I realize that this may sound obscure if you are not already into genetic genealogy and clusters, but hopefully the readers understand the basics of DNA and clustering.
The New Cluster Results Format
Here it is:
I have my four grandparents in four colors. The thought is that even if I am lost as to what a cluster represents, I should know under which grandparent the cluster belongs. At the top, I show the cM cutoff for the clusters. I have a small column for the cluster number that the program produces. This is a relative number and changes for each analysis. To the right of the cluster number, I have the name of the closest surname that cluster represents and the date that ancestor was born. If I don’t know this, I may give a geographical hint. As far as which ancestor to use, it is somewhat subjective. On the top row I have Hartley going to Pilling 1802 and Snell 1866. It may have made more sense to use an earlier Hartley instead of Pilling, but I suspected that one of the people in that particular cluster went back to Pilling. Under Lentz 1900, that went to the two parents who were Lentz 1866 and Nicholson 1865. This new representation, so far, keeps everything close together where I can keep track of where the clusters are going.
Cluster 13 on the 30 cM Limit
This is where I left off on the chart above. There are only three in this Cluster. I have a note on one of the match’s that they have a Northern Ireland background. I’m going to peek forward to 25 cM to see if I get any more hints. This adds one more match and tree. This tree in addition to the match with the largest tree has Canadian ancestors. I’ll take a look at the largest tree in Cluster 13:
This particular match had Ontario ancestors. The parents had connections to Owen Sound, Ontario. That sounds familiar from one of my distant Frazer relatives. My hunch is that the connection is not on the McRae side as they are listed as being originally from Scotland and Presbyterian. I’m not aware of Presbyterian ancestry on my Frazer side.
Here is my best guess:
Jane or Jennie in the bottom right of the tree is from Inniskillen. I assume this to be a variant of Enniskillen, where a lot of my DNA match leads take me:
All that to make a guess at Cluster 13.
Cluster 16 on the 30 cM Limit
I’m a bit stuck on this one. I think it is on my maternal grandfather’s side. Here is what I have so far:
I know that at 20 cM, I have 50 Clusters, so I have a way to go.
25 cM Clusters
Here I have 27 clusters, but some may be compound clusters.
Clusters 1 and 2
These split out the previous Cluster 15 which I had assigned to Fanny McMaster born 1829. Let’s take a second look:
Whitney in Cluster 1 matches everyone in Cluster 2. That is because he is a closer relative to me than I am to others in that Cluster:
What is also confusing is that Margaret McMaster had two McMaster parents. I had the previous Cluster 15 correct on Fanny McMaster. However, it would be easier for me to think of this now as having the old Cluster 15 on Margaret McMaster 1846. Then I could assign Fanny to BV and mt and James McMaster to Whitney. Here is how I’m related to Whitney:
I am Whitney’s third cousin once removed. Here is one case where I changed an earlier analysis based on a later one:
After going through some more clusters, I came up with this:
I mentioned that I Had 27 clusters at 25 cM and 50 clusters at 20 cM, so I gave up doing this for now. I think I have an easier way to go about this which I will explore in my next Blog.
One interesting thing above is that the orange Rathfelder line jumps from 1921 to 1819 in the above cluster summary. My explanation is that there were not many DNA matches for that line at Ancestry. That line represents my mother’s father who was from Latvia. He jumped ship and came to the US in 1916. I have has one Rathfelder 2nd cousin once removed who tested at Ancestry, but one person is not enough to form a cluster at that level.
Going All the Way to 6 cM with Shared Clusters
The creator of the Shared Cluster Program commented on my previous Blog and recommended I take the clusters down to 6 cM. John Brecher tells me I won’t get any more clusters but more matches associated with those clusters. At first I thought that I had to leave the “Lowest centimorgans in shared matches” as 20, but that gave me the same results as my last run using 20 for both values in that row. So now I have both values set to 6 cM:
This kicked up my spreadsheet from 912 rows to 2453 rows. I suspect that this is where the Shared Cluster Program really shines.
Filtering the 6 cM Results
Excel has a filter button. I would like to filter my results on Common Ancestors:
When I choose Filter, an arrow appears in each column’s heading. I click on the arrow under “Common Ancestors” and unclick the ‘Blanks’ option which is at the bottom of the list:
That will give me each row that has a common ancestor:
I couldn’t get all my results in one screen shot, so the top is cut off. Cluster 7 appears to have many of my 2nd cousins, so it shows other more distant clusters that they are related to. The 13 is highlighted in the Correlated Clusters column because it gives a clue to Cluster 13 with common ancestors Snell and Luther that I didn’t have before for Cluster 13. The same is true for associated Cluster 19 with common ancestors of James McMaster and Fanny McMaster. If I add up all the clusters plus associated clusters that have Common Ancestors, that adds up to about 20. Those will be a good clues to identifying my 50 clusters.
I highlighted Nigel because he is an interesting case. He has a fairly high DNA match with me. He’s my 5th cousin, once removed:
I don’t recall Nigel being in a cluster before due to the distant of his relationship to me. So it is good to see him in Cluster 40 now.
On to the Next Blog
Part of the difficult part of comparing these Clusters is cross checking between say, a 25 cM analysis and a 20 cM analysis. For example, Charlie was in Cluster 35 at 20 cM. What Cluster was he in at 25 cM? I hope to figure out a way to make that a little easier in my next Blog using MS Access. There may be other ways. It makes sense to me also to walk the Clusters Forward instead of back. That is because the older clusters have more people in them. As noted above they also have about 20 identified Common Ancestors.