There is a new report on Gedmatch called AutoSegment. From my understanding it clumps together triangulated matches into clusters. If I were creating this report, I might have called it AutoTriangulator or something similar. I figured it was worthwhile putting down $10 to get one month’s worth of Tier 1 Subscription at Gedmatch
Running My AutoSegment Report
I ran it and was not sure if I was supposed to get an email back with the results. The first time I didn’t get any results, so I ran the report again and got results on the same page where I ran the report. I was asked to download files, so I did. My downloaded file looked like this:
I opened up the file and got this:
The first html file is the one to open:
That gave me 26 clusters of triangulated matches. I am quite sure that the purple cluster is my Chromosome 20 matches. That chromsosome is out of control for some reason. I have written about this before in 2016.
Chromosome 20
There is a more detailed report below:
The purple Cluster is # 24. This Cluster involves three chromosomes. Chromosome 20 has a part to play in four clusters. That makes sense as Chromosome 20 has a paternal component and maternal component.
Identifying My Triangulated Clusters
I recognize the first two matches in Cluster 1. They are two of my Hartley second cousins: Beth and Mike. That relationship goes back to my Hartley/Snell great grandparents. It looks possible that those connections could carry down through Cluster 11.
The other matches are:
- Charles – He shows up twice and tested at 23andMe
- Lori and Phyllis – These two are at Ancestry and are administered by the same person.
- Edith – Administered by the same person who administers Lori and Phyllis but Edith tested at FTDNA.
Lori has the best tree at Ancestry, but I don’t see any obvious connctions. It is possible that building out her tree would give some clues as to the connection.
Pat in Cluster 4
Pat is related to me in two ways. One is as a second cousin in my Hartley/Snell side. The other is Bradford/Hathaway as 4th cousins:
So from Pat’s point of view, she is related to me as a 2nd cousin on her mother’s side and 4th cousin on her father’s side.
A Cluster Spreadsheet
Here I put some of the information into a spreadsheet:
The matches are 2nd cousins except for Jim. These matches go back to Hartley. This family was from England. Or they go back to Snell who were Colonial Massachusetts. Cluster 5 is interesting as at least one match (Matt) is recently from England.
An Out of Place Cluster 8
Cluster 8 is between my paternal clusters but I believe that it is a maternal cluster:
Joshua is the first person in the light green Cluster 8. He matches with Mike in Cluster 1 and his sister Tracy in Cluster 6. Here is where I have Joshua on my mother’s side:
That fact that Joshua matches Mike and Tracy was a bit misleading. The other person in Cluster 8 is Brittany. It is possible that if I built out her tree, I would get back to Nicholson or Lentz.
Lee in Cluster 11
I would like to connect to Lee in Cluster 11:
Lee has Hartley ancestors from the same part of England where my Hartleys are from. That does not necessarily mean that the connection is through those Hartleys, but may be through another set of ancestors. This appears to be the end of the Hartley Clusters for now:
Martin in Cluster 12
Martin is on my mother’s side. His ancestry is from Latvia, so that goes to my mother’s father who was also from Latvia. The connection is on Martin’s paternal side, but his genealogy stops with his father who was born in Latvia.
Cluster 17 – Rathfelder
Cluster 17 is easy as I can identify all Rathfelder relatives there:
Martin was in the green cluster above. I find it interesting that this group contains triangulation in the X Chromosome:
Clusters 18 an 19 – Nicholson
The light green and light blue clusters above are both Nicholson Clusters. I am not sure why they are separated:
Without getting into the specifics, my guess is as follows. Matches and triangulated matches go back to one of the two common ancestors. That means that for each segment where I match any of these people, the DNA we share is actually either from William Nicholson or Martha Ellis. Let’s say that my match with Carolyn favors the Nicholson side. That would mean that the other matches might favor the Ellis side. That would also mean that one cluster is a Nicholson Cluster and the other one an Ellis Cluster.
More on Cluster 19
Cluster 19 has Carolyn, Joan and Iain. Iain has contacted me and I told him the general direction of where the DNA was leading (to Nicholson in Sheffield). The other match tested at FTDNA and appears to be Joan. Joan has a tree at FTDNA. However, it is very basic. I can build this out to see if there is a connection. Joan’s roots are in Alabama mostly:
I make my trees at Ancestry, and here is how Joan’s tree is shaping up:
According to the 1910 Census, Tilden’s father was from North Carolina:
My attempt to make a connection by building out Joan’s tree failed:
This is not unusual. If the connection with Joan is at the fourt cousin level, then the connection would be out one level beyond what I have above. If it is at the 5th cousin or perhaps 4th once removed, it would be out two levels from what I show. The other problem is that the female line identifications become more scarce the further out you go.
Steve in Cluster 22
The next person I recognize is Steve:
Steve (or Stephen in my chart above) is important, because his primary connection with me is on the Clarke Line. I’m a bit stuck on this line beyond John Clarke. Unfortunately, Steve connects on my McMaster side also further back. Steve is a fifth cousin on that line. When I hover over Cluster 22, I see this:
It looks like Susan is the next largest match to Steve.
Finishing the Clusters with Known Names
I mentioned Cluster 24 as the large cluster. Cluster 25 is on my Frazer side also:
Susan, Doreen and Ken are in the orange square and G is Gladys in the yellow area. Our common ancestors are James Frazer and Violet Frazer. Gary is from an area near the Frazers and Shelly has not shared her ancestry to my knowledge. This is what I have so far:
I have 7 clusters on the paternal side and three on the maternal side. I have Lee who seems to be on my paternal side.
Checking by Phased Results
I have some phased kits that a genetic genealogist Martin made for me, so I can tell by the matches at least what grandparent side these clusters should be in.
Joshua’s Confusing Match on Cluster 8
My phased Hartley grandfather kit shows to match Joshua. However, the genealogy shows that he matches my mother’s mother’s side. Here is some more detailed AutoSegment information for Joshua in Cluster 8:
This indicates that the two triangulate with each other and me. Next, I’ll check my paternally and maternally phased kits. These were generated at Gedmatch based on my mother’s DNA test. Joshua matches me there on my maternal kit. When I recheck, it appears that Joshua does indeed match on my mother’s mother’s phased kit. So I don’t know what I was seeing before. I keep these mistakes and corrections in my Blogs to remind myself how easy it is to get off track with all the information out there.
Here is a continuation of my spreadsheet:
Here I have also color coded the grandparents. Haretly and Snell are blue, Frazer/Clarke is green. Lentz/Nicholson is orange and Rathfelder/Gangnus is yellow.
Clusters 14-16: Frazer Side
Cluster 14 seems to favor the McMaster side:
The first blue line in each case is Marshall and the second is Craig.
The key is with Keith. My common ancestor with him is with James and Fanny Mcmaster. My common ancestor with Susan and Katherine are Frazer/McMaster. However, their overlap with Keith seems to mean that the connection is on the McMaster side. Marshall and Craig have a match overlapping with Katherine but starting at 15M, a little later than Katherine’s which starts at 7M.
Margaret is in Cluster 16. She has some ancestors near Enniskillen:
This location comes up a lot. This may be on my Clarke side or from an unidentified Frazer wife’s family. She also has a Henderson and MacGregor in her tree:
Henerson was a second wife of Clarke (though not known to be related to me). McGregor comes up as a possible ancestor on my ThruLines at Ancestry:
From what I can tell, the MacGregor name comes from a George MacGregor Frazer who is in some people’s Ancestry Trees:
Filling in the Rest of My Spreadsheet
Here I have under the GP column, blue for Hartley, green for Frazer, yellow for Rathfelder and orange for Lentz. It looks like I only have four clusters to go.
Cluster 20
Cluster 20 is on Chromosome 4. Another way to check on these clusters is by DNAPainter or Visual Phasing.
The match seems to go through a maternal crossover, so my guess is that this match is on my paternal Frazer side. The last of the trhee matches in Cluster 20 is Gabrielle. She tested at Ancestry. At Ancestry, her match is reduced to 18 cM. She shows no shared DNA matches, probably due to the low match level.
From DNAPainter, these Cluster 20 matches are probably from my Clarke side:
This is a side with a brick wall.
Cluster 21
Cluster 21 is from Chromosome 22.
That area between the two arrows is not well mapped on my Chromosome 22. Orange Cluster 21 is medium sized:
Although the are of Chromosome 22 is not well identified, I can identify that the connection is through my Frazer grandmother.
Cluster 23
This Cluster has two triangulated matches at the beginning of Chormosome 18:
The matches are with Patricia and Carl.
I’m leaning toward Patricia and Carl being in the Clarke/Spratt section of my DNA. I have Patricia on my paternal side so that means I must be right.
Cluster 26
That leaves one last cluster.
The first three matches are from Ann and her close relatives. From my spreadsheet of matches, Ann is on my Lentz side:
That match is around the red arrows and would be more specifically on my Nicholson/Ellis side.
The Completed Key
Based on Visual Phasing, my match spreadsheet and DNAPainter, I was able to identify all my clusters at least back to one grandparent.
Summary and Conclusions
- Out of the 26 clusters, 6 were maternal clusters.
- Of the remaining 20 clusters, they were split fairly evenly between Hartley and Frazer
- It seems like I could have differentiated my Hartley clusters more.
- Every tool seems to give some new understanding to my DNA matches
- It would be interesting to look at other of my 5 siblings’ AutoSegments to see how they differ from mine