Cluster Overlap

Do you ever feel like you’ve figured out a new way to cluster matches and then realize that those clusters you thought you had found included matches from distinctly separate areas of your tree? Well, that was my experience when I expanded my mountain/valley analysis of my Crawford line.

As explained in my DNA Clusters: Mountains and Valleys post, I had a ‘missing mountain’ for the parents of Nelson G. Crawford in my original diagram.

Nelson Crawford’s parents were James Crawford and Sally Smith Duggins Crawford. Since Sally had two sons from her marriage to Alexander Duggins prior to her marriage to James Crawford, Nelson had two step-brothers. Knowing that I had matches whose lines go back to Nelson’s sister and a step-brother, I decided to expand my Mountain and Valley diagram to include those lines.

My thought was that I could learn more about this SMITH branch of my tree from these Duggins matches and the matches we share. I only have two matches on ThruLines thru the Duggins side of Sarah’s family. Both of those matches have ONE shared match. Thus, I’m only working with FOUR matches to potentially learn more about this SMITH line. When I followed the link to the profile page for one of my matches to view the shared matches, I was excited to see the shared match marked as having a ‘Common Ancestor’!

When I clicked to view the tree so I could see who the common ancestor was, I was disappointed. The common ancestor wasn’t on my grandfather’s CRAWFORD line, but was Osmond Bland on the line of his wife’s CURREY line.

I discovered a similar situation when I looked at the shared matches for descendants of my ancestor’s sister: Polly Crawford. Two of those descendants had shared matches that were labeled ‘MOTHER’S SIDE’. 

Since my CRAWFORD line is my dad’s side of the tree, these matches labeled ‘Mother’s side’ clearly indicate that I have some sort of cluster overlap going on. I am NOT a DNA expert and can only explain this overlap based on what I know about my tree.

  • I don’t have first cousins marrying each other in the first six generations of my tree.
  • James Crawford and Sally Smith are in the 7th generation of my tree, making them my 4th great grandparents.
  • I’m looking at matches for FIFTH cousins and thus share a smaller amount of DNA.
  • My tree is colonial. The vast majority of my ancestors were in the colonies at the time of the Revolutionary War.
  • My ancestors tend to come from New England, Pennsylvania, Maryland or Virginia.

So, I have learned that I can’t make any assumptions from shared matches of identified cousins. Since I’m wanting to learn about distant generations, those shared matches could descend from a different common ancestor.
Thank you Donna Leeds for prompting me to look at my data this way! Following your blog post, Who Is in a Cluster, helped me discover the potential for cluster overlap in my tree.

DNA Clusters: Mountains and Valleys

Dana Leeds’ recent blog post about “Who Is in a Cluster” demonstrates that although all people in a cluster share DNA with one person in the cluster, they don’t necessarily share DNA with another member of the cluster.

Curious as to what this would look like with my Crawford branch of my tree, I looked at the DNA matches I share with a first cousin once removed. At first, I was trying to map out all of my known matches. However, this became cumbersome since I have a lot of 2nd cousin matches. As I was creating my matches diagram, I realized that I needed to be looking for matches that were for the different surnames.


The blue box represents myself and the yellow box is the cousin used to create the cluster.

  • My ‘Valley’ is the common ancestors between myself and E: Judson Foster Crawford and Josie Winifred Hammond. I have shown two other matches thru these common ancestors: M and M2nd. Not only are we all cousins (myself, E, M and M2nd), but we would all share DNA with everyone else mapped out in the diagram.
  • I have two ‘hills’ in my diagram: Washington Marion Crawford/Mary Foster and Richmond Hammond/Sarah Ralston. PrivateLida  and Private Hattie are both related to the four cousins in the valley but are not related to each other.
  • I have diagrammed 3 mountains representing three lines: Foster/Ostrander – Hammond/Fisk – and Ralston/McCormick.

By visualizing the cluster in this way, I have identified matches that I can use to figure out some of those matches whose trees are sparse or missing. For example, if I look at matches shared with PrivateEliza, I will be identifying matches who are either part of the valley with a common ancestor of Zebulon Foster and Caroline Ostrander or who descend from an ancestor of Zebulon Foster or Carolne Ostrander. I could then create another diagram for this cluster 
This diagram also pointed out a missing ‘mountain’. I don’t have a ‘mountain’ for my Crawford line back to the Nelson G. Crawford or James Crawford generations. Although ThruLines shows that I have some matches who descend from Nelson G. Crawford and James Crawford, only one of those matches shares DNA with the cousin I used to create this diagram. Unfortunately, I didn’t recognize this match when creating the diagram, thus leaving out a ‘mountain’ for Nelson G. Crawford.

Working thru the creation of my diagram, I have made two observations:

  • I should have checked ThruLines to see whether there was a match to fill in the hole in my diagram.
  • Excel likely isn’t the easiest software to use to create this diagram. 

Thank you Dana Leeds for prompting me to look at my DNA results in this way.

Timelines – Cluster Research

Do you use timelines in your genealogy research? I’m guessing that most people who have been searching their family history for some time have used some sort of timeline.

When I retired and starting spending more time researching my family tree, I created a spreadsheet for my Crawford family research.

In the spreadsheet, I was able to color code the people, locations and time periods. All of this information was copied from my RootsMagic data into the spreadsheet.

It wasn’t until I read Randy Seaver’s Saturday Night Genealogy Fun – Make a Timeline Report blog that I seriously looked at the Timeline Report in RootsMagic. Thanks to his post, I started looking at the built-in timelines on Ancestry and within RootsMagic.

As I was learning about this report, I wondered if I could use RootsMagic Software to automatically generate a timeline similar to my Excel spreadsheet. Since the report allows one to use a ‘marked group’ to create the report, I decided I would try to create a marked group for my Crawford cluster research.

To experiment, I clicked on groups and then clicked on the folder to create or edit a group.

That opened a ‘Named Groups’ window.

When I clicked on ‘NEW’ that opened a window to allow me to select people. To start marking my cluster, I highlighted the first person in my cluster and clicked on the ‘MARK GROUP’ tab.

For this cluster, I wanted to show information for two generations of several different families. Thus, I selected “DESCENDANTS OF HIGHLIGHTED PERSON” from the menu. I then selected “DESCENDANTS AND SPOUSES” and changed the number of generations to 2. Then I clicked OK to mark that family for my cluster.
I then highlighted the next person I wanted to include in my cluster and repeated the process to mark the descendants of this next person. When finished marking the various families, I clicked the OK button at the bottom of the ‘select people’ window. 

That opened a window prompting me to name the group.

Once I had the group marked, I could then create the timeline report.

I pulled down the tab by EVERYONE and scrolled down to find my Crawford Cluster group.

When I clicked on Generate Report, it produced a report pulling together all of the events that I had entered for these various families. The report includes the date, the event, the person and the place of the event.

This report has several advantages over my original spreadsheet. The RootsMagic report

  • can be generated at any time
  • includes all events entered and not just selected ones
  • can be expanded to include others in the cluster by modifying the marked group
  • can be exported as a text file and then imported into Excel

THANK YOU Randy Seaver 

for blogging about this timeline report!

Sub-Clustering

I just watched Blaine Bettinger’s video, Sub-Clustering Shared Matches. As I was following the video, I was also trying to do this with my brother’s DNA matches.

The match I started with is listed on my match page as a second cousin. In reality, she is a first cousin once removed. Our common ancestors are Judson Crawford and Josie Hammond.

Her shared match list contained 157 matches. I marked them all with a yellow dot labeled: ICW EB.

When I got to the part about creating Sub-Group 1, I wasn’t sure how to proceed. Other than my brothers, the first match on the list was a first cousin to ICW EB. The next 4 matches are listed as 3rd cousins. They are all known descendants of my great-grandparents.

Unsure on how to proceed, I searched the Facebook group, Genetic Genealogy Tips & Tricks, for the term sub-group. That helped me find the original post on May 1 about the video and this technique. In the comments, I found the answer to my question about using these close cousins to create subgroup 1.

Thus, I skipped my second cousins who I know descend from Judson and Josie Crawford. To create sub-group 1, I started with the first match after these cousins. My brother shares 143 cM with this match and has 27 shared matches with her. These shared matches were marked with an Orange dot labeled, “ICW EB – Sub-group 1”.

For ‘sub-group 2’, I used the next match that does not descend from Judson and Josie Crawford. This match descends from Judson’s father, Washington Marion Crawford. These 32 shared matches were marked with a Green dot labeled, “ICW EB – Sub-group2.”

I skipped another match with Judson and Josie as common ancestors for ‘sub-group 3’. The next match was also a descendant of Washington Marion Crawford. I added these shared matches to Sub-group 2.

The next match is believed to be another descendant of Washington Marion Crawford. Since I haven’t been able to verify the ancestor of this match, I created another subgroup on this match.

I continued working thru these shared matches, creating 11 subgroups. The common ancestors of my ICW match are Judson Crawford and Josie Hammond. Thus, I expected to find subgroups for CRAWFORD, FOSTER, HAMMOND and RALSTON. The subgroups revealed CRAWFORD and FOSTER matches. However, the HAMMOND and/or RALSTON matches are not obvious in these 11 subgroups. 

This is an interesting method to study DNA matches.