Thursday, August 16, 2018

Color Clustering: Top 25 Fourth Cousins

For more on Color Clustering & DNA, please visit my new website at: www.danaleeds.com 

For another look at how Color Clustering works with 4th cousins, I created a Color Cluster chart then added the test taker's top twenty-five 4th cousin matches. I was able to easily sort all but one of these 4th cousins into Color Clusters!

Original Color Cluster Chart (click here for Color Cluster Method)

Color Cluster chart for actual test taker.
Names changed for privacy.

This test taker's AncestryDNA 2nd & 3rd cousins fell into 4 Color Clusters, labeled C1 through C4, with three "unclustered" cousins.

  • 2nd & 3rd cousins who are in more than one cluster are in redBarbie, Ken, & Mark.
  • 2nd & 3rd cousins who did not have shared matches with other 2nd & 3rd cousins are in "unclustered" columns: Lena, Sue, and Mike.

Color Cluster chart LABELED

Labeled Color Cluster chart

The test taker's four sets of great grandparents were identified from her research as follows:

  • G1 - Bailey/Bowman (father's father's line)
  • G2 - Stark/Dunn (father's mother's line)
  • G3 - Hillard/Morris (mother's father's line)
  • G4 - Washington/Manning (mother's mother's line)
I looked at each cousin's tree and did one of the following:
  • Put a "NO" in the cell if there was not a tree and I couldn't easily identify to which cluster(s) the cousin belonged.
  • Typed G1, G2, G3, G4, or a combination of those in the colored cell if the person had a tree and I could determine which surnames they fit in under OR if the genealogist had done research and discovered the relationship herself. (Note:  the "unclustered" cousin, Lena, was identified as belonging to G4.)
I then labeled the columns according to the cousins found in them: G1, G1, G2, G3/G4, G4, unclustered, and unclustered.

Adding 4th Cousins

Twenty-five 4th Cousins added into Color Clusters

Directly below this Color Cluster chart, I added the names of the first twenty-five 4th cousins. For each cousin, I looked at AncestryDNA's "Shared Matches" and determined which 2nd/3rd cousins they were matching. I colored in the appropriate cell and labeled the cell with the number of shared centimorgans (cM).

2nd/3rd/4th cousin Color Cluster chart

Above is the final chart which includes all of the 2nd/3rd cousins (sharing <400 cM) and, below it, the first twenty-five 4th cousin matches. A few things to note about the 4th cousins:

  • Owen - at this point, Owen is still not in a cluster
  • Mary & Bill - they both matched previously "unclustered" cousin Sue, so the three created a new cluster. We do not know what part of the family this cluster belongs to at this point
  • Others - a few did not match any 2nd/3rd cousins, but when I opened their top match, they DID match a 2nd/3rd cousin, so I added them to that column
NOTE: Trees were not used to match the 4th cousins to the appropriate Color Clusters. The sorts were based only on shared matches. This Color Cluster method is  a quick, visual way to see how your cousins are related.

If you give this method a try, please let me know what you think and how it works for you.

Happy Sorting!

10 comments:

  1. Thank you Dana for this terrific method--it has worked extremely well for me! I've been experimenting with those 4C matches who match only other 4Cs - no 2C or 3Cs. I followed your method and took a look at each of a 4C's shared matches. None of this 4C's shared matches matched to 2C or 3C. So then I tried to figure out how I might assign this 4C to one of the clusters.
    I ended up "leap frogging" from top shared matches back to a 4C who matched a 3C. So for example, let's call the 4C in question Mac. Mac's top shared match was Lane, Lane's top shared match was Van, Van's top match was Jim, Jim's top match was a 3C. I very tentatively assigned Mac to that 3C's cluster.
    I'm convinced this is way too attenuated to rely on. But since these 4C matches seem to be clustering over and over again in different permutations tentatively assigning Mac to a cluster might provide a clue when I start to examine the trees associated with some of the 4C matches.
    Thanks so much for all the innovative work you are sharing!


    ReplyDelete
  2. Alex, thanks for sharing your experience! I would like to work more with 4Cs who do not show a 2C or 3C match. How small are you going on cM when looking into 4C? I think there will be a point where it isn’t worth while, though I do not know what that number is at this point.

    ReplyDelete
    Replies
    1. I only worked with 4Cs greater than 20 cMs--I think the largest was 67 cMs--and then I just kept working my way down the list to see what might happen. I agree that number could get too low to be worthwhile--it already feels fuzzy enough!

      Delete
    2. Dana, how do you know what their top match is? Is it the one that comes up first when you look at Shared Matches? Thanks!

      Delete
    3. Hi. Yes, the top match is the one who shares the most DNA and, therefore, the greatest number of centimorgans (cM). AncestryDNA sorts them from highest or most shared DNA to lowest. So, the "top match" is the match that comes up first when you look at Shared Matches. You're welcome! :)

      Delete
  3. Hi Dana, I love your new analysis method. I only have 3 matches in the 2nd-3rd cousin category, but I thought I'd try it with my 'high' confidence 4th cousins. See my blog for the results. https://blackravengenealogy.blogspot.com/2018/08/colour-clustering.html

    ReplyDelete
  4. Hi, Dara. Thanks for sharing your experience when working with primarily 4th cousins. It really looks great! And, I hope you'll let me know what you find out as you work with the Green Cluster more. That looks potentially quite exciting!

    ReplyDelete
  5. Thanks. I posted this at the AncestryDNA Matching group on Facebook.

    ReplyDelete
    Replies
    1. Thanks, Jason. I've been meaning to do that!

      Delete
  6. I've found all your colour clustering posts interesting, but I'm not sure where to go from here. I'm trying to help an adoptee figure out who her birth parents were and I'm pretty stumped. I ended up with 3 clusters from her matches & could pinpoint 2 last names for each. But that still doesn't narrow down her place in the tree at all or separate out which parent is which.

    The most distant relative with a tree in my red cluster had a last name that matched a name for the tree of the closest match in the green cluster. I'm not too sure what to make of that.

    Guess I'm off to try to see if any other methods will help me.

    ReplyDelete

Color Clustering: Top 25 Fourth Cousins

For more on Color Clustering & DNA, please visit my new website at: www.danaleeds.com  For another look at how Color Clustering works...