Graphing DNA

My geminate Shelley from Twigs of Yore blog has been giving us all lessons in how to graph or group our Ancestry DNA matches. She’s done a great job of simplifying each step so it’s hard to make mistakes (but it occasionally happens, due to user-error).

In following this process I was lucky on a few counts:

  • I have only 64 matches for 4th cousins or closer
  • I have readily identifiable cousins in the short list at 2nd or 3rd
  • Overall I have about 1000 matches.
  • My ancestors mostly come from different countries or regions so I expected little overlap between groups/graphs.

As a result I have yet to need to do much to play with graphs which are crazy busy with lines. As yet, I haven’t updated my matches download, so I haven’t tackled the deletion of duplicates. I decided to take the process one step at a time.

Having followed the process, I ended up with 12 graphs in my screen (see below). There are another 12 as well which I can focus on, but they have only one or two linkages, and no identified cousins, so I’ve left them for the time being.

Graph 1 DNA

A quick glance showed me some decidedly interesting clusters within particular graphs connecting cousins who I know to be on particular lines. I decided to use the Kunkel graph as my example here, partly because I’ve followed up some of the connections and because I had known cousins in the mix.

KUNKEL DNA Matches relnships2

With Shelley’s guidance I removed names from the graph, cut and pasted the graph into Photoshop, and added some relationships. For these lines to link up, I’m assuming (yes, I know!) that they belong to one of my Kunkel lines but it’s important to realise that some links might be through the other surnames on that line: Happ (Dorfprozelten, Bavaria) Gavin (Kildare), Murphy (Wicklow), O’Brien and Reddan (Clare). Is this a logical assumption?

The known 2C and 3C cousins in the single graph above are on different branches, descended from George Kunkel and Mary O’Brien. The a2C cousins are on my Kunkel-Gavin branch while the b3C cousins are on the Lee branch, and the c3C are on another branch. Interestingly some of those intersect with matches who seem to be descendants of my Mary O’Brien’s sister, Bridget Widdup from New South Wales.

I find it fascinating how the DNA “lottery” varies so that some will match while others don’t. Similarly some link to the cluster in the coloured lilac area (more anon). What I need to remember is that they also match me, since these linkages derive from my results. What the graphs introduce are links which may be weaker for me and stronger for other cousins. Shelley reminds us that when we search ICW on a match, it only gives up to 4th cousins on that match. These graphs extend the links beyond that.

I’m most intrigued by the lilac cluster, all from the USA as far as I can tell. Many include the surname Kunkel, though unfortunately many do not know where their Kunkel ancestors were born. As you can see from the size of the dots there are strong links with a couple of these in particular. Despite emailing and working on their trees I’m still no wiser about where they fit into my Kunkel family but it seems inevitable that they do, because of the geographic separation. It seems likeliest that they tie to the Kunkels who lived in Laufach or Neuhütten in Bavaria, where my own earlier ancestors came from, and it fits that they would be at the 5th cousin or upwards range. I am fortunate that I should be able to identify relevant 4th cousin families – provided they are shown on a match’s tree.

Has this helped me? Yes, I think it has, because it’s identified where the strong links are. It also lets me target matches who I might otherwise ignore because they’re too far down the match ladder. The clustering with known cousins on particular lines gives the researcher confidence that they are focused on the correct area of their tree.

As always it’s not easy when the matches you want to look at have private trees (or don’t respond), no trees or minimal trees, or when the background information just can’t be found easily. I do feel some sympathy for American researchers because with Australia’s semi-centralised civil registers, it generally makes it easier to track ancestry (more assumptions behind that). On the other hand the US has decennial census records (apart from that very annoying 1890 census) and wider naturalisation information. Swings and roundabouts I guess.

I wonder how often people use the local church records to find where their ancestor may have come from – if the register even states that. Without that one strategy I’d never have found my Kunkels in Bavaria. How appropriate that I wrote that post in response to a geneameme by Shelley all those years ago!

Thanks Shelley for coming up with this bit of Excel magic to help us out. Thanks also to my cousins who’ve tested, either at my request or off their own bat.