Finding 3XGG with DNA: Splitting Genetic Networks

Melanie Mohler

Share with a friend: 

Finding 3XGG with DNA is easier when you use proven genetic genealogy strategies. Here, we see how splitting genetic networks and the Leftover Strategy are useful for finding mystery ancestors. But there was a key thing Clare did first that improved her chances of success!

Clare came to the DNA Skills Workshop with a question: Who were her 3x great grandparents, the parents of Robert Buchan, her 2x great grandfather? Clare only knew their names, Robert Buchan and Janet McRae, which she found on Robert Buchan, Jr.’s marriage certificate. This case study shows how she applied strategies taught in the Workshop to make progress toward answering this question.

Finding 3XGG using DNA

1. Using the closest possible generation

The first thing Clare did that boosted her chances of success was to use the DNA kit from her father’s first cousin, Pat, to help answer this question. This was an excellent strategy, as it reduces the distance back to that ancestor by a full generation! It’s a more reasonable task for autosomal DNA to help find Pat’s 2XGGs than it is to use Clare’s DNA to search for her 3XGGs.

We made this diagram with Lucidchart.

2. Splitting a genetic network

Clare realized during the DNA Skills Workshop that she had used a mishmash of several techniques to organize Pat’s DNA matches and needed to start over. She explained in the Workshop’s discussion board that she needed to split the descendants of her great-grandparent couple (Buchan-Bain) into separate genetic networks, one representing the paternal Buchan-McRae side and one representing the maternal Bain-Mathieson side. 

Clare started with Pat’s ten DNA matches at the second cousin level (these were second cousins once removed for Clare) who could help with this task. There were a lot of shared matches, so this part was easy. To label her AncestryDNA matches, she used the dot system Diahan Southard teaches, summarized in this video tutorial:

After sorting her matches, she could tell that several second cousins twice removed were definitely related to Robert Buchan. So then she tried to split off the unwanted Bain-Mathieson line of his wife Margaret Bain. ThruLines (AncestryDNA’s tree reconstruction tool) showed 19 matches descended from Margaret at this level, and four additional matches who descended from Margaret’s sister. (Other siblings didn’t show any descendant matches on ThruLines. There could be more matches who descend from the siblings, but who haven’t identified themselves as such in their own family trees, so they won’t be in ThruLines.)

One of the four additional matches was a known third cousin to Pat, so they should be ideal for finding descendants of the maternal line. ThruLines identified three more matches with tree relationships identified as a third cousin, a third cousin once removed, and a third cousin twice removed. When Clare looked at the shared matches for these four individuals, Clare found 66 that matched the Buchan-Bains and 37 that matched the Bain-Mathiesons. And 35 of these were also matching the second cousin matches of the Buchan-Bain couple. So she found only two people who didn’t share a dot for the paternal Buchan line. As she understood it, these were supposed to be her Bain-Mathiesons whom she could remove from the analysis so she could focus on Robert’s Buchan-McRaes. But there were only two, and neither had Bains or Mathiesons, except one in the wrong part of Scotland (and a tree with 35,000 people). 

Clare thought she’d made a mistake. “Just not enough Bain-Mathiesons have tested?” she wondered on the discussion boards. “Is it possible that there were just so few descendants anyway, let alone those that have tested?” She was afraid she had done something wrong.

3. Using the Leftover Strategy

Lori Napolitano, one Your DNA Guide’s genetic genealogy coaches, responded to Clare in the DNA Skills discussion board and assured her that she was on the right track:

Are you concerned that those shared matches of third cousins already have the dot from your large first second cousins group?  If that is the issue, then no problem! Most will have that first dot as they share one of the ancestor lineage (Bain) in that first couple.  Anyone with a Bain ancestor is going to likely match descendants of the Buchan-Bain couple, since they too are Bains.

The idea when splitting is to give a lot of people a second dot, hopefully half or more. The Leftovers group is now who you have left in that first group, who only has your original dot, ignoring all people with both dots.  By your numbers 66 in the main group, you now have 37 with a second dot – that leaves leftovers of 29 with just the original dot and they should be your Buchan group.  If this sounds right, then give those 29 a second new dot for your Leftovers Buchan group and take them into lesson 3.

You can also split that 29 leftovers group again if you have a fourth cousin descendant of one of the Buchan couple, or do “Bottoms Up” [another strategy taught in the DNA Skills Workshop] on the leftovers group to create smaller groups, hopefully for each person in the Buchan couple.

Lori’s reply had reminded Clare that she forgot to go back into the Buchan-Bains matches and re-dot those who did not match Bain-Mathieson as a separate group. Clare realized she should then focus on the leftovers from that group and call them Buchans.

The Leftover Strategy is typically used to find matches from unknown ancestors. Another Your DNA Guide genetic genealogy coach, Michelle Leonard, suggested that Clare also treat the Buchan-Bains as missing parents due to the absence of any records, except for the son (Buchan’s) marriage certificate in another country. 

Clare later reported that she dotted only the Buchan matches, which contained 14 non-Buchan-Bain people. She’s already been able to put five of these matches onto her tree as fourth cousin descendants of three siblings: George, Isabella and her Robert Buchan, all born between 1800-1810, according to matches’ trees and one document. That’s great progress!

Clare noted that she initially caught these five matches, but it was not done methodically. Now when she reaches out to these distant matches, she can be more targeted and specific about the connections she’s looking for between a small number of matches. She’s already had at least one person respond to her second message with a lot of interest and information!

Get Our FREE Guide on Contacting DNA Matches

You can DO the DNA, too! Get started by purchasing a copy of Your DNA Guide–the Book, which has step-by-step explanations for splitting genetic networks and the Leftover strategy. Or join the next DNA Skills Workshop, which uses Your DNA Guide–the Book as a textbook, but also includes tons of expert guidance, practice exercises and more! (Those discussion board conversations really helped Clare understand what she did RIGHT and where she went WRONG–so she could correct her work and go in the right direction of her relatives!)

Get the Book


Get More DNA Inspiration

Our free monthly newsletter delivers more great articles right to you.


  1. Victoria Jones

    I’m trying to sort out my dot system Diahan says each ancestral couple should have a dot my question is where do you start the dots from ? grand parents, great parents or later?
    Or does it not matter

    • Your DNA Guide

      Hi Victoria – It doesn’t really matter, it depends on your specific situation and what will help you answer your DNA questions. Ideally, you should mark your matches with the furthest back ancestral couple that you can determine.


Submit a Comment

Your email address will not be published. Required fields are marked *

Send this to a friend