If you’re of a certain age, you may remember Milton Bradley’s Chutes and Ladders game. Even if you don’t know it, its use as a metaphor to autosomal DNA research will not be lost on you. In this post, I demonstrate how to use Ancestry’s Enhanced Shared Matches to “ladder” back to an unknown ancestor.
Chutes and Ladders is a dice game of chance where one progresses up the board by climbing ladders and falls back toward the beginning on chutes (see the image below). The game is analogous to how we can use Ancestry’s Enhanced Shared Matches to review a group of shared matches to discover our unknown ancestors. The research strategy is detailed shortly but permit me to elaborate on the game association for greater comprehension of DNA analysis.
To construct the metaphor, the identity of an unknown ancestor represents the endgame for the Chutes and Ladders game. Within our autosomal DNA matches, we use shared segments between ourselves and our matches to discover the identity of an unknown ancestor who passed down the segment to us. For each child born to an unknown ancestor, the parent contributes 50% of their DNA. However, each of their children receive different segments but still totaling 50%. Which segments the child inherits in pure chance – akin to a roll of a die.
DNA Chutes
With each subsequent generation, the original DNA representing the unknown ancestor becomes diluted by the contribution of DNA from a spouse to create the next generation. Segments from the original ancestor become smaller or disappear altogether due to the randomness by which DNA is inherited. Whether a segment reaches you or another DNA tester is comparable to the chutes in the Chutes and Ladders game where not all chutes are of the same length. Sometimes a segment fragment proceeds down the chute to you or the segment is lost in an earlier generation.
DNA Ladders
If a DNA segment from your unknown ancestor is within your genome and it similarly “traveled down the chute” to another cousin who has tested, then you might be able to climb a ladder up toward your ancestor by building out your respective family trees to identify the most recent common ancestor. One match by itself may not be enough to take you as far back as you desire, but if you use other shared matches to systematically progress backwards one rung at a time, the ladder may take you to the end of the game, which is the identity of an unknown ancestor. Ancestry’s Enhanced Shared Matches, which is part of their Pro Tools, can aid in the progression up the genealogical ladder.
Laddering in 5 Steps
In an earlier post about viewed match switching, I found a related genetic network for my original Hill genetic network that included an unlinked family cluster with the surname of Matkins who were from Dorchester Couty, Maryland, which is where my Hill ancestors are believed to have also originated. An unlinked family cluster is is a large group of matches who all descend from a single ancestor but for whom you are unable to establish your genetic relationship. See the image below.
One of the shared matches in the Matkins genetic network was for an individual whose birth surname was Hill. For obvious reasons, my curiosity was peaked for this so-called “Hill” match. While building out the match’s tree, I quickly surmised that its Hill line was not likely related to my own as the geographies were all wrong. However, I recognized another surname in the match’s tree that I had previously seen in other shared matches’ trees within the same genetic network. This surname was Rhoads.
Step 1: Create the Ladder
To systematically investigate the possible Rhoads connection, I used Ancestry Pro Tools’ Enhanced Shared Matches feature. While viewing the “Hill” match, which I now call “Match 1”, I sorted the list of shared matches by centiMorgans (cM) with those sharing more cMs with Match 1 at the top of the list. This permits me to see those matches who are more closely related to Match 1 rather than my DNA tester. See the image below.
Step 2: Climb the Ladder
One match (or rung) at a time, I methodically review and/or build each match’s family tree searching for how they are connected to one another. To help me quickly review and/or build their family trees, I use the forest management tree triangulation and search principles discussed in an earlier YouTube learning module.
Sorting shared matches by those closer to Match 1 permits you to see how these closer matches connect to Match 1. It’s best to have a few of the shared matches who are first or second cousins (229-866 cM) to the viewed match (e.g., Match 1) as it is easier to determine how these matches are related because their common ancestor will be more recent within the first few generations.
If your viewed match does not have close shared matches, then it becomes more effortful to find the common ancestor, but not impossible. It is akin to having to climb a ladder two to three rungs at a time rather than one. If this is the case, it may be better to select another viewed match from which to begin climbing the genealogical ladder.
To effectively climb the ladder one rung at a time back to a most recent common ancestor among these close matches, you similarly progress one match at a time. Because the match list is ordered by shared cM, each subsequent match becomes more genetically and generationally distanced from the viewed match. That is not to say that each match on the list represents a different and separate generation from one another. That depends on the number of matches you have and how closely related they are to the viewed match. They could be brother and sister, or parent and child, or first and second cousins, etc.
If you look at the previous image, you can see that the match on the first rung of the ladder is projected to be a first cousin once removed to the viewed match, while the second and third rungs are both predicted to be second cousins once removed. So, each rung does not represent a separate generation where the common ancestor between the matches resides – although it could. Rather, as you progress further down the list (or up the ladder), the most recent common ancestor should become further back in the viewed match’s family tree.
Step 3: Construct an Unlinked Family Cluster Tree
After I review matches on the first several rungs for the viewed match and confirm the ancestral line to which the shared matches connect with the viewed match, I build a family tree showing how these close matches connect to one another. While I am quite adept with software, I find it easier to construct the tree first on paper. Later, if necessary, I can add it to my online family tree if I’m successful in connecting the “unlinked family cluster” to my tree or use Lucid charts for greater visualization if I cannot yet make the connection (see the image below).
As you can see in the above image, I was able to construct a basic (anonymized) family tree for the unlinked family cluster. What started out as a “Hill” match (Match 1) but later seemed to be a Rhoads match ended up being a genetic network for Edward (1713-1784) and Eleanor Rumley from Dorchester County, Maryland, which is where my prior research has suggested my Hill family also originated.
At the center of the image is Match 1, which is the viewed match. The matches spatially and genetically closest to Match 1 are in blue and represent the Rhoads branch of the Rumley unlinked family cluster. The predominance of these matches (numbers 2 – 8) was at the top of Match 1’s shared match list and formed the base of the ladder. The pink cluster is connected to the blue cluster several generations further back and therefore share fewer cMs with Match 1 and are located further down the shared match list with Match 1.
Step 4: Triangulate Results
Once a genetic network for an unlinked family cluster has been identified, its validity should be proven. Genetic genealogist Blaine Bettinger identified Two Hurdles[1] researchers must overcome to use the genetic network as evidence when applying it to their own family tree.
First, all members of the genetic network should be valid, which means all members share an ancestral path to a common ancestor. An additional check here is to ensure the identified ancestral path is the one in which the DNA tester, the viewed match, and shared matches collectively share together, and not another common ancestor. Researchers can use Ancestry’s SideView™ to ensure the match is associated with the correct parental side of the DNA tester’s family tree. You can also use viewed match switching.
Second, a triangulated or shared segment on a chromosome(s) among the DNA tester, the viewed match, and the shared matches should be identified. Many of us do most of our DNA analysis within Ancestry, which does not provide segment data. So, we need to find our Ancestry matches in other websites where segment data is provided, such as MyHeritage, FamilyTreeDNA, and Gedmatch. I’ve written a previous blog post for how to efficiently do this.
However, it is the opinion of other genetic genealogists that the second hurdle of segment triangulation is unnecessary. Diahan Southard reasons that the only concern genealogists should have when using a shared matches filter is that a group of shared matches (e.g., Match 1, Match 2, Match 3, etc.) share DNA.[2] However, the opinion of this author is the need for segment triangulation depends how you intend to use the genetic network. If the purpose is to guide your documentary research and the genetic network is one of several pieces of evidence in a proof argument, then segment triangulation is helpful but not required. However, if the genetic network is the only piece of evidence or the primary piece of evidence, then segment triangulation is necessary.
In my Rumley example, I have cleared the first hurdle, but I have yet to find any of the Rumley matches from Ancestry in other databases. However, I found a different Rumley match not in my Ancestry matches but in MyHeritage. It shares a 16.6 cM portion of the same segment (50018165-64418315) on chromosome 10 that I previously identified in my recent Gephi network graph post for the Clark, Keel, Harris, Linn, and Vickers unlinked family clusters associated with my Hill genetic network. The new MyHeritage match is a first cousin to Match 1 presented earlier and also matches other kits in the same Hill genetic network on MyHeritage. The second hurdle is indirectly cleared.
Step 5: Formulate a Research Plan
Depending on what you discover in the preceding steps, you will take one of two paths – additional DNA analysis or documentary research – or perhaps both.
DNA Analysis
If you’re up for additional work with your matches, I recommend viewed match switching. This entails selecting another match within the genetic network for the unlinked family cluster and viewing its list of shared matches for “new” matches that might provide further insight into how the cluster connects into your tree. The underlying logic here is that while all your matches match you, they won’t necessarily match each other because of the randomness to which we inherit DNA. So, the genetic network for the unlinked family cluster is actually larger than the list you see by viewing one of its shared matches – you need to visualize the shared matches of the shared matches!
To perform viewed match switching, review the current list of shared matches from the original viewed match. Select the largest match in the list in terms of shared cM with the DNA tester (not the viewed match). In my Rumley unlinked family cluster, I would select one of the matches having 17 cM such as Match 4, 7, or 8. However, because Match 4 and 8 are genetically closer to Match 1, I might gain greater insight if I select Match 7, which is more distantly related to Match 1. With Match 7, I’m more likely to find new matches you have not seen previously, which represents additional opportunities to learn new information.
Documentary Research
Your discovery of an unlinked family cluster presents an opportunity to review documentary records for where your family and the unlinked family cluster appear jointly in records or communities. This will help you learn more about your direct ancestors’ lives and experiences. If you’ve not been successful in linking the unlinked family cluster to your tree, documentary records can help you build the direct and indirect evidence for the hypotheses and proof arguments for how they are related.
What About Small Matches?
If you’ve spent any time working with DNA matches, you’re probably aware of the danger in relying on small matches – those with less than 15 cM of shared DNA. Many of my Rumley matches were below 15 cM. Should I be concerned that they are false matches, i.e., inherited small segments that are similar in composition by chance rather than by descent?
In my situation, concern is not warranted. Small matches with less than 15 cM are not problematic if they are part of a genetic network of shared matches who all share the same common ancestor. I’ve written about this in a past blog titled, “Small DNA Matches as a Compass in Genetic Networks”, noting that a small match must meet three conditions to be a valid match:
- The small match must be part of the same genetic network of shared matches being investigated (genetic membership), e.g., the Rumley genetic network.
- The small match must possess a most recent common ancestor with another kit in the network that is at least 15 cM or greater (conditionally linked to the genetic network), e.g., Match 1 (9 cM) and Match 4 (17 cM) in the Rumley genetic network.
- The small match should not be a direct lineal descendant of the larger 15+ cM match (i.e., child or grandchild) as it does not reduce the number of ancestral lines needed to be investigated. The match should be a collateral descendant, i.e., through another child of a great grandparent (collaterally linked).
All the small matches in the Rumley genetic network meet the above three conditions. They have genetic membership and are conditionally and collaterally linked.
Conclusion
While I have yet to identify the father of my ancestor, William Hill (1775-1836), I feel confident that I have identified the correct Hill family in Dorchester County, Maryland. The prior genetic network research identified two other lines who originated in Dorchester County – Matkins and McKeel (Keel). The newly identified Rumley unlinked family cluster is also from Dorchester and are part of the same genetic network as Matkins and McKeel.
Within the larger Hill network, to which Matkins, McKeel, and Rumley belong, are the Linn, Harris, Vickers, and Clark unlinked family clusters. The challenge before me is to figure out how all these lines all converge.
There is no silver bullet when it comes to genealogy research. DNA and documentary research is incremental. Before the discovery of the Hill genetic network with the Clark, Harris, Linn, Matkins, McKeel, Rumley, and Vickers unlinked family clusters, I was stuck in 1805 in Lycoming County, Pennsylvania where my ancestor William Hill (1775-1836) first appeared in records. At least now, I have new places and people to research who lived in the 1600s and 1700s!
Sources
[1] Bettinger, Blaine (2024, June 24). I know it isn’t popular, but you MUST be careful with shared matches below the 20 cM threshold (now visible with Enhanced Shared Matching at AncestryDNA). [Status update] Facebook.
[2] Southard, Diahan (2023 March 4). Shared DNA Matches – the only DNA Tool You will Ever Need. A presentation at RootsTech 2023, available at https://youtu.be/btQwRIyhuns?si=YNHgXbr5FUSGOW2r.