Small DNA Matches as a Compass in Genetic Networks

We’ve been told not to rely on segments/matches having small amounts of shared DNA. This warning is true, but in genetic networks, small matches can act like a compass and direct us where to research.

I’m not talking about small matches below 6 to 8 cM in length. Testing companies like Ancestry, FamilyTreeDNA, and 23andMe do not display matches in our test results below this threshold. Many of these matches are considered false matches.¹ While the origin of false matches is outside the scope of this article, interested readers are encouraged to watch the companion YouTube learning module on using Small Matches as a Compass in Genetic Networks where false matches are discussed.

Rather, I’m referring to matches below 15 cM (but above 6-8 cM threshold noted previously). Matches at or above 15 cM are what genetic genealogist Blaine Bettinger calls the “safe zone” where he estimates 99.3% of matches are likely valid cousins.² Bettinger’s assertion is also supported by a FamilyTreeDNA white paper.³

So, when are these 6 to 15 cM matches useful in genealogical research? Despite their small size, these matches can be a big help when evaluating a group of shared matches within a genetic network. Genetic networks are a group of matches that the test taker shares with another match where most of the shared matches in the group have a common ancestor. See my last blog post for examples of genetic networks and how they are used in genealogical research.

Analyzing Family Trees within Genetic Networks
One of the challenges we face when reviewing our matches’ family trees is determining which of their ancestral lines is where our common ancestor lies. Nowhere is this more important than when reviewing an unlinked family cluster within a genetic network. An unlinked family cluster is a large group of matches who all descend from a single ancestor but for whom you are unable to establish your genetic relationship.⁴ If you could find the connection, you likely could open up new areas for research and potentially break down brick walls.

One of my more active research projects is to establish the lineage of my Boyd line back to the Earls of Kilmarnock in Ayrshire, Scotland. The last Earl was William Boyd (1705-1746), who supported the Jacobite Rising to get a Scot on the English throne.⁵ Y-DNA results for my Boyd line indicate we are part of the same haplogroup (R-FT102166) as the Earls of Kilmarnock, but the exact connection to them remains elusive.

When reviewing the autosomal DNA matches for one of my cousins who descends from our common Boyd ancestor, William Boyd (1753-1836), I discovered an interesting genetic network of 23 matches. Within the genetic network is a small unlinked family cluster of Benson descendants as well as several other matches that have no obvious connection. These latter “unknown” matches might be the missing link I need to learn about my Boyd ancestry, but how do I know which of these matches’ ancestral lines to spend my precious time investigating? I need a compass to point me in the right direction.

Case Study: The Boyd Genetic Network
James Boyd, my 6th times great grandfather, resided in Bucks County, Pennsylvania in the 1750s. In 1751, James Boyd witnessed the will for Henry Benson.⁶ Somewhat recently, I discovered a 30 cM autosomal DNA match to a descendant of Henry Benson’s probable brother, Thomas Benson (1725-1793). When I reviewed the shared matches between the descendant of Thomas Benson and my cousin, whose DNA matches I was reviewing, there were 22 other matches (see the table below, which includes the 22 matches and the original Benson match).

Including the original Benson match, five matches in the genetic network descended through two different children of Thomas Benson (i.e., the Benson unlinked family cluster). Nine descended through three different children of William Boyd, who was the son of James Boyd. The observation that so many different child lines are represented for both families provide good evidence for a genetic link, but the nature of the relationship is undetermined. The image below presents the largest matches for each of the separate child lines.

Of the remaining nine “unknown” matches, four had private trees, two had no trees, and three had trees. For the matches with trees, I was unable to establish a connection to either the Bensons or Boyds, which tells me that perhaps one of these three matches might help me discover a new Boyd or Benson cousin to bridge the gap between Pennsylvania and Scotland.

Unfortunately, because I’m investigating ancestors who lived in the 1700s, each of these “unknown” matches with trees each have potentially 96 ancestral lines at the 4th great grandparent level where the Boyd (or Benson) connection may reside. To make my research more efficient, I need to determine if any of the three matches have some lines in common that could act as a compass and help me narrow down the ancestral line where I need to research further. I need more matches to help evaluate the ones I have.

Finding More Matches Using DNA Gedcom Client
I’m not sure about you, but most of my time analyzing DNA matches is done on Ancestry. It’s easy to use and there are more matches with usable trees than other sites. The problem is that Ancestry only displays shared matches having 20 cM or greater. While I appreciate Ancestry’s decision not to show matches that can be distracting (i.e., potentially false matches), the 20 cM cut off feels arbitrary given other evidence indicating most matches down to 15 cM are valid.⁷

Enter DNA Gedcom Client, which enables you to download your match list, segment information, and tree data into a spreadsheet for offline analysis. For a nominal cost and downloaded software on your computer, you can view all matches within your genetic network, even those between 8 cM and 19.9 cM (Ancestry only captures matches above 7.9 cM). View the companion YouTube learning module to see how to use DNA Gedcom Client.

After downloading my cousin’s full match list, I discovered an additional 37 shared matches between 8 cM and 19.9 cM within the same genetic network between my cousin and the original Benson match. In addition to analyzing these new 37 matches, I also reviewed the shared matches between my cousin and the other three matches previously mentioned having family trees posted but for whom I could not establish a connection. This produced an additional 88 matches to evaluate.

For the new matches with trees or for those I built out their trees for them, I look for two items. First, can I connect them into the two family lines I’m investigating – Boyd and Benson? It would be great if I can identify another child line for either of these families going back to the respective common ancestor (i.e., William Boyd or Thomas Benson), which would increase the confidence in the genetic connection.

Second, if I cannot connect these new matches into my investigated Boyd and Benson lines, can I connect them into one of the ancestral lines for the other matches? In other words, can I find a smaller “micro” unlinked family cluster of related matches within the overall Boyd-Benson genetic network who share a common ancestral couple but for whom are not in my Boyd tree nor in the Benson tree?

Identifying the common ancestor among this smaller group of shared matches may help me narrow down on which of the ancestral lines to concentrate my future research efforts. For example, taking my match’s 96 possible ancestral lines at the 4th great grandparent level down to one or two.

Identifying a Micro Unlinked Family Cluster within a Genetic Network
Within the expanded genetic network made available through the DNA Gedcom Client, I found one match in particular that opened a new research door. The match was only 9 cM in length, which is clearly below Bettinger’s safe zone. I would never use this match alone as genealogical evidence, but in conjunction with one of the stronger matches (21 cM), I found a common ancestral couple among the two matches that formed a micro unlinked family cluster (see image below).

The ancestral couple was for Joel T. Stewart (1818-1889) and Sarah Ann Price (b. 1813) from Bucks County, Pennsylvania. In fact, Joel T. Stewart was from Warwick Township, which is adjacent to where my Boyds lived in Northampton Township.

With this information, I was able to narrow down the possible 32 ancestral lines of the 21 cM match’s family tree (at the 3rd great grandparent level) down to two lines representing a 96% reduction in future research effort! The discovery of the 9 cM match also helped to reduce my fears of confirmation bias by highlighting where I should research rather than where I want to research. The 21 cM match’s Bucks County ancestral lines were tempting, but I wanted the data to point me where to research, not the shiny object.

Summary
Progression in genealogy research is incremental, and it requires persistence. I have not immediately learned more about my Boyd line, but I have discovered a new micro unlinked family cluster to research – Joel T. and Sarah Ann (Price) Stewart. Often, brick walls block our paths forward, but if we dig a little deeper and discover a few more relevant DNA matches, we just might find another way around the wall. This is what I’m hoping happens with my Boyd research and the Stewart-Price family micro cluster.

For me to use matches between 6 cM and 14.9 cM in my research, there must be an indication that the small segment/match is valid, or what I call “collaterally and conditionally linked genetic membership”. That is, the small match must meet three conditions:

1. The small match must be part of the same genetic network of shared matches being investigated (genetic membership), e.g., the Boyd-Benson genetic network.
2. 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., Joel T. and Sarah Ann (Price) Stewart.
3. 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).

If you’re working in Ancestry, it’s also important that the small match is associated with the correct parental side of the tester’s tree as the ancestor being investigated (i.e., using Ancestry’s SideView™ technology). It is also worth repeating that I never use a kit between 6 cM and 14.9 cM by itself as genealogical evidence.

If you would like to learn more about using the DNA Gedcom Client software to discover other relevant matches, please watch my YouTube learning module on Small Matches as a Compass in Genetic Networks. If you want to learn more about how to efficiently review the family trees within a genetic network, please watch my YouTube learning module on Forest Management, which relies on “fertilizing, pruning, and harvesting” of genetically relevant information in your matches’ family trees. Associated with this project, I also recently appeared as a guest presenter on Family Locket’s Research Like a Pro® Webinar Series, where I discussed how I discovered William Boyd’s (1753-1836) father using only indirect evidence, which included DNA analysis. The research report on William Boyd’s father is available for viewing and downloading on my website.

---

---

 

Sources
1 – Bettinger, Blaine (2022), An In-Depth Analysis of the Use of Small Segments as Genealogical Evidence. Accessed 10 October 2023 at https://thegeneticgenealogist.com/2022/08/07/an-in-depth-analysis-of-the-use-of-small-segments-as-genealogical-evidence/. 
2 – Bettinger, Blaine (2017), The Danger of Distant Matches. Accessed 10 October 2023 at https://thegeneticgenealogist.com/2017/01/06/the-danger-of-distant-matches/. 
3 – Hu, Rui, et al. (2021), Family Finder Matching 5.0: Matching Algorithm and Relationship Estimation, White Paper 2021-08-18. Accessed 8 October 2023 at https://blog.familytreedna.com/wp-content/uploads/2021/08/Family_Finder_Matching_WhitePaper.pdf.
4 – Bettinger, Blaine (2023), The Growing Phenomenon of Unlinked Family Cluster. Accessed 17 October 2023 at https://thegeneticgenealogist.com/2023/03/16/the-growing-phenomenon-of-the-unlinked-family-cluster/.
5 – Wikipedia (2023, July 20), William Boyd, 4^th Earl of Kilmarnock. Accessed on 10 October 2023 at https://en.wikipedia.org/wiki/William_Boyd,_4th_Earl_of_Kilmarnock. 
6 – Bucks County, Pennsylvania, estate file, no 789, Henry Benson (1752), Recorder of Wills, Clerk of Orphans’ Court, Doylestown.
7 – See Bettinger (2017) and Hui et al. (2021).