BackgroundAll living persons and test subjects have been anonymized for privacy reasons.
In this particular study, AncestryDNA performed all DNA tests.
Although the tool itself was not employed in this investigation, DNAPainter.com's What Are The Odds (WATO) tool was used to generate the horizontal inheritance diagrams featured throughout this study.
Abstract
Progress on long-standing genealogical brick-walls can take many forms — often involving happy accidents, dedicated research, and inspired discoveries. Occaisionally, multiple approaches to achieving the same goal may present themselves. This case study documents how I identified a familial connection to my Great-great-great Grandmother, whose pedigree remains uncertain, despite being unable to follow through on my original research plan.
Introduction
Every genealogist eventually runs smack into the dreaded "brick wall" ancestor — most pedigrees have several such roadblocks — and being the sort of researchers who seldom take "no" for an answer when confronted with little documentary evidence and a maddeningly common surname, my Shaw cousins and I turned to DNA in the hope of making progress on our family tree.
Of course, "progress" is seldom as straightforward as just taking a DNA test and waiting for the results. This case study will document how I employed Discovering New Ancestors' CMA process to generate AI, or "Actionable Intelligence" as a basis for making significant connections towards solving an intractable problem. A full discussion of CMA is beyond the scope of this study, but instructional materials and downloadable resources may be found here.
CMA uses the DNA matches of a nexus individual (labelled as A) and the DNA matches of known family members (B, C, D, etc.) to identify individuals with significant connections to the nexus' extended family. Let's begin with a diagram of members of my Shaw family who have had their DNA tested at AncestryDNA:
Note the three brick-walls in the diagram:
- brick-wall A: the ancestors of my GGGG Grandparents Farquhar Shaw and Jenny Mackintosh
- brick-wall B: the ancestors of my GGG Grandmother Susannah Jones
- brick-wall C: the ancestors of my GG Grandmother Catharine Mardell
(note that I'm not personally shown on the roster of DNA test takers, as I'm descended from a half-cousin of A and B).
The slim prospects of finding numerous DNA matches above 40cM for the descendants of any of these brick-wall ancestors makes using probability-based tools something of a non-starter. However, the CMA process, which employs axiomatic set theory to organize and prioritize actionable individuals, may still be able to provide us with answers.
Preliminary Analysis via CMA
CMA uses set theory to organize and tabulate the full sets of DNA matches for up to 26 individuals. The CMA Master Workbook allows us to sort this roster of individuals by average shared linkage, membership in ancestral genetic complexes, and by the number of test subjects an individual shares DNA with.
The image at right ranks individuals according to the number of Shaw family test subjects they match (Member names have been abbreviated for privacy). The ACS Classification lists the letter names of the Shaw family test subjects which share DNA with the abbreviated individual.
Member names are color coded:
- Black font applies to Shaw family DNA test subjects.
- Blue font signifies known Shaw family members not included among our test subjects. (For instance, D.M. is JM's son)
- Green font signifies an unknown individual.
So our Shaw family CMA has identified individual JA as the unknown individual most widely connected to our Shaw test subjects.
In terms of "Actionable Intelligence" this information qualifies as "intelligence" (in the CIA sense) in that it's something we didn't previously know — but to what degree is it "actionable?"
Further Analysis
Messaging Ancestry members ranked lower in our ACS list revealed that JA's son (DA) and first cousin (CW) also share DNA with our Shaw family test subjects, so we were able to construct an inheritance diagram of their family group (AncestryDNA ProTools would facilitate this discovery in the present day):
The cM linkage shown is the amount shared with Shaw family member GSM (test subject P).
Because the shared DNA in JA's family group falls below a 40cM threshold, Tools such as DNAPainter.com's What Are The Odds tool won't be useful to us in this particular case. Further, the Most Recent Common Ancestor (MRCA) of this Rivron family group does not suggest a connection to our Shaw family — and the MRCA of this family group is realistically too recent for such small DNA matches to connect to our Shaw test subjects through an NPE.
CMA also assigns a Genetic Complex to JA based on the Shaw family test subjects that JA and her family group match. In this particular case, JA was assigned to the [Shaw-Jones] genetic complex — which is to say that (referring to the diagram of our Shaw family test subjects) JA and family are either direct descendants of one (or both) of my GGG Grandparents John Shaw and Susannah Jones or share a common ancestor with one or the other of John or Susannah.
Because JA and her family don't share significant linkage with any of our Shaw family test subjects, it's extremely doubtful they could be direct descendants of John and Susannah — and since we know that John Shaw was born in Scotland and none of JA's family has any discernable Scottish ethnicity (and their recorded pedigree doesn't suggest any such connection) — it's reasonable to conjecture (at least for the time being) that JA's connection to our Shaws is through Susannah Jones' pedigree.
So, just where do we go from here?
Analytic AlternativesDespite the low centiMorgan linkage of JA's Rivron family group, a number of analytic alternatives present themselves as we attempt to connect JA to our Shaw family:
- Perform a "Reverse CMA"
- Use Shared Match predigrees
- Find a "reciprocal JA"
1. Perform a "Reverse CMA" — my original research methodologyThe CMA process used the DNA matches of our nexus individual (Shaw family group subject A) and the DNA matches of related test subjects (Shaw family group subjects B through P) to identify JA as an individual significantly connected to our Shaw family, most probably through the ancestors of Susannah Jones. CMA does this by evaluating the Most Recent Common Ancestral Couple (MRCAC) shared by subjects B through P with our nexus A.
If we were able to download the full set of JA's DNA matches, we could begin a "reverse CMA" by designating JA as our nexus individual, and using the DNA matches of Shaw family test subjects A through P as family members connected to JA through a "mystery MRCAC".
Although JA does not match every Shaw family subject A through P, the CMA process would provide us with a subset of JA's DNA matches which also match our Shaws. We could then build up and/or examine the pedigrees of this subset, beginning with those individuals sharing greatest DNA with JA, in order to identify which of JA's grandfather's ancestral lines are common to our subset of Shaws, and hopefully identify a connection with our [Shaw-Jones] family.
Unfortunately, although JA's connection to our Shaw line was first identified in late 2021, and we have reached out to JA, we yet to receive a reply from her, so this approach is something of a non-starter for the time being.
2. Use Shared Match pedigrees
A survey of the individuals listed as "Shared Matches" between JA and each of the Shaw family subjects might generate a list of candidates whose pedigree might connect one of JA's grandfather's ancestral family lines to our Shaw family. Unfortunately, JA's Shared Matches with each Shaw family test subject included only known members of our Shaw family.
AncestryDNA's ProTools Extended Shared Matching provided us with the lower cM matches of JA's son DA and cousin CW, but nothing in the way of further "actionable intelligence" — so our efforts to identify which of JA's grandfather's ancestral lines might connect with our Shaws remained unfulfilled.
3. Find a "reciprocal JA" — file this solution under "never stop testing; never stop looking"
Back in 2021, when JA's significant connection to our Shaws was first identified, our roster of Shaw family test subjects only went as far as subject I (individual AS) and, while JA clearly shared DNA with a broad range of our Shaws, there was nothing to suggest that any of our Shaw family test subjects shared anything similar with JA's family — as JA only appeared as a low-cM Shared Match amongst other known Shaw family members.
This changed when we were able to obtain the DNA test results of Shaw family subject P (individual GSM) in September of 2024. In addition to a 21cM match with JA, GSM's DNA matches included a 92cM match(!) with individual DC, whose Shared Matches with GSM inluded several heretofore unknown individuals whose pedigrees converged in the ancestral lines of the Bowman family, descended from Christopher Bowman (1788–1871) and his wife Christiana Jones (1787–1871):
From her dates and place of birth (Westminster, London) Christiana Jones would appear to have some significant connection to my GGG Grandmother Susannah Jones (1781–1859), who married in Westminster in 1800.
With the ancestral relationship of these shared matches established, it became easier to attach additional family members to our Bowman family, including JA's family, as her grandfather's mother was a Bowman:
With the ancestral relationship of these shared matches established, it became easier to attach additional family members to our Bowman family, including JA's family, as her grandfather's mother was a Bowman:
- In the above diagram, horizontal lines of the same color (black, magenta) indicate full-siblings; green and yellow connections are half-siblings.
- Individuals in the diagram have been color coded as follows:
- Beige individuals share DNA with Shaw family test subject GSM (linkage to GSM is shown), in addition to other Shaw test subjects.
- Green individuals share DNA with Shaw family test subjects, but not with GSM.
- Red individuals are Shaw family test subjects.
Observations
Although other relationships are possible — for instance, Susannah and Christiana Jones could be paternal first cousins — at the present moment, I'm inclined to treat them as sisters, particularly because of their geographical proximity.
With regard to our Shaw family test subject group, it's interesting to note that even though we'd previously tested GSM's sister, CSM, her DNA test results in no way suggested the sort of connection to the Bowman family group that GSM's results made abundantly clear — so the utility of "never stop testing" cannot be overemphasized.
The 92cM match between GSM and DC — with a 4th Cousin 2x removed relationship — places the match in the top 0.3% of reported cases (per the Shared cM Project), suggesting that DC may have a connection to GSM somewhere on her maternal lines, in addition to her paternal connection through the Bowman family. Since DC appears to share an elevated amount of DNA with a cluster of individuals belonging to neither the Bowman nor Shaw families, it's entirely possible this supplementary ancestral connection is through a different Jones sibling.
Had we been able to aggregate the In Common With matches of JA and our Shaw family test subjects using a reverse CMA process, we very likely would have obtained a similar grouping of Bowman family members through which we'd connect to our Shaw family group, but that remains something to attempt at a later date.
Conclusion and Future Considerations
With DNA matches involving a total of 42 documented individuals between the Bowman and Shaw familes, the significance of the relationship between these two family groups appears assured.
A further collection of DNA matches, shared by both Bowmans and Shaws but without significant linkage to either group, suggests the existence of one or more as yet unidentified Jones siblings whose descendants may assist in identifying the parents of Susannah and Christiana Jones.
Can Discovering New Ancestors solve your genealogy mysteries?
Read what our clients say — or download our Intake Form!