Last week, I looked at the current state of research into my ancestors – focusing on my assembled DNA evidence. (This is a modification of Yvette Hoitink’s original Level-Up Challenge for genealogy research). Here, I’ll walk through the steps I take to bring my genetic genealogy research to the next level. My goal for this process is to compile and analyze the DNA evidence I have pertaining to a given ancestor or ancestral couple, and at least write a summary of it. My eventual plan is to include genetic evidence (or a note about why it’s not there) in proof arguments for each ancestor’s connection with the generation before in my tree.
These are not necessarily the steps I would take to identify a completely unknown ancestor; rather, this is the process I use to turn hints from autosomal DNA matches into a stronger argument for a genetic connection to an ancestor I have already named. Even though I considered documentary evidence and DNA evidence separately in the two different Level-Up Challenge summaries I posted for my ancestors, these areas of research come together when I’m writing a research report about the identity of the parents of a given ancestor.
Here are the steps I take to level up the DNA evidence for the connection of an ancestor to their parents:
Step 1. Evaluate basic documentary evidence for this ancestor and their parents.
Even if I haven’t done reasonably exhaustive research, or if I know there are still particular records I need to track down, I want to make sure that my ties to this ancestor are documented back through the generations. I also want to make sure I have some decent compiled evidence about an ancestor’s origins, where possible. This can give me a good foundation for my hypothesis about an ancestor’s parents, which I can test against the DNA evidence.
Step 2. Identify possible DNA matches who might also descend from the proposed parents of this ancestor.
I use spreadsheets to keep track of information about my best DNA matches (“best” being somewhat arbitrarily defined as those that share the most DNA with me or those who otherwise seem helpful for my analyses). See the end of this post for a list of the columns I include in my spreadsheet. Information about these matches may get recorded in the spreadsheet in bits and pieces over time; I can refer back and fill in any missing details as I compile information about matches relevant to a particular line.
I’ll also go back to each testing site to look for new matches who I haven’t seen (or kept track of) yet. Often, if I’m looking for ties to a particular ancestral couple, I’ll start with AncestryDNA’s ThruLines and MyHeritage’s Theory of Family Relativity. With ThruLines, I can look for the matches suggested to be related to me through a given ancestor. I often make a list of the cousins shown, and then prioritize which ones I will try to understand better and possibly attempt to confirm a relationship with.
Outside of these algorithms, I also check the regular match lists at each testing company (some of my most useful matches do not show up in ThruLines, if their test is not linked to a tree, for example). I focus my efforts by looking at shared matches with known cousins also descended from the ancestral couple of interest. Matches that I have in common with those cousins may be worth exploring in search of evidence supporting a genetic connection with a shared set of ancestors.
Note – by taking this step, I do not assume that the presence of these cousins on my match lists is proof of a connection with our suspected common ancestor(s), but this lets me highlight some good candidates to investigate further.
Step 3. Thoroughly examine possibly relevant matches
For each match who has been added to my list of possibly relevant cousins (Step 2), I start by checking their family tree. If I can find a tree online for some (or all) of their ancestors back to those who they share with me, I will certainly start with what research has already been done. If their tree does not reach all the way back to our common ancestors, or if their tree has little or no information or sources included, I will likely try to research or confirm their connection to our shared ancestor myself, seeking documentation to support each generation’s connection to the one before it.
If/when I find the person or couple who appears among both the match’s ancestors and my own, I next check to see what our supposed relationship should be, based on the paper records (1C2R, 2C, etc.) I check the amount of DNA we share to see if it is consistent with our possible relationship, using the Shared cM Project tool at DNA Painter. Often, I can also compare the amount of DNA that this match shares with other known testers among my relatives. This gives me some additional confidence in the proposed relationship, if all of the known testers fall into the expected range(s) of DNA shared with the match.
If there is a reasonably sound paper trail documenting a connection between this match and our possible shared ancestors, and if the amount of DNA we share supports that level of relationship, I check the rest of their pedigree to see if there is likely overlap with other lines of my pedigree (see my earlier blog post about tree completeness). This is an important step in evaluating if this DNA match provides evidence of a shared genetic connection going back to our proposed common ancestor(s), or if it is merely evidence of some sort of genetic connection with the match, through one or more unknown lines.
I may not be able to fully research every match that could be relevant to a branch of my tree. Some matches share so little DNA with me that I do not make it a priority to understand our connection, especially if it appears that there are other matches who may provide stronger or clearer evidence. Sometimes I won’t be able to determine my relationship to a match (even if I suspect which part of my ancestry I might have in common with them). Or sometimes I can’t confirm all of their pedigree to rule out additional sources of shared DNA. If I can’t fully understand a match, I will likely leave them out of my compiled evidence for the time being, unless they present a conflict or there’s some other reason that I really feel that they need to be addressed. Like documents with illegible text, not every potential source of information can provide meaningful evidence.
Step 4. Map well studied matches onto a cousin chart
Once I have evaluated a match enough to determine that our connection seems likely to be through the common ancestors of interest, I map that match onto a chart of cousins who all appear to descend from this ancestral couple. Here are a couple of examples of cousin charts I have used to keep track of DNA matches, their documented lines of descent from our common ancestors, and their DNA sharing with me (or other known testers) and with other cousins on the chart.
This first cousin chart shows a simple example of how I keep track of data from DNA matches supporting a genetic connection between the three testers (1, 2, and 3, shown in green) and the great-grandparent couple at the top that they all have in common. Tester 1 is the individual whose DNA results I have access to; Testers 2 and 3 both match Tester 1 at AncestryDNA. The tree shows the three testers’ relationships (according to paper records). The table summarizes their relationship levels on the bottom: Tester 1 is a second cousin (2C) to both Testers 2 and 3, while Tester 2 and Tester 3 are first cousins (1C) to each other. The top of the table shows details of their genetic connection: Tester 1 shares 133 cM with Tester 2 and 125 cM with Tester 3; both of these amounts are consistent with a second cousin relationship.1 Since I don’t have access to the DNA results for Testers 2 and 3 at AncestryDNA, I do not know how much DNA they share with each other. They do, however both appear as shared matches with Tester 1 and the other of the pair, so, I do know that they share enough DNA to be considered matches. I could ask Tester 2 and/or Tester 3 for details of their matching, or ask to view their results, but I have not at this point. So, this graphic serves as a way to summarize the relationships of these testers, as supported by both documentation and by DNA evidence.
As an alternative way to display this information, I really like the setup of the McGuire Method. For me, it’s easier to start with just the tree and tables while the graphic is a work in progress; I might add in the McGuire labels as a later step. Here’s what this first cousin chart looks like, though, using the McGuire Method:
Here’s a second example of my cousin charts. This more complex graphic displays a larger group of cousins representing multiple lines of descent, of different generations, and with test results from multiple companies:
As in the first chart shown above, the tree diagrams the relationships of all of these cousins according to documentary evidence. The tables below summarize the DNA match results for those cousins and how they correlate with predicted relationships. As the tables show, all predicted cousins match each other as expected, either with the amount of DNA shared, or based on whether or not they match each other according to AncestryDNA’s algorithm. The only non-matching pair at AncestryDNA, Tester 8 and Tester 10, are predicted to be second cousins once removed; cousins at this level of relationship are not expected to share detectable DNA in all cases.2
While many testers from different families are included here, in most cases if a parent and child both appear on a list of DNA matches, I won’t include both of them in a cousin chart since the child’s results are unlikely to provide any additional information compared with those from their parent. In this instance, though, I do have one parent-child pair both depicted – myself (Tester 1) and my parent (Tester 2). I included both of us on here because while my parent has test results at AncestryDNA and MyHeritage (and is therefore the better one to use for analyses at those sites), they have not tested at 23andMe, but I have. So, my own results are shown to be able to include analyses with matches at 23andMe.
I like using charts like these as a way to check cousins’ relationships to each other and whether their shared DNA (or lack thereof) makes sense. These charts help me keep track of which branches have well studied descendants, and if there are any poorly supported connections, so I can determine if and how further research should be done into a particular ancestral line.
Step 5. Write up a summary of the DNA evidence and analysis
If multiple cousins have appeared as DNA matches covering multiple lines of descent from our common ancestors, and if no conflicts or unexplained connections have arisen (or any that have can be explained), then I write a summary of these findings in my research report regarding the ancestral connection in question. If any issues or questions remain for the DNA evidence, I would make a note of that in the report, too.
This analysis, when combined with other types of evidence, may or may not be enough to prove parentage. Just as the documentary findings and analysis in reports may have gaps (an original record yet to be retrieved, or a record set yet to search), the DNA evidence may or may not yet be strong enough to be “complete.” But if I do have at least decent supporting evidence from autosomal DNA cousin matches, and I can compile it, analyze it, and record what I’ve put together, I’ll at least know where I stand moving forward.
Columns for Match Spreadsheet (used in Step 2)
In my match spreadsheet, I record one match per row, with the following columns:
Source – DNA testing company for this match data
Display Name – the match’s name (or initials or pseudonym) as on the testing site
cM_Shared – the amount of DNA shared (in cM) with this match, according to the testing site. In my spreadsheets, I actually have more than just one cM column, with one column for each known tester who might share DNA with this match (I have columns not just for myself, but for other family members whose results I can access, such as parents and aunts/uncles/cousins where relevant).
Likely MRCA – MRCA = “most recent common ancestor”; I usually list the most recent ancestral couple I share with this match when I have enough information about them to enter this. Occasionally I might put a note about the likely line(s) we have in common if I just have a guess at this point.
MRCA Child & MRCA Grandchild (two columns) – These two columns name the child and grandchild of our common ancestral couple from whom this match appears to descend. (Note: Many of the items in this spreadsheet were likely inspired by or modified from similar spreadsheets I’ve seen online, but these two columns in particular I credit to Jim Bartlett, from the list of columns he shows for his Common Ancestor Matches spreadsheet.)
Relationship – Here, I list my relationship to this match, as suggested by the documentary research.
Trees Checked? – This is where I note whether or not I have checked this match’s pedigree thoroughly enough to rule out the possibility of them matching me through more than one ancestral line. Sometimes this will be noted as “part” or with some other caveat to remind me that this still needs work.
Name – the person’s real name if I’ve figured it out, and especially if it differs from their user name at the testing company’s website.
Contact Name – name (or user name) of the person handling the test, if it’s not the tester themselves.
Email – email address for the tester/test manager
Contacted? – If I’ve reached out to the match (or been contacted by them), I’ll either record the date I send a message, or maybe a brief note about the message.
Responded? – If I reached out to them and they (eventually) responded, I’ll record the date or a brief note about it.
Tree – I’ll paste a link to an online tree with their pedigree. If they have a thorough, well documented tree of their own, I’ll link to their tree. If not, I’ll link to an online tree I’ve generated for their family (kept private and unsearchable), with whatever I’ve found about their pedigree so far.
Notes – In this last column, I’ll include whatever other information seems worth keeping track of about the match. Many matches won’t have anything here. For some, if I have traced their pedigree back for only some of their ancestral lines, but the untraced lines all appear to be from geographical regions for which I have no known ancestry (and therefore it is unlikely that we would have shared ancestry from those lines), I might note this here.
1. Relationship estimates from “The Shared cM Project 4.0 tool v4,” DNA Painter (https://dnapainter.com/tools/sharedcmv4 : accessed 27 January 2022), 2C relationship.
2. Relationship estimates from “The Shared cM Project 4.0 tool v4,” DNA Painter (https://dnapainter.com/tools/sharedcmv4 : accessed 27 January 2022), 2C1R relationship.