I use DNA almost every day. I use it to research distant ancestors where research has hit the brick wall, to help people who have been adopted or have other mysteries in their family, and to generally enhance my research. Yet, I really struggled with what to write about DNA. I could write about a given case, interesting finds, etc. Finally, I decided to share an overview of my approach to digging deep into DNA when I am looking for someone several generations back with a brief look at each step using my research to find my great-great-grandfather Lemuel McCracken’s parents.
Define The Questions
I find that it is important to define very specifically what I am trying to find. And, writing it down is useful to keep the research on track. It might seem like an unnecessary step, but it is easy to get distracted following many different tangents.
Lemuel Case Study:
The search for the parents of my Great-Great-Grandfather Lemuel McCracken has been going on for decades. When I started working on finding Lemuel’s parents, all I had were family stories and a handful of records. Through research I learned many more pieces of information, but I still didn’t know who Lemuel’s parents were or even where he was born. He had simply shown up around 1850, seemingly alone, as a young man in Lee County, Iowa. From there, more of the story of his marriages and children have been discovered using traditional genealogy. That left me with two questions: Who were his parents? Where was he born?
Logic Problem
I approach genealogy brick walls as logic problems. I write down the facts that I know about the person, what I don’t know, and any reasonable assumptions. When beginning a case, this is based on traditional genealogy research and family stories. Thus, I include items, such as, “The family believes . . . ” I find it important to give ownership to each item since family stories aren’t facts, but can be hints. And, records can be wrong. As I move through the process and learn more, I add (and sometimes remove) statements.
Lemuel Case Study:
I started my DNA search with the following data and assumptions:
-
- Records indicate Lemuel was born in 1831.
- According to records, Lemuel married in Lee County, Iowa in 1854.
- Records imply Lemuel was in Lee County, Iowa as early as 1850.
- Records indicate Lemuel was born in either Ohio or Pennsylvania.
- Family stories indicate Lemuel was a Campbellite (Christian Church).
- Etc.
Testing, Testing, Testing
Depending on the situation, who and where you test may be different.
Y-DNA
When looking for a male ancestor following an all male line, so doing Y-DNA testing is a logical step. Even when I am looking for both parents, I will follow this step if it applies as learning the father is one part of the question.
Lemuel Case Study:
From Lemuel to my dad is an all-male line. Thus, Y-DNA testing was an option for determining Lemuel father. So, both my dad and the son of one of his brothers did Y-DNA tests with FamilyTree. Each joined the McCracken project as well as other projects related’s to Ireland and Scotland, where the family is believed to have lived prior to the United States. The projects agreed that my cousin and my dad are of the same family. But, they did NOT match any other McCrackens.
This implies several possibilities:
-
- No one in our branch had taken the test.
- Perhaps Lemuel’s father was not a McCracken. Did he make up that name? Was that his mother’s name?
- Assuming that his father was a McCracken, it eliminated several major McCracken families from consideration.
Autosomal DNA
Most of my work with DNA relies on autosomal DNA (e.g. Ancestry, 23andme), which provides a lot of flexibility for testing and analysis (e.g. both males and females can test and match any family line).
When trying to find people back several generations, I prefer to test multiple people if possible. Since DNA inherits unevenly each person will have some different DNA matches. Additionally, they will match them at different levels, which gives a range for the match to the person, which can be very helpful. When possible I try to have people that descend from multiple of the earliest known person’s descendants. This allows me to narrow down matches more quickly and also increases the possibility of people inheriting different pieces of DNA. However, it is possible to find a distant ancestor using only matches from one of their children.
For this type of analysis where you are working with known persons, I have found Ancestry to be the most useful, followed by 23andme, MyHeritage, and Gedmatch. When starting with a known person and looking for their ancestors, shared matches are the best place to start. Chromosome browers, what are the odds, and other tools can be useful for detailed work.
Lemuel Case Study:
I have access to 24 different people’s DNA tests on Ancestry, 23andme, MyHeritage, FamilyTree, and Gedmatch with the largest number on Ancestry, Basically, I kept asking anyone of my generation or my dad’s generation to test and asked people who had tested if they would share their DNA results with me. Those 24 people descend from four of Lemuel’s seven children.
The Analysis
Once I have even one test in hand, I go to work analyzing DNA matches. Some of the techniques that I use include:
- I make sure each person that tests matches to the family in the expected manner. If a person matches with DNA that is out of bounds for their relationship, this must be investigated before their DNA can be used in the analysis.
- I create clusters of people that are related to the specific line of research. I manually plot relationships as tools for clustering that I have seen to date aren’t designed to work for solving this type of problem where you are looking some distance back and want only certain people, but where relationships may need to be manually analyzed.. Additionally, since I never seem to find simple clusters that are completely independent, I often change how the people are clustered to be able to analyze things differently.
- Within clusters, I first look to see if I can determine relationships between any of the people in the cluster. Second, I use the Leeds Method, visual scans, and other techniques to look for a name that connects the people. Sometimes I add them to my tree and do more research on them to connect them. If no connection emerges from the group, I re-evaluate the group and look in detail at the connections between different people in the group. When the groups don’t yield answers, I do special deep dives as if they were the puzzle I am solving.
- I make lists of names that show up frequently where I haven’t been able to fit them into the puzzle. Having the lists gives me a reminder that I have seen them before and notes about them can help me connect the dots.
- I record all descendants of the person in question who have done DNA testing. I show the direct line back to the person of interest. This allows me a quick visual of how the descendant DNA matches fit together.
- I record the groups of people showing the direct line back along the line where they connect (e.g. if I am looking for a Smith connection and they connect to a Smith, I will record that direct line). Occasionally, I have disconnected groups where I can connect several people back to an earlier ancestor, but where I haven’t yet been able to connect them to any other group. I do this in a visual chart so that I can see everyone together. I find this more effective for seeing gaps and possible connections than simply having these people in a tree.
- When I have a lot of information, I create summary charts to pull pieces of the information together in a compact way to share with others or for additional analysis.
- When I am not seeing anything new, I do what I call “turning it on its side and upside down.” What this means is that I revisit information. I review the facts and assumptions, the DNA matches, and other notes. I may even take work that I have done and set it aside and do it again to see if looking at it with a fresh perspective helps. I will also consider if there is a different angle for analyzing the data. I also ask myself if there is some additional traditional research or additional genetic genealogy tools that I can apply.
Lemuel Case Study:
Direct Descendants
I have found nearly 80 descendants of Lemuel, many that I didn’t know about, through this process. In addition, it led me to 10 descendants of his wife Louisiana and her first husband. This includes some surprises, such as, a pair of siblings that had no idea that they descend from our family line.
McCracken
I determined that Lemuel’s descendants are related to 32 descendants of James McCracken and Rachel Kelly. Additionally, my family matches 35 additional people tied to the McCracken family that I have reasonable evidence to believe descend from the same couple, but where it is not yet proven. I also have multiple large groups of people who, based on DNA, appear to be related, but whom don’t have enough data to tie them to the McCracken family. One of these groups is huge with well over 100 people in it.
Peppard
In addition, I determined the other primary name associated with Lemuel is Peppard. Likely McCracken is one parent and Peppard is the other. I determined that member of our family that tested are related to over 100 identified descendants of Jonathan Peppard and Elizabeth Gilmore. It seems likely that these are Lemuel’s grandparents or great-grandparents. I have not yet decided which is his paternal family and which is his maternal family. However, I have analyzed all of their children and grandchildren to determine who could possibly be his parents. The list is very small.
I have also determined that these two families lived very close to each other in Wayne County, Ohio. Thus, this seems to be the likely place that his parents paths crossed. These pieces of information are now part of the assumptions.
It should also be noted that this information is the result of years of work. It did not happen overnight.
Complications
Complications typically arise when there connections between families. Generally, the further back the more connects that exist. For communities with lots of intermarrying (e.g. Quaker, Jewish, certain localities), it can be really complicated. I have worked some cases where it seemed like everyone was related to everyone else. The more inter-related families are, the more adjustments that may be needed in thinking about relationships indicated by DNA.
Lemuel Case Study:
I have encountered a limited number of connections between families. In some cases, two people who tested are related two different ways. In other cases, DNA matches are related to the family line I am researching plus another key family line. Fortunately, no significant intermarrying has been found on this project. Thus, the multiple connections have had a limited impact on my analysis.
Additional Reading
Note: For any of you that do a quick search on Ancestry or Family Search and see that people show John and Mary McCracken as Lemuel’s parents. I believe this to be incorrect information. I have explained to many people the flaw with this theory.