Monday, December 29, 2014

Wondering, Pondering, and Theorizing with DNA

With DNA and genealogy research, I talk to myself all the time. No joke! I’d bet that I am not the only one! Wondering, pondering, and theorizing out loud is actually the nature of documental and genetic genealogy research. To wonder is to think about it. To ponder is to consider thoroughly and deeply. Then, to theorize is to suggest scenarios or situations about what is possibly true or real. Talk it all out. Then, write it down! This helps one to map out a research plan in order to find the answers to pertinent questions that surface based on the information at hand or based on an interesting DNA match.

With that said, this blog post is threefold. First, I want to demonstrate this wondering, pondering, and theorizing using the case of a recent DNA match, as I work my way towards figuring out the family connection. Since I often talk with my fingers, I am talking it out in the writing of this blog post. Secondly, I want to stress the importance of utilizing GEDmatch. Thirdly, I want to put this research out there in hopes that it may lead to a major clue from someone who recognizes the families in question.

Ms. Cargill is a "Very High" DNA match to me in AncestryDNA. She has been in my list of matches since I’ve been on AncestryDNA, and I have wondered how she is related to me. The predicted relationship is 4th cousins. According to AncestryDNA, a confidence score of “Very High” means that Ms. Cargill and I share approximately 20 to 30 cM (centiMorgans) of DNA. AncestryDNA further assesses that there’s a 99% chance that my match and I share a single recent common ancestor within 5 or 6 generations. Thankfully, Ms. Cargill has a family tree that’s viewable to the public. Therefore, the wondering, pondering, and theorizing began. I ponder my DNA matches further by asking myself the right questions.

Q1: Do we have any common surnames in our families?

I looked at her family tree and recognized one common surname: MILLER. My father has Miller ancestors from Warren County, Mississippi. However, her Millers were from Arkansas, and her oldest-traced Miller ancestor, her great-grandfather Joseph Miller, was born in Tennessee around 1829. My father’s oldest-traced Miller ancestor is his great-great-grandfather, Fredrick Miller, who was born around 1827 in Mississippi, according to the censuses. However, the 1900 census-taker reported Virginia as the birthplace of Fredrick Miller’s parents. The differences in our ancestors’ locations do not mean that our Millers aren’t related. Many enslaved African Americans had family members that were taken or sold to other states. However, we lacked more information to connect our Millers.

Q2: Did she have ancestors who lived in the same area as my ancestors during and after slavery?

I viewed Ms. Cargill’s family tree thoroughly and determined that the answer to that question was NO. Since AncestryDNA does not provide any analysis tools, I couldn’t compare her to known relatives in order to narrow down how we may be related. At this point, I could only speculate that maybe the connection is on my father’s side and through our Millers.

Q3: What about GEDmatch?

More in-depth research is required to try to connect the dots. Also, since AncestryDNA lacks analysis tools, maybe we can turn to GEDmatch for possible answers. GEDmatch is a free, third-party DNA utility site that helps to find the family connection between people who have uploaded their autosomal DNA raw data file from 23andMe, AncestryDNA, or Family Tree DNA's Family Finder. Think of GEDmatch as “DNA Central” – the central location where DNA testers from 23andMe, AncestryDNA, or FTDNA's Family Finder can upload and compare DNA data. Therefore, in GEDmatch, one will gain many more DNA matches and have many analysis tools that allow people to compare, analyze, and determine how someone may be related and their ancestral backgrounds. GEDmatch is highly recommended (see

Thankfully, Ms. Cargill had uploaded her AncestryDNA raw data file to GEDmatch. I was able to do a “One-to-One” comparison and determined that she shares 20.6 cM with me. Since I have also uploaded both of my parents’ data files to GEDmatch, I was also able to determine that Ms. Cargill matches my mother and not my father. Therefore, our Millers aren’t related. She shares 20.7 cM with my mother with a MRCA (Most Recent Common Ancestor) of 4.7 generations back. Therefore, the connection goes back to slavery.

GEDmatch’s analysis tools as of 12/29/2014

Q4: What else can we do to help us figure out this connection?

Utilizing the DNA triangulation feature in GEDmatch, I decided to compare Ms. Cargill to other known relatives on my mother’s side in order to determine if the connection is on Mom’s mother’s side or Mom’s father’s side. Unfortunately, Ms. Cargill did not match them.

Ms. Cargill also asked herself the same question. Her parents are deceased. But as a substitute, she decided to test her father’s brother to garner more clues about her paternal side and to assist with DNA analysis. However, since her father and her uncle were half-brothers who shared the same father, any mutual DNA matches between her and her uncle would be related via her paternal grandfather, Roland Miller (1875-1940) of Augusta, Arkansas. Interestingly, testing a parent’s half-sibling can provide greater specificity since it narrows the family connection down to one grandparent.

Lo and behold, her uncle matched Mom!  According to GEDmatch, they share a whopping 81 cM over 2 segments with a MRCA of 3.7 generations back. Her uncle also shares 89 cM over 2 segments with Mom's sister. Therefore, GEDmatch is estimating that her uncle and my mother and aunt probably share the same great-great-grandparents, i.e., 3rd cousins. This is close kin, in my opinion!

Q5: Who were Roland Miller’s parents, grandparents, etc. and where were they from?

Ms. Cargill has only been able to trace back to her grandfather Roland's parents. They were Joseph Miller (born c. 1829) and Sarah Roddy (born c. 1843). Both had been enslaved near Augusta, Arkansas and remained in the area after slavery. The family connection is via Joseph or Sarah.

Q6: What can I garner from the censuses?

According to the 1870 and 1880 Woodruff County, Arkansas censuses, Joseph Miller was born in Tennessee and Sarah Roddy Miller was born in Arkansas. According to the 1880 census, Joseph and Sarah’s parents were born in Tennessee. Therefore, for both Joseph and Sarah, there was a forced family migration from Tennessee to Arkansas. Of noteworthiness, the same migration pattern existed for other black Miller and Roddy families that lived nearby. Presently, Ms. Cargill does not know the names of her great-great-grandparents.

Q7: Where is the Tennessee connection in my mother’s family?

My only known, Tennessee-born maternal ancestor was my mother’s great-great-grandmother, Peggy Warren Milam. After years of research, I was finally able to determine that she was born in Williamson County, Tennessee around 1829. Grandma Peggy, her parents, and siblings were enslaved by Edward Warren, who transported them to Marshall County, Mississippi during the 1830s. Grandma Peggy was eventually sold to Joseph R. Milam of present-day Tate County, Mississippi. I recently determined that Grandma Peggy had a sister who was sold to Ed Warren’s son-in-law, Erasmus J. Ellis; he took her to Ouachita County, Arkansas around 1858. For further details about that discovery, which was revealed from a close DNA match, read DNA Does It Again – Another Long Lost Sibling Found!.  

Q8: Who were Joseph and Sarah Miller’s last enslavers, and are there any obvious links to Edward Warren or Erasmus J. Ellis?

After researching the 1850 and 1860 Jackson County, Arkansas Slave Schedules, the 1830 – 1870 censuses, and searching through, as well as finding an informative online source, I was able to tie both Joseph & Sarah Miller to the Alexander & Agnes Roddy Family of Jackson County, Arkansas. (Note: Woodruff County was created from Jackson County in 1861.) Six key facts about the white Roddy family included the following:

     (1)   Alexander and Agnes Roddy were from Spartanburg County, South Carolina. They had 7 sons and 1 daughter named Rose.
     (2)   Their daughter Rose Roddy married John MILLER in 1807 in Spartanburg County, and the Millers subsequently had seven children.
     (3)   The Roddy and Miller families both moved to Tipton County, Tennessee by 1825. The Roddys and Millers were reported in the 1830 Tipton County, TN Census.
     (4)   By 1835, the Roddy and Miller families both moved to Jackson County, Arkansas and established a plantation called Walnut Woods near Augusta, Arkansas.
     (5)   Alexander Roddy died in Jackson County, Arkansas in 1840 at the age of 81. According to the 1840 census, two sons, Elias and John Roddy, owned 25 slaves, collectively.
     (6)  His daughter, Rose Roddy Miller, died in 1851. Per the 1850 Jackson County, Arkansas Slave Schedule, three of Rose’s sons, Wilson, James, and Henry Miller, were the only Miller slave-owners in the county, owning 24 slaves, collectively.

Although I haven’t been able to find any online Arkansas probate/estate records or wills for members of the Roddy and Miller families, chances are very high that people in this family were the last enslavers of Joseph & Sarah Miller and their parents. First, the migration pattern matches (TN>ARK). Secondly, the white Roddys and Millers lived in the same township as Joseph and Sarah, and other black Miller and Roddy families in 1870. Nonetheless, I could not find any ties between the white Millers and Roddys to Edward Warren and Erasmus J. Ellis.

Q9: Could the connection go back to South Carolina?

After learning that the white Roddy and Miller families had been living in Spartanburg County, South Carolina, I diverted my attention to my mother’s great-grandparents, Edward Danner (1832-1876) and Louisa “Lue” Bobo Danner (1842-1921). Both of them were born in Union County, South Carolina, which is adjacent to Spartanburg County. Their enslaver Dr. William J. Bobo took them to Panola County, Mississippi in 1858. Thomas Danner Jr., the previous enslaver of Grandpa Edward and his parents, and David Boyce, the previous enslaver of Grandma Lue’s mother and her family, as well as Dr. William Bobo, had all resided in the Cross Keys and Sedalia area of Union County. This area is less than 10 miles from the Union/Spartanburg County line.

Q10: Do Ms. Cargill and her uncle match my Mom’s 2nd cousin, too?

Fortunately, my mother has a second cousin, Orien, who has taken the AncestryDNA and the 23andMe DNA tests. Edward and Lue Danner are Cousin Orien’s great-grandparents, too.  Therefore, I asked Ms. Cargill to see if she matches her in AncestryDNA. She confirmed that both she and her uncle also match Cousin Orien! Major clue!

Q11: Is DNA triangulation possible?

Since Cousin Orien hasn’t uploaded her data to GEDmatch yet, I decided to see if all of the matches were on the same chromosome. Using the chromosome browser tool in GEDmatch, I was able to do some DNA triangulation between Mom, Ms. Cargill, and her uncle. As the following diagram shows, Mom matches Ms. Cargill and her uncle on overlapping segments on chromosome 17, from 7.2 to 39.2 Mbp (50.4 cM) with her uncle, and from 18.5 to 39.2 Mbp (20.7 cM) with Ms. Cargill. Mom also matches her uncle on chromosome 8 (30.5 cM).

Via 23andMe, Mom shares 324 cM across 17 segments with Cousin Orien. Utilizing the Family Inheritance Tool in 23andMe, I determined that Cousin Orien matches me and Mom on chromosome 17 from points 6.0 to 38.0 Mbp.  See diagram below. Since Ms. Cargill confirmed that she and her uncle match Cousin Orien in AncestryDNA, I am confident that they all will be matching in the same area on chromosome 17, indicating that they all descend from a common ancestor(s). 

UPDATE (12/31/2014): Cousin Orien has uploaded to GEDmatch. As anticipated, she indeed matches Ms. Cargill's uncle on chromosome 17 in the same region at 48.1 cM. She also matches him on chromosome 10 at 15.3 cM for a total of 63.5 cM.

Q12: What did this additional DNA triangulation conclude?

It confirms that there’s a close family relationship between Joseph Miller (born c. 1829 in TN) or Sarah Roddy (born c. 1843 in Ark.) and Grandpa Edward Danner (born c. 1832 in SC) or Grandma Louisa “Lue” Bobo Danner (born 1842 in SC). Perhaps two of them – Joseph and Grandpa Ed, Joseph and Grandma Lue, Sarah and Grandpa Ed, or Sarah and Grandma Lue – were first cousins who shared the same grandparents? If that’s the case, Mom and Ms. Cargill’s uncle would be 3rd cousins once removed who happen to share above-average DNA even for 3rd cousins. See DNA sharing chart here.

Q13: Can a closer look at the time frames and locations give more insight?

The answer to this question is YES, and hopefully this blog post demonstrated that. It helps to analyze where people were and when they resided in various locations. Mapping out time frames and locations can greatly assist in figuring out family connections.

Grandma Lue’s family is the foundation of my first book, Mississippi to Africa: A Journey of Discovery, which goes back to her mother Clarissa Bobo (born c. 1823), Clarissa’s parents, and three of Clarissa’s four grandparents. Grandma Lue’s father was white. Therefore, I am placing my bet that the connection is via Grandpa Edward Danner since I have never found any ties to the Miller/Roddy Family of Spartanburg County, SC in my Bobo/Boyce research. But you just never know sometimes!  Many unknowns remain prevalent, even within well-researched family lines. More genealogical research is necessary to solve this mystery. More DNA matches will greatly assist. However, asking the right questions out loud and researching to find the answers, in conjunction with DNA triangulations, enabled me to narrow it down to this point, which is good progress.

Stay tuned for updates to this DNA mystery. Hopefully!


  1. I have family ties of Miller from around Earl Arkansas and Jackson County Mississippi. Luckett sisters married brothers. Very few if any would know any details. I have some people I can talk to maybe.

  2. Melvin, once again an excellent post. I really appreciated how you took us through your step by step process, which was very instructive.

  3. Very important to write all of this for the world, Melvin. You are not just talking and theorizing with yourself. You are showing countless others your data and your inferences. Different people will notice different points. Some will suggest other paths of research, make you aware of a website or repository you did not know about, or point out an omission or an unsupported conclusion.
    Rare but possible: someone may be sitting on the documentation and proof you need and decide to come forward.
    Engaging the analysis of others is one of the greatest benefits of blogging, I think.

  4. I was just talking with a dna cousin yesterday & we were talking about like matches. Your post is right on time -- very helpful. Thanks!

  5. Wow, Cousin. Thanks for laying it out for me. I didn't make it to the Archives as I had intended, but when I do, you have given me a good place to start looking. Thanks again, Melvin. What you are doing is such a blessing to so many!!!

  6. Once again Melvin, excellent post & awesome research! You have given so many a restored faith in searching for our Ancestors as well as bridging the gaps between our present & our past! Thank you so much for sharing your gift with us all!

  7. Melvin an instructive, informative post. When I first read your book, I wondered if we were related. Time passed and I just received an email from your DNA cousin J.McNeil. She said we were related through DNA. (she and I)
    Documented research the family surnames are Floyd, Tribble, Glenn from Newberry, Union, Saluda ,Counties SC. ( @1822-1917) My maternal ggrandparents Rev George and Fannie Glenn migrated from Greenwood County, SC to Philadelphia (1917)
    Undocumented research theory through the internal slave trade descendant(s) were transported from Mecklenburg or Lundenburg County, VA @1790 to Silverstreet, Newberry, SC
    Although not research based I will add the ancestors are talking. I will message you FB the kit numbers.

  8. Melvin, Great story. I wonder if we are related as I am researching my Miller lineage in Spartanburg,South Carolina.


Thanks for visiting and commenting on Roots Revealed!