RLP with DNA Study Group Part 5: Exploring DNA Tools

What DNA tools can we explore to help us make progress in our genetic genealogy? Part of the Research Like a Pro with DNA process is exploring tools both for pedigree triangulation and tools that work with segments. As we work on specific projects, we don’t need to know how to use every company and 3rd party tool, but we do need knowledge of what tools are available and how they can help us in our research.

In the hopes that DNA can help to discover an unknown 4th great-grandfather, I’m returning to my brick wall of identifying Cynthia (Dillard) Royston’s father. The Research Like a Pro with DNA study group is a perfect way to tackle a persistent research question. The structure of the assignments helps me stay on track and the process keeps me moving forward. I’ll be sharing my progress in this series. Will I finally be able to answer the question of who was Cynthia’s father? I don’t know. But I do know that I’ll be one step closer by the end of the study group.

When working with DNA, using the DNA matches of the closest generation to the target ancestor can make all the difference. For this project, I’ll be analyzing the DNA matches of my second cousin, twice removed (2C2R), Victor Parker. While Cynthia is my third great-grandmother, she is Victor’s great-grandmother. He received approximately 12.5 % of her DNA, whereas I only received about 3% or less.

I’m working on this project in phases. Previous phases included documentary research to eliminate Dillard candidates for Cynthia’s father and analysis of the DNA to find a cluster of DNA matches to research. Here is my current objective.

The objective of this research phase is to test the hypothesized biological sibling connection between Elijah Dillard and Cynthia (Dillard) Royston. Elijah Dillard was born about 1814 in Georgia and died on 6 September 1886 in Coffee County, Alabama. Cynthia was born about 1816 in Georgia and died in 1882 in Collin County, Texas. Cynthia married Thomas Beverly Royston about 1833 in Georgia or Alabama.

Up to this point, I’ve completed the following steps and written about them in these blog posts.

RLP with DNA Study Group: Part 1 Assess Your DNA Matches & Analyze Your Pedigree

RLP with DNA Study Group Part 2: Organize Your DNA Matches and Create an Objective

RLP with DNA Study Group Part 3: Timeline, Source Analysis, and Citations

RLP with DNA Study Group Part 4: Locality Research and Ethnicity

Exploring DNA Tools

We have two lessons in the study group about DNA tools. Because there are so many – we divide the lessons into learning about tools that help with creating genetic networks and pedigree triangulation and the tools that work with DNA segments. Robin Wirthlin has created a useful chart showing a bell curve of DNA tools, starting with the easier tools on the left and moving to more difficult or time-consuming tools on the right. We will spend most of our time using tools in the Bell Curve’s highest point.

Exploring Genetic Networks & Pedigree Triangulation – GWorks

Up to this point in the project, I had already analyzed ethnicity estimates and created genetic networks using the Leeds Method, DNA2Tree AutoClusters, and a Gephi network graph. I had diagrammed relevant DNA matches for my tester, Victor Parker, and now it was time to explore a new tool.

Because I’m working with the cluster discovered in the network graph with the common ancestor of Elijah Dillard, I wanted to use a tool to help me explore Ancestry trees. Most of the DNA matches in the cluster have no family trees or small trees. As I build the descendancy tree of Elijah Dillard, I want to seek connections among the DNA matches. DNAGedcom’s GWorks sounded like an excellent choice since it takes the tree data and allows you to work with it in the online database to find surnames, given names, birth dates, death dates, and places. GWorks has a direct link to the match page for further analysis.

Here is the tool description from DNAGedcom.

GWorks is a powerful suite of tools custom-designed to help genealogists find buried treasure in family trees. GWorks can sort, filter, compare and perform Boolean searches on the trees you
upload and match. You can then tunnel down into the details and find exact matches in your trees, generate surname lists and create Boolean searches based on any criteria you select.

GWorks can analyze GEDCOMs you upload and trees posted by your autosomal DNA test matches on Ancestry DNA or Family Tree DNA.

In the screenshot below, you can see the database containing the AncestryDNAfamily tree information for Victor Parker’s DNA matches. Shown are the results from my surname search of “Dillard.”

GWorks groups together the various individuals with the same birth and death information. The first category is the file name (grayed out) which is a live link that takes you to the DNA match page where you can open up the entire tree.

If a DNA match has alternate information for the name, birth, or death, the individual will be grouped separately. This lets you see immediately variations for the same people. For example, some DNA matches have a birth date of 1814 for Cynthia and they were grouped separately.

To use GWorks, I carefully followed the tutorial from DNAGedcom. It worked perfectly! I’m looking forward to searching for specific surnames in DNA match trees when working with the descendants of Elijah and Josiah Dillard to find more matches.

Exploring Chromosome Browsers & Segment Data – Genetic Affairs Hybrid AutoSegment Cluster

After exploring genetic networks and pedigree triangulation, our next step in the RLP with DNA process is to explore chromosome browsers and segment data tools. I decided to run the Genetic Affair’s hybrid AutoSegment cluster analysis that combines segment data for MyHeritage, FamilyTree DNA, 23andme, and GEDmatch. I had my tester, Victor Parker’s data from all those companies except 23andme. Note, because Ancestry DNA does not provide chromosome segment data, it can’t be used in the comparison.

Setting Parameters

An important part of the analysis is understanding what parameters to set. You select the parameters for the overlapping segment size and how many matches you want to be included in each cluster. You can choose as small as 8cM or go as high as 50 cM for the overlapping segment size.

I ran two tests – one for segments that overlapped at least 12 cM and one for segments that overlapped at least 20 cM. I knew the size of the shared segments would be relatively small, 20 cM or less, so didn’t run a test for larger segments. Although Cynthia (Dillard) Royston is Victor’s great-grandmother, she would be the 3rd or 4th great-grandmother for the majority of his matches and they would not have inherited very large segments of DNA from her.

The analysis for matching segments of 20 cM or more resulted in 69 clusters. In contrast, the analysis for matching segments of 12 cM or more resulted in 150 clusters. If I were to increase the overlapping segment size to 50 cM I’d have fewer clusters and if I were to decrease the overlapping segment size to the lowest threshold of 8 cM, I’d likely have more than 200 clusters. My main takeaway was to experiment with the matching overlapping segment size.

You also choose settings for each company for the maximum and minimum shared DNA. You can just leave the default settings or adjust them depending on your project. I chose to leave the default settings.

Each project will warrant a different approach to using this tool. If you’re working with an adoptee case, you’d want to go with 50 cM because you’re looking for close cousins. However, if you’re looking for 3rd to 4th cousins, you’ll need to use smaller values for the segment overlapping.

The screenshot below shows the defaults. By clicking the arrow on the right of each white box, you can change the default. Genetic Affairs explains how to obtain the DNA match and segment files in this tutorial. I was pleasantly surprised with how easy it was to retrieve the data, plug it in, then run the analysis by clicking the green box.

My results were emailed to me within a few minutes and when I opened my Zipped file this is what I received.

The folder titled “chromosomes” contained the details for each cluster. The next file took me to the report powered online by Genetic Affairs. As I waited for it to load, I saw the clusters appear before my eyes. The following screenshot shows the first 26 clusters.

Because the clusters were so small, I found the Excel spreadsheet was the easiest to use in analyzing each cluster. The Excel spreadsheet listed all the clusters and the DNA match information. For example, the brown Cluster 15 shown very small above is clearly laid out on the spreadsheet, in between the purple Cluster 14 and the pink Cluster 16. Each line represents a DNA match.

To the left of the clusters is the information for each line. The screenshot below corresponds to the above clusters with the DNA match names grayed out. The spreadsheet lists the DNA testing company, match name, amount of total shared cM, the cluster number, any notes that you’ve recorded in the testing company for that match, and a direct link to the matches online tree. An important consideration for the analysis is whether the segments are matching on the maternal or paternal chromosome.

Clicking on the cluster number takes you to the Genetic Affairs page for that AutoSegment Cluster. For cluster 16, I saw that the matches were on chromosome 5. I was also able to view the start and stop points for each segment. Two of the matches are on MyHeritage and I can check there to see if the segments triangulate. Why wouldn’t they triangulate since the image looks like they should? Remember each chromosome has two parts -maternal and paternal. One segment could be on the maternal chromosome and the other on the paternal chromosome.

How do you decide if the matches in the cluster are really on the same portion of the chromosome – such as maternal or paternal? You can view the matches on the website and see if they are on the known maternal or paternal line. You can also check for triangulation using the company tools. In the example above, I found the first two matches on My Heritage and saw that they did triangulate on chromosome 5.

Always do a follow-up on the analysis to see if the segments truly triangulate.

How am I using the hybrid AutoSegment tool from Genetic Affairs? I’m hoping to eventually discover a Dillard cluster and a segment that could point to my unknown Dillard father of Cynthia.

To learn more, watch this video featuring Everett Jan Bloom, the creator of Genetic Affairs discussing Auto Segment. There is also a Facebook group for Genetic Affairs where you can ask questions and get help with the hybrid AutoSegment tool as well as the others. See the complete manual for Genetic Affairs to get started.

Best of luck in all your genealogy research!