Exhaustive MRCA Analysis for GEDmatch Matches

May 15, 2025May 26, 2025 ~ Patricia Coleman ~ Leave a comment

A new import tool has recently been added to AutoLineage, enabling the use of GEDmatch Tier 1 data. This powerful addition allows users to identify potential common ancestors based on tree data linked to up to 7,500 matches. Unlike existing tools (even the ones on GEDmatch) that typically only analyze trees for a small number of top DNA matches, this new capability dramatically expands the scope of common ancestor analysis.

Start by selecting the One-To-Many – Full Version.

Enter your kit number and select 7500 matches for the comparison. Click on ‘Search’ and a message will come up for you to wait while the data is collected. When the site has loaded completely, save this page using the ‘Save Page As…’ command. Make sure the complete site is loaded, it might take a while to fully process.

Enter kit number and select 7500 matches.

Next click on the ‘Select all with Gedcoms’ command. This will select all DNA matches that have a linked GEDCOM.

At the top of the list of matches and gedcoms is the ‘Visualization Options.’ Click on that next.

List of visualization tools for the matches and gedcoms.

Select ‘GEDCOM’ that I’ve circled in red in the figure above, which will bring up the tools.

Select the ‘Find matching GEDCOMs’ which brings up a new tab with a detailed list of the gedcoms. This page will allow for linking the DNA matches (represented by the kit numbers) to their GEDCOM trees as well as which root person they represent.

‘Save Page As….’ using the filename that GEDmatch gives it. Make sure that the page is fully saved before the next step. Then go back to the previous page and select ‘Find Matching GEDCOMs (anc)’. This step takes some time to come up. It will quickly display a heading, but you don’t want to save the page until after that X, that I’ve circled in the figure changes to a circle with an arrow and the table has been filled in.

The anc2 page before it has collected the data.

The anc2 page when all the data has been collected.

Once the page is complete, save it using ‘Save Page As….’ again using the filename that GEDmatch gives it.

Now you are ready to add the matches and their trees to the AutoLineage profile. If you’ve not made a profile yet, log into Genetic Affairs, select AutoLineage (local) and click on register a new profile.

Creating a new profile on AutoLineage.

After the profile has been created, select the register DNA test.

Next select the GEDmatch as the DNA test.

Next select Import matches and the Wizard will appear. The ‘One to many HTML file’ is what you want to select.

Wizard to import the matches from GEDmatch.

Clicking on the blue question mark explains what you want to collect, which you’ve already done.

Instructions for collecting the 7500 GEDmatch match files.

The list of matches is then displayed.

The next step is to add the gedcoms that you collected. Go to ‘Tree Management’ and select ‘Import trees.’

Clicking on the blue question mark gives the detailed instructions for using this GEDmatch import tool.

The trees are now automatically linked to the matches and with the matches and linked gedcoms in place, you can run ‘Find Common Ancestors.’ Select Judy’s profile and go to the overview page. Select the find common ancestor button.

I know that Judy’s closest matches are her siblings and nieces. Then her next highest match is less than 100 cM. So I ran the ‘Find Common Ancestors’ wizard down to a birth year of 1700.

FInd Common Ancestors Wizard set for 1700 birth year.

After waiting for quite some time, a large number of trees where found. The first one, shown here, connects Judy and her siblings to their fifth cousins once removed. Judy’s mother was a Hocking, whose family was originally from Cornwall, England. Some of the Hocking family immigrated to Australia, and Judy’s side immigrated to the United States. The Hocking line is well documented by the Hocking Descendants Society based on Australia.

Judy’s family tree on the Hocking side..

In many cases when I find unknown matches and want to find the connection to the tester, I start building a private, unsearchable tree on Ancestry. After a short while Ancestry provides hints to records and other trees, and I also research other sources for records and information to make the connection. Luckily, a new GEDCOM export capability in AutoLineage can facilitate the tree building!

On the ‘Tree Management’ ‘Tree Overview’ Page scroll to the bottom to ‘Manage Tree Data’ and select ‘Export Trees to Gedcom.’ This option will generate a new GEDCOM file using all of the trees that have been found, and optionally you can add additional DNA matches that do not have trees. All trees and DNA matches are provided in the GEDCOM as floating trees or floating individuals. In addition, tree persons linked to DNA matches as well as the individual DNA matches are provided with the DNA match tag, allowing for easy retrieval when searching for people.

Profiles for Judy and another tester, Linda, are available. Judy’s profile has 898 trees that came from GEDmatch, which is the only test site I’ve added to Judy’s profile currently. The wizard shows trees linked to available profiles as well as showing all trees (the all-trees option). Moreover, it’s possible to deselect individual trees if they do not need further researching in Ancestry.

I’ve checked the ‘Integrate DNA matches’ which allows me to add up to 1000 additional DNA matches that are not associated to an existing tree. These would show up in the resulting gedcom as floating trees of one person.

I selected 100 additional DNA matches and clicked on the export. It takes some time for the wizard to finish and it puts the gedcom, named ‘example.ged’ into your download folder.

Next in your Ancestry ‘Trees’ menu, scroll down to ‘My Trees.’

Select ‘Upload a gedcom file’ and add the ‘example.ged’ file. You can then name the tree whatever you like. I unclicked the ‘make the tree public’ and after it was processed I turned off the ‘make tree searchable’ under the Privacy settings.

It took several hours for Judy’s large gedcom to finish processing.

List of all the people there were in the example.ged file.

Since I want to find details of Judy’s third great grandfather, David Coleman, who was born somewhere in Ireland around 1801, I searched Coleman.

It appears there are two listings for David Coleman born around 1801 in Ireland. David, born in 1835 is his son. Carefully review the trees before merging them, or all the people in the tree will be duplicated. David has 16 hints and potential parents that someone has found. In the list of hints there are several US Federal Census, homestead records in Michigan and a death record for his daughter.

Summary

Using GEDmatch Tier 1 One-To-Many – Full Version, selecting 7500 matches and then finding the gedcoms that these matches have provides an easy way to find common ancestors using AutoLineage. This approach also works for profiles that do not have a linked tree, in that scenario, the common ancestors will be identified using only the trees of DNA matches.

The new export to GEDCOM functionality provides an easy method to boost your tree finding efforts, without the need to manually copy existing tree data.

Convert Ancestry ProTools to Trees using AutoLineage and AutoKinship!

November 3, 2024November 4, 2024 ~ Patricia Coleman ~ Leave a comment

From ProTools to AutoKinship tree using matches shared by my cousin, Tish¹, and me.

Genetic Affairs has unveiled a powerful new feature in AutoLineage—now you can apply AutoKinship to each individual cluster generated from Ancestry ProTools shared matches, unlocking a whole new level of insight into your family connections! It invokes the functionality of AutoKinship on the site directly at no additional cost, to provide reconstructed trees based on the shared DNA between shared matches.

In a previous blog post, I described how to get started with Ancestry and AutoLineage. In short, these are the steps involved. First, a profile is created or selected after which the same is done for a DNA test. Select a generic DNA test in order to process Ancestry matches. Next, we import DNA matches by saving each match page (make sure to set the number of matches to 50) to an HTML file. Repeat until you have enough matches. Next, we download shared matches for the matches of interest followed by a clustering analysis. Once we obtain the clusters, we can select an individual cluster for which we want to have a reconstructed AutoKinship tree.

In this blog post I will demonstrate how I cluster the shared matches of my 2nd cousin Trish. For this purpose, I will visit each DNA match page in the shared matches list with Trish and download their shared matches.

First, I imported my direct DNA matches. Next, I visited Trish her match page and downloaded the shared matches by using the Chrome “Save page as” to save the shared matches. Trish and I have three pages of shared matches, and I saved them as Trish-p1, Trish-p2 and Trish-p3. I used the “Import shared matches” to add these matches that Trish and I share. Next, I visit each DNA match in the shared matches with Trish and download the shared matches for each match.

In order to get a cluster that only showed Trish and my shared matches, I selected “DNA matches” from my DNA test linked to Ancestry page and then clicked on Trish’s name.

Clicking on “perform clustering analysis” brings up the Cluster wizard. Again, this clustering approach will only take all DNA matches that are shared with Trish, basically allowing me to only cluster matches that share an ancestral connection with myself and my 2nd Trish.

Trish has 42 matches that she shares with me.

Clusters of Trish’s and my shared matches.

Notice the cM values listed in each of the colored cells in the cluster. These are the result of the new Ancestry ProTools data!

Trish and I share great grandparents, Thomas Byrnes and Bridget Fenton.

Thomas was born in County Roscommon, Ireland and Bridget in County Limerick. They met and married in Virginia. From building out family trees we know that the orange cluster has ancestors from Counties Roscommon and Mayo, and the green cluster has matches with surname Burns. Those matches would be on Thomas’ side. The brown cluster has ancestors who lived in County Limerick and would be on Bridget’s side. Very few of the matches in the large red cluster have any trees or other family information.

Clicking on “Matches” at the top of the screen brings up a table with the list of matches in each cluster.

The next step is to click on cluster 3, which brings up just the red cluster. Again clicking on “Matches” at the top of the screen brings up the list of matches for cluster 3 only.. In the upper left of the screen is the “Run AutoKinship” button.

Cluster 3 matches and AuroKinship button.

Once AutoKinship analysis starts a message appears at the bottom of the screen telling you that the results will be sent to your email. The zip file should be saved to the computer and then unzipped. There are ten probability trees in the AutoKinship.

AutoKinship tree gives hints as to how matches can be related based on the amounts of shared DNA. At the top of the tree beau is listed as child of M, and M and P are listed as siblings. This is identical to what Ancestry said about these matches. But not every relationship is correct. Trish and I are second cousins, but we share a larger amount of DNA than would be expected compared to the average for second cousins, and we’re often miss assigned as first cousins or first cousins once removed. However, the AutoKinship tree can be used to get valuable hints for connections for the matches.

At the bottom of the page with the first AutoKinship tree is a matrix of the DNA matches used for the tree. This matrix can be downloading into Excel and saved.

Part of the DNA matrix for the red cluster.

In conclusion being able to organize you Ancestry ProTool matches into a probability tree and generate a matrix provides addition methods for viewing your shared matches.

Note that the current AutoKinship approach will be replaced with a more versatile version that will run directly in AutoLineage. This version will allow for using pre-determined relationships or integration of known trees into the reconstructed trees.

Patricia Harris Anthony, Trish, has given me permission to use her real name and her data.

Linking GEDmatch and FTDNA – AutoLineage for iGG

May 7, 2024 ~ Patricia Coleman ~ 6 Comments

AutoLineage is a powerful tool that allows you to cluster your matches at a particular testing site but also to find common ancestors across multiple sites. In this blog, we will discuss a scenario when only data from FamilyTreeDNA (FTDNA) and GEDmatch can be used as sources, for instance for an investigative genetic genealogy search.

An example reconstructed tree using only FTDNA and GEDmatch data.

We will be importing and analyzing data from FTDNA and GEDmatch, and use the common ancestor identification tool to find common ancestors. Also, we’ll show how to use the hints tool to get insights about ancestral lines that could hold a potential common ancestor.

From the landing page select ‘Register a new Profile’ in the left panel. Start AutoLineage by making a profile for the case you are investigating and for whom DNA matches are obtained. In this blog, we shall be using Dave as an example.

Starting a new profile.

Family Tree DNA

Next, a DNA test is registered.

The first thing to do is to import Dave’s matches from FTDNA. Click on the ‘import matches’ button.

Flowchart showing Import matches.

Each testing site has its own list of methods for collecting matches.

Options for uploading matches from FTDNA.

Here we are using the first one, a CSV file from Genetic Affairs. Clicking on the question mark explains what file this option needs.

Explanation of the CSV from Genetic Affairs.

For this option, we first ran the AutoCluster analysis with the AutoTree feature enabled for FTDNA on Genetic Affairs.

Starting AutoCluster with AutoTree enabled on Genetic Affairs.

After starting the analysis, an email appeared with a download link. Alternatively, this link is also available in the notification section on the members page of Genetic Affairs (top right corner, under the bell icon).

After unzipping the results file, we found all the data we needed for AutoLineage. In the Gephi folder, two files are present; nodes.csv and edges.csv. The nodes file contains the matches that would be selected in the dialog box of the ‘Wizard – import matches’.

The imported DNA matches are shown after the import process has been completed.

Clicking on ‘DNA test overview’ at the top brought back the flowchart where we selected ‘Import shared matches.’. Shared matches are required for the clustering process.

The wizard explains that the edges.csv file from the Gephi folder is used for the shared matches.

After importing the shared matches, which can take some time, the number of shared matches for each match is now displayed in the table.

Matches list showing the number of shared matches each has.

With matches and shared matches available, we ran the cluster analysis for FTDNA. Going back to the ‘DNA test overview’ brings up the flowchart where we selected ‘Cluster matches.’

The ‘Clustering wizard’ allows us to select the range of matches for the clustering, which type of clustering, and which color scheme should be used.

We selected to run all of Dave’s matches in the clustering analysis. After selecting the “start clustering” button, the clustering chart appears.

GEDmatch

For the GEDmatch data, we added data from an AutoKinship analysis (Tier 1).

Going back to Dave’s profile page we registered another DNA test and selected GEDmatch from the drop-down list. Next we ‘import matches.’ An unzipped AutoKinship report also has a Gephi folder containing the nodes.csv and edges.csv files. Nodes.csv contains Dave’s matches, and edges.csv contains the shared matches. Now these matches can be clustered. We ran 500 kits for AutoKinship and 362 of these have shared matches.

Adding Trees

Now it’s time to add trees. The trees are typically associated with the matches, but they are managed independently of the DNA testing site. Clicking on ‘Home’ in the left panel brings up a list that includes ‘Tree Management.’ Clicking on ‘Tree Management’ brings up the flowchart for Trees.

In ‘Gather Trees’ we clicked on ‘Import Trees’ which brought up the ‘Tree Wizard.’ There is a setting for the trees associated with the GEDmatch DNA matches, and a different setting for trees associated with the FTDNA DNA matches. In both cases, they retrieve tree information from the ‘match’ files associated with the GEDmatch AutoKinship or the FTDNA AutoTree AutoCluster.

‘Import tree wizard’ showing the two options for importing the trees associated with both testing sites.

In both scenarios, we selected all the HTML files in the match folder of the unzipped AutoCluster FTDNA report or unzipped AutoKinship GEDmatch report. Using these methods also ensured that the trees were automatically associated with the FTDNA and GEDmatch DNA matches

The GEDmatch AutoKinship folder contents.

The list of trees inside the matches folder.

Typically, iGG profiles are not associated with trees, but in our example, Dave does have a tree, and it needs to be linked to his profile. For other cases where the tester does not have a tree, such as people looking for birth parents, this step would be skipped, and you’d go on to ‘Find Common Ancestor’ from the various DNA matches.

To link Dave’s profile to a tree, we visited the tree management and select Dave in the ‘link profiles to tree’ section (alternatively, go to the tree view in Dave’s profile and associate the tree from there).

We clicked on Dave’s name and brought up the Wizard to link Dave’s profile to his tree. Next we searched for Dave’s tree in the name field.

Selecting Dave brought up his tree where we selected him as the root person of the tree.

Selecting Dave as the root person in his tree.

This brought up Dave’s tree. Notice the person image over Dave’s name that tells the tree is connected to his profile.

Dave’s tree. Names of living persons or those who died since 1973 have been hidden,

Going back to Dave’s profile page we scrolled down to ‘Find Common Ancestors’ near the bottom of the page. There are now two DNA tests associated to his profile, and some of the DNA matches are linked to trees that we imported. We selected the ‘find common ancestor’ button to start the common ancestor wizard.

Running the analysis with the default birth year of 1800 identified eighty-seven common ancestors for Dave across FTDNA and GEDmatch. A problem arises when a match is on both sites and has a tree at both locations. The match on the right of the reconstructed tree shows both FTDNA and GEDmatch. She and her tree are duplicated which causes brown squares, indicating that tree persons are duplicated. Seeing a lot of these brown squares is often indicative of a problem.

DNA match who is on both sites and has trees on both.

Clicking on ‘Linked trees’ shows the trees that are attached.

Right-clicking on the tree brought it up in a new tab. Since the GEDMatch tree has more people we kept that one and deleted the FTDNA one. Next, we attached Rachel’s FTDNA DNA match to the remaining GEDmatch tree.

Connecting Rachel’s FTDNA match to her GEDmatch tree.

Rachel shared 30 cM with Dave at FTDNA and 27.7 cM at GEDmatch. Looking through the list of matches in the Wizard there’s only one Rachel that matches Dave at FTDNA with 30 cM, so we selected that one.

Clicking on the DNA symbol over Rachel’s name shows she is connected to both GEDmatch and FTDNA matches to Dave.

Finding Hints

Since Rachel was the only match on both FTDNA and GEDmatch we used her tree and ‘calculate hints.’

Calculate hints is basically a common ancestor identification with a focus on the tree that is shown on the screen. The tool compares the names and dates in Rachel’s tree with all the other trees that have been loaded into AutoLineage. It can take a minute or so for this calculation depending on how many trees there are. Contrary to the regular common ancestor tool, the hints tool also shows hints about shared surnames and/or surnames and first names.

The calculate hints from Rachel’s tree showed a yellow hint for Phyllis May Thompson and for Grace Tonkin. Yellowish hints indicates a shared surname and might mean that there might be something useful farther back in time. So we expanded the tree out first from Phyllis. The common surname, Barrett, was found in another DNA match’s tree, but the dates were several hundred years different. Next we looked at the hint for Grace.

Grace Tonkin’s hint showed that Dave had several people in his tree with the Tonkin surname, but her date is more recent than the people in Dave’s tree. We followed the Tonkin line past Grace in Rachel’s tree and found that Thomas Tonkin’s dates were in the same range as those in Dave’s tree.

Thomas Tonkin’s wife, Margaret Hattam, also has a hint.

Looking at Margaret Hattam’s hint we find that Dave’s tree has several Hattam ancestors.

Margaret Hattam was born in 1785 and Dave has an ancestor, Mary Hattam born in 1783. When we added Margaret’s parents to our tree, and rerun the hints tool we found new green hints linked to Margaret’s parents.

John Hattam and Elizabeth Eddy both have green hints.

Looking at John Hattam’s hint we found that he is in both Rachel’s tree and Dave’s tree. We’ve found the common ancestor!

John Hattam in both Rachel’s and Dave’s trees.

Since John Hattam had a birth date of 1754, we changed the birth year to 1700 in the parameter of the ‘Find Common Ancestors’ wizard and then reran ‘Find Common Ancestors.’

Changing the birth year in the Find common ancestor Wizard.

Now a total of fifty-four ancestors were found, and the reconstructed tree showed the connection between Dave and Rachel. They have common fourth great-grandparents, John Hattam and Elizabeth Eddy.

Reconstructed tree connecting Dave’s and Rachel’s trees.

Dave’s tree is showing brown squares which often means there is a duplicate tree or a date problem, however for Dave’s family this is correct. He has two Hocking lines. John Hocking’s parents were James Monteith Hocking and Martha Murrish. James Monteith is one of the Hocking lines, and Martha’s mother was Eliza Hocking, which is the other Hocking line. They both go back to John Hattam and Elizabeth Eddy, so they are Dave’s fourth great-grandparents twice!

Summary

We ran AutoLineage using only FTDNA and GEDmatch data to look for common ancestors to Dave and his matches. AutoLineage generated clusters for the FTDNA matches and also for the GEDmatch matches. Trees from both testing sites were loaded and attached to the DNA matches. Next, we identified common ancestors across all of the data. We found one match who had a tree on both sites. By associating her DNA from both sites to one tree and deleting the extra tree we avoided her showing as a common ancestor to herself. Next, we ran hints on her tree to get clues on which ancestral lines to prioritize for our research. After expanding the tree, and re-running the hints tool, we found the common ancestor. This led us to find that she and Dave had common fourth great-grandparents!