July 2020 Update: Please note that Ancestry issues a cease-and-desist order against Genetic Affairs, and this tool no longer works at Ancestry. The great news is that it still works at the other vendors, and you can ask Ancestry matches to transfer, which is free.
If you’re an adoptee or searching for an unknown parent or ancestor, AutoPedigree is just what you’ve been waiting for.
By now, we’re all familiar with Genetic Affairs who launched in 2018 with their signature autocluster tool. AutoCluster groups your matches into clusters by who your matches match with each other, in addition to you.
A year later, in December 2019, Genetic Affairs introduced AutoTree, automated tree reconstruction based on your matches trees at Ancestry and Family Finder at Family Tree DNA, even if you don’t have a tree.
Now, Genetic Affairs has introduced AutoPedigree, a combination of the AutoTree reconstruction technology combined with WATO, What Are the Odds, as seen here at DNAPainter. WATO is a statistical probability technique developed by the DNAGeek that allows users to review possible positions in a tree for where they best fit.
Here’s the progressive functionality of how the three Genetic Affairs tools, combined, function:
- AutoCluster groups people based on if they match you and each other
- AutoTree finds common ancestors for trees from each cluster
- Next, AutoTree finds the trees of all matches combined, including from trees of your DNA matches not in clusters
- AutoPedigree checks to see if a common ancestor tree meets the minimum requirement which is (at least) 3 matches of greater to or equal to 30-40 cM. If yes, an AutoPedigree with hypotheses is created based on the common ancestor of the matching people.
- Combined AutoPedigrees then reviews all AutoTrees and AutoPedigrees that have common ancestors and combine them into larger trees.
Let’s look at examples, beginning with DNAPainter who first implemented a form of WATO.
Let’s say you’re trying to figure out how you’re related to a group of people who descend from a specific ancestral couple. This is particularly useful for someone seeking unknown parents or other unknown relationships.
DNA tools are always from the perspective of the tester, the person whose kit is being utilized.
At DNAPainter, you manually create the pedigree chart beginning with a common couple and creating branches to all of their descendants that you match.
This example at DNAPainter shows the matches with their cM amounts in yellow boxes.
The tester doesn’t know where they fit in this pedigree chart, so they add other known lines and create hypothesis placeholder possibilities in light blue.
In other words, if you’re searching for your mother and you were born in 1970, you know that your mother was likely born between 1925 (if she was 45 when she gave birth to you) and 1955 (if she was 15 when she gave birth to you.) Therefore, in the family you create, you’d search for parents who could have given birth to children during those years and create hypothetical children in those tree locations.
The WATO tool then utilizes the combination of expected cMs at that position to create scores for each hypothesis position based on how closely or distantly you match other members of that extended family.
The Shared cM Project, created and recently updated by Blaine Bettinger is used as the foundation for the expected centimorgan (cM) ranges of each relationship. DNAPainter has automated the possible relationships for any given matching cM amount, here.
In the graphic above, you can see that the best hypothesis is #2 with a score of 1, followed by #4 and #5 with scores of 3 each. Hypothesis 1 has a score of 63.8979 and hypothesis 3 has a score of 383.
You’ll need to scroll to the bottom to determine which of the various hypothesis are the more likely.
Using DNAPainter’s WATO implementation requires you to create the pedigree tree to test the hypothesis. The benefit of this is that you can construct the actual pedigree as known based on genealogical research. The down-side, of course, is that you have to do the research to current in each line to be able to create the pedigree accurately, and that’s a long and sometimes difficult manual process.
Genetic Affairs and WATO
Genetic Affairs takes a different approach to WATO. Genetic Affairs removes the need for hand entry by scanning your matches at Ancestry and Family Tree DNA, automatically creating pedigrees based on your matches’ trees. In addition, Genetic Affairs automatically creates multiple hypotheses. You may need to utilize both approaches, meaning Genetic Affairs and DNAPainter, depending on who has tested, tree completeness at the vendors, and other factors.
The great news is that you can import the Genetic Affairs reconstructed trees into DNAPainter’s WATO tool instead of creating the pedigrees from scratch. Of course, Genetic Affairs can only use the trees someone has entered. You, on the other hand, can create a more complete tree at DNAPainter.
Combining the two tools leverages the unique and best features of both.
Genetic Affairs AutoPedigree Options
Recently, Genetic Affairs released AutoPedigree, their new tool that utilizes the reconstructed AutoTrees+WATO to place the tester in the most likely region or locations in the reconstructed tree.
Let’s take a look at an example. I’m using my own kit to see what kind of results and hypotheses exist for where I fit in the tree reconstructed from my matches and their trees.
If you actually do have a tree, the AutoTree portion will simply be counted as an equal tree to everyone else’s trees, but AutoPedigree will ignore your tree, creating hypotheses as if it doesn’t exist. That’s great for adoptees who may have hypothetical trees in progress, because that tree is disregarded.
First, sign on to your account at Genetic Affairs and select the AutoPedigree option for either Ancestry or Family Tree DNA which reconstructs trees and generates hypotheses automatically. For AutoPedigree construction, you cannot combine the results from Ancestry and FamilyTreeDNA like you can when reconstructing trees alone. You’ll need to do an AutoPedigree run for each vendor. The good news is that while Ancestry has more testers and matches, FamilyTreeDNA has many testers stretching back 20 years or so in the past who passed away before testing became available at Ancestry. Often, their testers reach back a generation or two further. You can easily transfer Ancestry (and other) results to Family Tree DNA for free to obtain more matches – step-by-step instructions here.
At Genetic Affairs, you should also consider including half-relations, especially if you are dealing with an unknown parent situation. Selecting half-relationships generates very large trees, so you might want to do the first run without, then a second run with half relationships selected.
I ran the program and opened the resulting email with the zip file. Saving that file automatically unzips for me, displaying the following 5 files and folders.
Clicking on the AutoCluster HTML link reveals the now-familiar clusters, shown below.
I have a total of 26 clusters, only partially shown above. My first peach cluster and my 9th blue cluster are huge.
That’s great news because it means that I have a lot to work with.
Next, you’ll want to click to open your AutoPedigree folder.
For each cluster, you’ll have a corresponding AutoPedigree file if an AutoPedigree can be generated from the trees of the people in that cluster.
My first cluster is simply too large to show successfully in blog format, so I’m selecting a smaller cluster, #21, shown below with the red arrow, with only 6 members. Why so small, you ask? In part, because I want to illustrate the fact that you really don’t need a lot of matches for the AutoPedigree tool to be useful.
Note also that this entire group of clusters (blue through brown) has members in more than one cluster, indicated by the grey cells that mean someone is a member of at least 2 clusters. That tells me that I need to include the information from those clusters too in my analysis. Fortunately, Genetic Affairs realizes that and provides a combined AutoPedigree tool for that as well, which we will cover later in the article. Just note for now that the blue through brown clusters seem to be related to cluster 21.
Let’s look at cluster 21.
In the AutoPedigree folder, you’ll see cluster files when there are trees available to create pedigrees for individual clusters. If you’re lucky, you’ll find 2 files for some clusters.
At the top of each cluster AutoPedigree file, Genetic Affairs shows you the home couple of the descendant group shown in the matches and their corresponding trees.
I don’t expect you to be able to read everything in the above pedigree chart, just note the matches and arrows.
You can see three of my cousins who match, labeled with “Ancestry.” You also see branches that generate a viable hypothesis. When generating AutoPedigrees, Genetic Affairs truncates any branches that cannot result in a viable hypothesis for placing the tester in a viable location on the tree, so you may not see all matches.
On the top branch, you’ll see hyp-1-child1 which is the first hypothesis, with the first child. Their child is hyp-2- child2, and their child is hyp-3-child3. The tester (me, in this case) cannot be the persons shown with red flags, called badges, based on how I match other people and other tree information such as birth and death dates.
Think of a stoplight, red=no, green are your best bets and the rest are yellow, meaning maybe. AutoPedigree makes no decisions, only shows you options, and calculated mathematically how probable each location is to be correct.
Remember, these “children,” meaning hypothesis 1-child 1 may or may not have actually existed. These relationships are hypothetical showing you that IF these people existed, where the tester could appear on the tree.
We know that I don’t fit on the branch above hypothesis 1, because I only match the descendant of Adam Lentz at 44.2 cM which is statistically too low for me to also inhabit that branch.
I’ve included half relationships, so we see hyp-7-child1-half too, which is a half-sibling.
The rankings for hypotheses 1, 2, and 7 all have red badges, meaning not possible, so they have a score of 0. Hypothesis 3 and 8 are possible, with a ranking of 16, respectively.
Looking now at the next segment of the tree, you see that based on how I match my Deatsman and Hartman cousins, I can potentially fit in any portion of the tree with green badges (in the red boxes) or yellow badges.
You can also see where I actually fit in the tree. HOWEVER, that placement is from AutoTree, the tree reconstruction portion, based on the fact that I have a tree (or someone has a tree with me in it). My own tree is ignored for hypothesis generation for the AutoPedigree hypothesis generation portion.
Had my first cousins once removed through my grandfather John Ferverda’s brother, Roscoe, tested AND HAD A TREE, there would have been no question where I fit based on how I match them.
As it turns out they did test, but provided no tree meaning that Genetic Affairs had no tree to work with.
Remember that I mentioned that my first cluster was huge. Many more matches mean that Genetic Affairs has more to work with. From that cluster, here’s an example of a hypothesis being accurate.
You can see the hypothetical line beneath my own line, with hypothesis 104, 105, 106, 107, 108. The AutoTree portion of my tree is shown above, with my father and grandparents and my name in the green block. The AutoPedigree portion ignores my own tree, therefore generating the hypothesis that’s where I could fit with a rank of 2. And yes, that’s exactly where I fit in the tree.
In this case, there were some hypotheses ranked at 1, but they were incorrect, so be sure to evaluate all good (green) options, then yellow, in that order.
Genetic Affairs cannot work with 23andMe results for AutoPedigree because 23andMe doesn’t provide or support trees on their site. AutoClusters are integrated at MyHeritage, but not the AutoTree or AutoPedigree functions, and they cannot be run separately.
That leaves Family Tree DNA and Ancestry.
After evaluating each of the AutoPedigrees generated for each cluster for which an AutoPedigree can be generated, click on the various cluster combined autopedigrees.
You can see that for cluster 1, I have 7 separate AutoPedigrees based on common ancestors that were different. I have 3 AutoPedigrees also for cluster 9, and 2 AutoPedigrees for 15, 21, and 24.
I have no AutoPedigrees for clusters 2, 3, 5, 6, 7, 8, 14, 17, 18, and 22.
Moving to the combined clusters, the numbers of which are NOT correlated to the clusters themselves, Genetic Affairs has searched trees and combined ancestors in various clusters together when common ancestors were found.
Remember that I asked you to note that the above blue through brown clusters seem to have commonality between the clusters based on grey cell matches who are found in multiple groups? In fact, these people do share common ancestors, with a large combined AutoPedigree being generated from those multiple clusters.
I know you can’t read the tree in the image that follows. I’m only including it so you’ll see the scale of that portion of my tree that can be reconstructed from my matches with hypotheses of where I fit.
These larger combined pedigrees are very useful to tie the clusters together and understand how you match numerous people who descend from the same larger ancestral group, further back in time.
Integration with DNAPainter
Each AutoPedigree file and combined cluster AutoPedigree file in the AutoPedigree folder is provided in WATO format, allowing you to import them into DNAPainter’s WATO tool.
You can manually flesh out the trees based on actual genealogy in WATO at DNAPainter, manually add matches from GEDmatch, 23andMe or MyHeritage or matches from vendors where your matches trees may not exist but you know how your match connects to you.
Your AutoTree Ancestors
But wait, there’s more.
If you click on the Ancestors folder, you’ll see 5 options for tree generations 3-7.
My three-generation auto-generated reconstructed tree looks like this:
Selecting the 5th generation level displays Jacob Lentz and Frederica Ruhle, the couple shown in the AutoCluster 21 and AutoPedigree examples earlier. The color-coding indicates the source of the ancestors in that position.
You will also note that Genetic Affairs indicates how many matches I have that share this common ancestor along with which clusters to view for matches relevant to specific ancestors. How cool is this?!!
Remember that you can also import the genetic match information for each AutoTree cluster found at Family Tree DNA into DNAPainter to paint those matches on your chromosomes using DNAPainter’s Cluster Auto Painter.
If you run AutoCluster for matches at 23andMe, MyHeritage, or FamilyTreeDNA, all vendors who provide segment information, you can also import that cluster segment information into DNAPainter for chromosome painting.
However, from that list of vendors, you can only generate AutoTrees and AutoPedigrees at Family Tree DNA. Given this, it’s in your best interest for your matches to test at or upload their DNA (plus tree) to Family Tree DNA who supports trees AND provides segment information, both, and where you can run AutoTree and AutoPedigree.
Have you painted your clusters or generated AutoTrees? If you’re an adoptee or looking for an unknown parent or grandparent, the new AutoPedigree function is exactly what you need.
Genetic Affairs provides complete instructions for AutoPedigree in this newsletter, along with a user manual here, and the Facebook Genetic Affairs User Group can be found here.
I wrote the introductory article, AutoClustering by Genetic Affairs, here, and Genetic Affairs Reconstructs Trees from Genetic Clusters – Even Without Your Tree or Common Ancestors, here. You can read about DNAPainter, here.
Transfer your DNA file, for free, from Ancestry to Family Tree DNA or MyHeritage, by following the easy instructions, here.
Have fun! Your ancestors are waiting.
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- FamilyTreeDNA – Y, mitochondrial and autosomal DNA testing
- MyHeritage DNA – ancestry autosomal DNA only, not health
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload – transfer your results from other vendors free
- AncestryDNA – autosomal DNA only
- 23andMe Ancestry – autosomal DNA only, no Health
- 23andMe Ancestry Plus Health
Genealogy Products and Services
- MyHeritage FREE Tree Builder – genealogy software for your computer
- MyHeritage Subscription with Free Trial
- Legacy Family Tree Webinars – genealogy and DNA classes, subscription based, some free
- Legacy Family Tree Software – genealogy software for your computer
- Charting Companion – Charts and Reports to use with your genealogy software or FamilySearch
- Legacy Tree Genealogists – professional genealogy research
Thanks Roberta! I enjoyed reading it. Concerning the large trees, they sometimes start much larger. I prune the branches if there are no valid hypotheses downstream so that already improves the visualization. Another tip for users is to look at the generation level of the proposed hypotheses. Some hypotheses might have a better rank but are not compatible with the expected number of generations of the tested person. Also, AutoPedigree only employs matches sharing at least 30-40 cM (based on the settings). However, there can be DNA matches sharing less than the min cM in the AutoPedigree that can help with the analysis. Imagine having a branch with some good hypotheses but also a DNA match that shares 25 cM. That particular match will then nullify the generated hypotheses of that branch because the match was ignored by AutoPedigree.
According to the manual, AutoPedigree is only available for Ancestry profiles, yet your blog instructions refer to using it with FTDNA profiles.
Hi Neil, sorry for the confusion. The AutoTree is also available for FTDNA but I see I used a screenshot of the Ancestry interface to illustrate how to use it. But if you click on the AutoCluster/AutoTree link for an FTDNA profile, you’ll see the AutoPedigree option.
Roberta, in the WATO tree example, you state that hypothesis 1 is irrelevant. The details below the graphic state that hypothesis 1 is 167 times more likely than the next highest hypothesis…which is hypothesis 3. Seems pretty evident that hypothesis 1 is the most likely relationship…no?
Yes, you’re right. DNAPainter calculates the probabilities with a higher number meaning a probable better match, where Genetic Affairs uses a lower number to indicate a better match. Fixing this now. Good eye, and thanks for reading closely.
Great review. I can’t wait to play with the results.
In the autocluster report, I had no idea what the gray squares meant. Now I will go over their names very carefully in comparison withe the 26 autoclusters listed in my report. KMC
Hi Kathy, Dana blogged some time ago about the grey cells: https://www.danaleeds.com/what-are-the-grey-squares-on-a-dna-match-cluster-chart/ 🙂
I am an adoptee and I am so excited about this I almost can’t stand it Thank you for Writing This and all that you do in the DNA WORLD!!! 🤞Alongside with DNA PAINTER this I am in hope of many answers I use the Shared cm project, however everytime I try DNA PAINTER I think I know what I’m doing and I’m not…lol
Hi Maria, I sincerely hope you get some good clues from there. Let me know if you have any questions or suggestions. Also, if you have some good results I would like to hear about it!
Pingback: Genetic Genealogy at 20 Years: Where Have We Been, Where Are We Going and What’s Important? | DNAeXplained – Genetic Genealogy