Of Mirroring And Shared Ancestors: Exploiting AncestryDNA To Find Biological Families


What we aim to accomplish in this post:

  1. Provide definitions and an overview of methods for finding biological families of those who don’t know one or both biological parents, using tools at AncestryDNA.
  2. Discuss possible future methods that may be helpful in reaching the same goals.


Finding a bio-ancestor is like finding the horizon on a foggy day - keeping looking and you might find it.

Looking for a bio-parent may feel like looking for the horizon on a foggy day – but keeping looking and you might find what you seek.

Edit 10 May 2016: The May 2016 change in the matching algorithm at AncestryDNA makes some of the details in the PDF out of date, but the basic ideas still apply.

One of the more common uses of autosomal DNA matching products such as 23andMe, AncestryDNA, and FTDNA Family Finder, is for people searching for their biological families, including:

  1. Adoptees searching for their biological ancestors;
  2. Those who discover that one of their presumed biological parents turns out not to be so – often a shocking surprise, but sometimes also a deeply and long held suspicion that turns out to be true;
  3. Those who never knew one or both parents (usually but not always a missing father.)


The processes described here are not new nor invented by me.  An entire community of people dedicated to genetic genealogy and helping others (such as adoptees) have arisen and the methods being used are continually improving, especially as creative software developers bring forward new tools to make use of autosomal DNA matches.

However, on forums and Facebook groups it seems that the same questions get asked continually about finding family using DNA, sometimes posters presenting the same question within minutes of each other. Thus we believe that there is still a need for more educational material on this topic.

In the presentation below, as a PDF file, is an overview of how to exploit one’s DNA test at AncestryDNA, including:

  1. mirror pedigrees;
  2. shared matches;
  3. ancestor harvesting.

This presentation is not an exhaustive exploration of these topics, but hopefully will be helpful to many seeking bio-families.

Data Mining Through Mirroring

Download (PDF, 3.35MB)



I believe there are two possible tool types that AncestryDNA could give us that would help in research, both for more traditional genealogy goals (i.e., finding ancestors beyond those we knew personally) as well as for those searching for birth families:

Surname analysis

See my post “Surnames, 23andMe, and AncestryDNA: Making the Most of Match Counts and “Enrichment”

Geographic analysis

Our ancestors lived in locales at specific times and much family history work depends upon locating our ancestors within their landscape.  Not enough of this is done by novices to family history but often analyzing locations are key to unraveling complex families.  Anyone who has spent some time reviewing family trees knows that the location fields for deaths and births and many of the events in-between are often left off, or given inaccurately.

AncestryDNA does provide a map tool for every DNA match, to see where the match’s ancestors were born, according to the tree to which the DNA kit is attached. This is very useful for my matches where surnames are of little to no use (such as my Norwegian ancestors).  AncestryDNA also provides a filter for searching matches by location.  Both are limited by the difficulties in conforming geographic names to standards, as well as the sad fact that even those of our matches who have family trees often do not have birth locations for their ancestors.

Still, it ought to be possible for an automated process to sort the ancestors of our matches into location groups.  In this regard even the rudimentary Map View in the DNA Relatives tool at 23andMe offers a capability not at AncestryDNA.  For example, there are  3,144 counties (or county equivalents) in the US – why can’t we have a heat map at the county level, for the birth and death counties of the ancestors of our matches?  There are analogous (to the US county system) geographical divisions in many nations that could also be heat-mapped.

In the bigger picture we are discussing what might be called “data mining”. We want to use automated systems to find patterns in large data sets. Here we have two large data sets: DNA matches, and Family Trees.

The collected family histories available both in traditional publications (books, journals, monographs) and electronically represent our families and their culture through the collective historiography work of many thousands of people, some professional genealogists, many of them not.

Connecting this family history to that of our physical inheritance – as presented through our genotypes, created by these direct to consumer DNA testing companies – is the goal of everyone in genetic genealogy.  Of the companies involved, so far AncestryDNA has proven for me and many others to be the most able to reach this goal (though as a product it is still wanting for some basic tools, such as a chromosome browser for those wishing to do segment mapping.)   I’m hoping someday they will add tools in the categories I list above.


  1. If I share a certain % DNA with someone who has copied my tree, the only element that is different from my own DNA is their DNA. Am I to assume that Ancestry has determined which side the DNA matches and through whom on each side? I find their adjectives such as “potential” “probably related” etc. misleading.
    With DNA I want definite answers about my DNA.

    • @JOY UPTON: In most cases the only way to tell from which side a DNA match arises is to have a known relative from either your mother or father also tested and match that same person. The most direct method of course is to have a parent tested but that is often not possible.

      An exception is for a male who has a match on the his X chromosome – he only got that from his mother.

      In some cases, if one can map deep ancestry (i.e., “ethnicity”) to chromosome regions, and one’s parents came from distinct (that is, continental level) population groups, a match on a chromosome region which you know is from one parent and the match only has ancestry from that continent, you may conclude that the match is from the side of the parent whose deep ancestry coincides with the match. However, the problem here is that many people have deep ancestry from places they would not ever have guessed (such as self identifying “white” Americans who have colonial-era African ancestors.)

  2. Thank you for your well written and illustrated article on Mirroring. I especially appreciate this ,”Our ancestors lived in locales at specific times and much family history work depends upon locating our ancestors within their landscape. Not enough of this is done by novices to family history but often analyzing locations are key to unraveling complex families. ” Which is the key to genealogical research period. People did things for reasons…..you must put their lives in context to understand what they did, where they lived and why they moved.

    • @WHEATONWOOD: Thanks for the compliment. There are so many people with so many questions (often the same) that I think much more needs to be done in reaching out to people seeking answers.

    • That is exactly how I could narrow down my matches to paternal/maternal sides. My maternal side had been in one area for a very long time. When I saw these matches from another part of the country going back hundreds of years I knew I had found a gold mine. I have a ways to go yet, but with patience and new matches every day I will get my answer soon enough.
      Kelly Jean

  3. How do I upload my info here?

    • @ KATHIE : Not sure what you mean. This is a blog. I presume you mean some sort of DNA data, and the usual sites for such things are gedmatch.com and FTDNA.

  4. Thanks for the very useful information. I did have one question…how does one “attach” dna to another person on Ancestry? (Given my ineptitude with all this, I’m sure that there is a big “Attach To Another Person” button just below the fold that I can’t see, so please forgive me in advance!)

    • On your AncestryDNA page for the kit there is a gear icon in the upper right. Click on that and you go to the administration page, and there you will find that you can attach the kit to someone in any tree you for whom you administrate.

  5. A very clear and thorough description of how to create a mirror tree. Great job! May I attach a link to the blog in an article I am writing for adoptees looking for family? I will link back to your original article.

  6. Okay My question is When I start a mirror tree Do I use the matches (home person) name to start or Do I put my name in as self then do tree?

  7. I did this for my top 4th cousin match that hasn’t responded to messages. I mirrored their tree and placed myself at his position. I now have 6 shared ancestor hints and every single one of them is along the same surname line… the closest was a 4th cousin match with great-grandfather at common ancestor. Is this where I fill out all the gg children and grandchildren? I guess I’m confused with who I’m suppose to attach my DNA kit to next.

    • The purpose of the mirror technique is to find out from which line in your match’s pedigree your most recent common ancestor with them lay. It sounds like you have done this.

      It is not clear to me when you write “…with great-grandfather at common ancestor” to which match the ancestor is a 1x great grandfather – your original match, or the second one.

      If you’re working with an estimated 4th cousin whose pedigree you are mirroring, since you’ve discovered a line on which you have shared hints, I’d follow that line to the 3rd great grandparents of your mirrored match. The for that couple work out their children, the spouses of their children, and then the next generation and their spouses. Then finally you can check the spouses to see if their family names or locations also show up in your match list.

      • I cannot print the mirror tree pdf file. Even though I can printer other files from my computer. I wondered if you would mind emailing me the file at dee.marsh@asu.edu I would appreciate your help since I am just starting a mirror tree and value this information.

  8. Immediately under the PDF viewing window in this post is a “Download” link. Clicking on that will open the PDF in your browser window. Or, you can right-click on the link and save the file to your own computer.

  9. Pingback: “I’m an adoptee. Can you help me find my bio parents?” | Oak Grove Genealogy

  10. Gosh, I wish there were some way of insinuating non-Ancestry.com DNA onto an Ancestry tree!

    I’m working on a 19th-century adoption. My great-grandmother was adopted as an infant in 1880 — no papers; no genetic link to the adoptive family. I’ve worked this puzzle since the days of microfilm, and, since 2013, using the DNA of my great-grandmother’s last surviving child, who volunteered to join the search at age 100. She and I both tested at 23andMe during the V.3 years.

    Two years ago we got our first break when a 2nd-to-3rd cousin showed up for my aunt. He was unwilling to share genomes but the little he shared about his ancestry, combined with the proven genetic connection, was usable as GPS data re time/place/opportunity and such. He didn’t open the lock but he pointed to others more willing to collaborate, and now we have some new 3rd cousins.

    BUT the birth was an NPE and we haven’t been so lucky with the father. Nothing closer than 4th cousins show up in our databases, and while the mother had to have given birth in a specific place (near where the adoption took place), the father could have been a transient.

    A mirror tree sounds useful, but although I was able to get my aunt’s data into FTDNA and GEDMatch, Ancestry won’t accept outside data files. And my aunt has since died, so what we already have (itself priceless) is all we will have to work with.

    I realize that the very size of Ancestry’s database is one of its best assets (its many inaccurate trees muddy that a bit!), but are there other tools “out there” that can work something similar?

    Feeling inspired by your article.



  11. Angela Blankenship

    Hello and thank you for the information. I’ve created a mirror tree for a DNA cousin who shares 397 cM and have gotten several shared ancestor hints but none of them actually lead to the same ancestor. In the chart, the two sides “appear” to lead to one person, but that shared ancestor is not actually in both trees. To be sure, if you click on the details you can clearly see that one side does not lead to that person at all. Why would this be happening?

    • In the chart, the two sides “appear” to lead to one person, but that shared ancestor is not actually in both trees.

      It’s not clear to me what you mean by “chart” or “both sides”.

      A Shared Ancestor Hint is only generated when two DNA matches have the same person in each of the pedigrees of the test.

      • Angela Blankenship

        I’m not sure how to explain it without a picture. Ancestry gives a shared ancestry hint and provides a diagram with the shared ancestor(s) at the top. It then has the descendants that lead to one of the DNA matches down the left side and the descendants that lead to the other DNA match down the right side. In my instance, the descendants on the right are in fact NOT descendants of the “shared” ancestors.
        I think what was happening is that Ancestry takes some time to update the DNA matches and trees and the results I got weren’t done yet.

  12. Hello, I have been reading about mirror trees and actually have started one. Still waiting for Ancestry to make the connection after 5 days. I am considering doing another DNA test to “play” with, so my original tree and link will stay together. Drawbacks would be that I have to pay for another membership. What do you think? Thank you for any help.

    • These days I mostly discourage people from doing the mirror-tree thing, as it messes up the hinting and DNACircle processes at Ancestry.

      Note that if you do go the route of getting another test you don’t need another Ancestry membership, you just need to buy another kit. Wait until a sale comes around on a kit.

      The idea of using another person’s pedigree as a proxy for your own may still have some merit for the adoptee with absolutely no knowledge of bio-parents.

      But, if you have any knowledge of a bio ancestor, even just one grandparent, I think it is best to make a tree with what you know to be true, and attach your DNA kit to you in a tree with that knowledge. Then, use the plethora of tools now available to suss out the other connections. For example, look at the DNAgedcom website.

      Trees are just databases that one can manipulate extensively. At Ancestry, one can attach and detach parents for an individual at will. I often make changes to families in doing regular genealogy.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.