Using AncestryDNA’s ThruLines to Further Your Genealogy Research

Have you heard about ThruLines? One of our genealogists shares how to use this tool to extend your family tree.

For many years, I’ve worked on building a large family tree. I always was interested in knowing who my second, third, and fourth cousins were, even though I might never have the opportunity to meet them. As a kid, I wondered if any of my classmates were somehow related to me.  Turns out that some of them were. (Hi, Cuz’n Todd!)

When I received my AncestryDNA results in late 2015, I instantly recognized a few names near the top of my match list, and set to work figuring out who the other close (and sometimes not-so-close) matches were. This effort has continued over time as new matches were added. Because I already had a robust family tree, this process was more straightforward than it might be for someone with little knowledge of their extended family tree. Yet, this process has been time consuming.  Really time consuming.

When AncestryDNA announced ThruLines this February, I checked it out, because I like all the new and shiny tools. I didn’t expect it to help me, but was really surprised when it correctly identified some of my tested relatives with barely-existent family trees, or trees that were incorrect in really unusual ways. With a click of a button, ThruLines identified some of my matches that I had spent a LONG time studying. It also found some connections that I hadn’t.

My family tree is offline, with just a subset of that tree (which has remained unchanged for years) attached to my AncestryDNA results. Any additions I have made to my family tree are not reflected in any of my trees on Ancestry; thus, ThruLines found these connections on its own, and not from my family tree.

What is ThruLines?

ThruLines shows identified descendants of a given ancestor who have tested with AncestryDNA, and share DNA with the tester whose results are being reviewed. The ancestral path between the common ancestor and each DNA match is provided, along with predicted relationship and amount of DNA shared.

ThruLines replaces Shared Ancestor Hints, and greatly expands upon the data provided by those hints. AncestryDNA has long been able to compare family trees for a tester and any given match, and note common ancestors between the two trees. ThruLines builds upon this by including data from all public or private searchable trees in the AncestryDNA network. When a genetic cousin’s tree is incomplete, ThruLines provides the opportunity to view connections that would take a great deal of research to find manually. Of course, any time data from family trees is utilized, all of the names/dates/places should be viewed as clues rather than fact. Comparisons between multiple family trees are only as accurate as the data contained in said trees.

ThruLines is accessible from “Your DNA Results Summary” under DNA in the top menu bar on Ancestry. To have access to ThruLines data, your family tree must be public, and linked to your DNA test. To check this, go to Your DNA Results Summary, and click on the Settings button (near the top right corner of the screen). Then follow the instructions in the Family Tree Linking section.

ThruLines is still in Beta testing. Several revisions were made shortly after introduction, but I have not noticed significant changes to my own data in recent weeks. Kudos to the programmers at Ancestry! This was an ambitious undertaking with a relatively smooth implementation. ThruLines is currently available to all AncestryDNA users, but eventually will only be offered to those testers with an Ancestry subscription.

ThruLines Example: John Berger/Eva Lenz family

John Berger and Eva Lenz were my great-great-grandparents. My research has uncovered a great deal of information about Eva’s parents and extended family, but I have not spent much time researching John. He came from Mecklenburg, Germany, to Minnesota as an adult, and I am not aware of any of his siblings or extended family who traveled with him. John is buried in the cemetery at the small Lutheran church where my family has generations of history. I have not personally viewed the church burial registers, but a distant relative reported that names for John’s parents were listed, but no other dates or specific details were provided about them.

Many descendants of John and Eva are in my match lists on the various testing sites. I’ve used the shared matches functionality to identify other DNA matches who share DNA with me and with these known Berger/Lenz descendants. I’ve added notes about these shared matches and have utilized AncestryDNA’s new grouping functionality to create custom groups for these matches.

Relationship View

ThruLines has two views: Relationships and List. Relationships provides a graphical view of descendants of a given ancestor. Here is the Relationships view of my ThruLines for John Berger:

The default Relationship View for a given ancestor shows your connection to that ancestor (although some details may be collapsed), and any other testers who are the only identified descendant of a child of that ancestor (S.B. is an example of this in the above graphic). Data is collapsed for all other children of the target ancestor, but can be expanded by clicking on the “# DNA Matches” link under the child’s name.

To view detail for a specific line (for example, Emma Berger), click on “5 DNA Matches” below her name, which results in the following view.

This provides detail of the ancestral path to each of my DNA matches from that family, my relationship to them, and the amount of DNA shared.

The default Relationship view for John shows that descendants of four of John Berger’s children have tested (and have been identified by ThruLines—in many cases, other descendants will have also tested). John and Eva Berger had a total of six children: the four listed here (Carl, Gustav, Emma, and Otto), along with a son, Henry, and a daughter, Bertha. Henry died around age 20 and had no known children. I reviewed my DNA results in search of Bertha’s descendants (in case they were not detected by ThruLines) and did not find any.

Bertha had a large family, and most of her children had children. This lack of tested descendants represents a research opportunity. Bertha most likely inherited some segments of John Berger’s DNA that her siblings did not, so it would benefit research efforts for this family if her descendants were also in the testing databases. In previous research, I have already identified 14 of Bertha’s grandchildren, born between 1912 and 1938. Focusing on the younger grandchildren is probably my best strategy to find available testers. If no grandchildren are available or willing, I can recruit great-grandchildren, but Bertha’s grandchildren are a better choice, if available.

List View

The List View shows the same descendants as Relationship View, but all in one list.  When using Relationship View, if you expand one line, the other lines will automatically contract. List View provides a great “at a glance” view, with all of the relationship and shared DNA information included. (To easily save the entire list, load this view in the Firefox browser, and then right click, choose “Take a Screenshot”, and then choose “Save full page” (top right of screen).)

Seeing the relationships and amount of shared centimorgans for all relatives in one place provides a good cross-check for your family tree (and the family tree of your matches). If the predicted relationships and centimorgans shared are not consistent with each other, there may be errors in your tree or the trees used to collect this data. This type of view can sometimes make half relationships more apparent than they might be by looking at individual results.

Because John and Eva Berger are my great-great-grandparents, their other great-great-grandchildren are the same generation as me, and except for my close relatives, will be my third cousins. Their great-grandchildren are my second cousins once removed, and great-great-great-grandchildren are my third cousins once removed (check out our handy DNA & Relationship Chart to help decipher who's who).

When choosing DNA testing candidates, it’s optimal to find the oldest available relatives. John and Eva’s children and grandchildren are all deceased, so the oldest available relatives are their great-grandchildren. My list of tested relatives shows one great-grandchild of John and Eva (a second cousin once removed). I happen to know of two other second cousins once removed from this family who have tested (but do not have family trees); however, they too are from the Carl Berger family. Before I started this process, I knew a few great-grandchildren of John and Eva were in my test results, but hadn’t realized that they were all from the Carl Berger family. Recruiting older testers from other lines in the John/Eva Berger family will hopefully reveal other matches that aren’t available in my own match list.

The lack of testers from the Bertha Berger family has already been noted. Further review of the list as a whole shows that coverage is minimal from the Gustav Berger family, and in my own Otto Berger line. If I were to prioritize further testing, I would choose testers in the following order:

• Great-grandchild of Bertha Berger (currently no known testers)
• Great-grandchild of Gustav Berger (currently only one great-great-great-grandchild tester)
• Great-grandchild of Otto Berger (currently two great-great-grandchild testers)
• Great-grandchild of Emma Berger (currently four great-great-grandchild testers)

Ideally, I would ask all of these testers to share their results with me, to allow for better visibility of their distant matches within our family.

When ThruLines Misses the Mark

Observant readers may have noticed in the list above that I am listed correctly in the family of my great-grandfather, Otto Berger, but my sister appears in the family of Otto’s brother, Carl. This type of misclassification might occur if my sister had an error in her family tree, but she does not. Her tree includes our mother and grandmother (with names and dates identical to my family tree), but does not include a father for our grandmother. I reviewed all the family trees in which I could find my grandmother, and none named Carl as her father, nor did Carl have a daughter named Evelyn. It’s a mystery why ThruLines created this connection.

This error is far more obvious when viewing my sister’s results, as people who are actually her second cousins once removed and third cousins display as first cousins once removed and second cousins, with impossibly low centimorgans shared.

Who’s In? Who’s Out?

A total of nine descendants of John and Eva Berger were identified in my ThruLines. This represents just over a third of the Berger/Lenz descendants I have personally identified. Why weren’t the others included? I reviewed data for the descendants that weren’t included, and found that all but two had no family tree at all, or had a family tree that is not linked to their DNA test. As such, we wouldn’t expect their inclusion in ThruLines, as they’ve provided no data to ThruLines other than their username.

Of the two cousins with family trees, one contains only four living people (although one of them is a Berger descendant). Because everyone in that tree is living, I can’t see the details provided. In the other tree, a Berger descendant was included, but without any dates. There are several family trees that include this individual, but her name is not distinctive enough to distinguish her without a date of birth. Based on limited information, I believe ThruLines’s non-inclusion of that match was the right decision.

I also reviewed data for the nine Berger/Lenz descendants who were included in my ThruLines. Four of them had family trees that included John Berger and Eva Lenz (although one had an incorrect name for Eva). One had a private family tree. I can’t see the details, but it’s large enough that I suspect these ancestors were included. Three of the other four family trees ended with one of John and Eva’s grandchildren, and the fourth extended back to one of John and Eva’s children. Four of the nine (nearly half) family trees did not include the common ancestors. These four trees would not have generated Shared Ancestor Hints, but do generate data included in ThruLines.

Why Is This Helpful?

I intentionally chose a simple example for this post, one with a small data set and with testers whose identities I had previously verified. Because I already knew who these people were, what is the benefit to me? Because of the large amount of time I’ve spent doing match identification, that part of ThruLines functionality may be less beneficial to me than it will be to others with less-developed family trees. However, ThruLines is finding connections that I haven’t, so identification of matches (when that identification is correct) still benefits me. I’ll say it again—all connections suggested by ThruLines must be verified!

The data organization and display are what I’m excited about. Many genealogists have documents, spreadsheets, and handwritten notes all over the house to help them organize their DNA matches. Any time the testing companies provide tools to achieve these purposes, I’m all for it.

I’ve previously identified a very small group of matches who share DNA with some of my Berger/Lenz matches, but who are not descendants of John and Eva Berger, and do not share DNA with members of Eva Lenz Berger’s extended family. This group of people may be distant relatives of John Berger. I’ve found common ancestors between them, but it’s a really small group, and the shared DNA falls in a range where they might be a generation or two back from John, but they also could be really, really distant relatives. I just don’t have quite enough data.

Review of my ThruLines data shows me there are significant gaps in the testing coverage of the John Berger family. One entire ancestral line is missing, and coverage is spotty in others. Most important, though, is the realization that the majority of the tested descendants are five or more generations removed from the ancestors of interest (John’s parents). That information was, of course, available to me all along, but the display of data (even when not all matches are included) made these trends much more obvious.

If I want to use DNA to make progress with researching this line, older DNA testers would provide a great benefit. Testers who are a generation older than me may have matches that I don’t have at all, or shared matches with people who are below the shared matches threshold for me. There may be a group of really fascinating matches from this line that I’m not seeing because they all share 18 or 19 centimorgans with me and aren’t appearing on shared match lists, but would be listed as shared match for an older tester if they share just a bit more DNA than I do.

Confirm, Confirm, Confirm!

My father-in-law has a ThruLines connection that surprised me, because it identified a potential ancestor with a surname I didn’t at all recognize. This particular match is said to be a half sixth cousin to my father-in-law, and the algorithm utilized eight family trees to make this connection. Have I added these new ancestors to my family tree?  Absolutely not! There’s a lot of research that needs to be done to confirm this, both because of the number of generations to verify, as well as the fact that this connection involves a couple of generations of people named Smith, which makes the verification even more difficult. However, shared matches with this individual are from the lines I would expect, so I will explore these results, and attempt to confirm or refute the connection.

ThruLines is making amazing connections for many AncestryDNA testers. Of course, there are plenty of erroneous family trees in existence, so data from ThruLines will be incorrect when the underlying family trees are incorrect. I would encourage caution whenever common surnames are involved, and with more distant connections. The fact that you share 6 or 8 centimorgans with ten people who descend from a common ancestor does not mean that you also descend from that ancestor. ThruLines data should never be accepted blindly, but with proper confirmation, the clues provided by this tool have the potential to guide research in new directions.

Our resident DNA experts would be happy to help you analyze and make sense of the results you get from any of the major DNA testing companies. Give us a call or drop us an email and let us know how we can help you!

Understanding Your AncestryDNA Matches

Your AncestryDNA results are in. Maybe you’ve always been interested in genealogy, or perhaps you received the test as a gift, and really have no idea what to expect. There is a learning curve to understand DNA test results, even for seasoned genealogists.

AncestryDNA’s test is an autosomal DNA test, which looks at chromosomes 1-22 and the X-chromosome and provides information about many relationships and ancestors from both paternal and maternal ancestral lines. This post focuses on AncestryDNA’s results, but several other companies which offer autosomal DNA testing, including MyHeritage, 23andMe, and Family Tree DNA, provide similar features.

Autosomal DNA results include ethnicity predictions, as well as a “match list” – a listing of other testers, or genetic cousins, with whom you share DNA. While the ethnicity prediction is the feature that prompts many people to test, the cousin matching features contain significant information about your biological family and their origins, and can provide much more information for genealogical research once you know how to navigate the results.

Getting Started

“What do I do with these results?  I don’t recognize any of the names on my match list.”

This is a common sentiment upon first viewing one’s match list, and it can be disconcerting to not immediately see expected surnames. Before you start formulating “switched at birth” theories, take a few moments to learn about how to interpret these results.

Your main AncestryDNA page will look something like this. Click on the green “View All DNA Matches” button to view your match list.

If you have matches with predicted relationships closer than a 2^nd cousin whose names you don’t recognize, you may have uncovered a surprise. Even then, don’t be too quick to judge, because your Close Family match with “surferdood456” may turn out to be Uncle Henry, who hasn’t mentioned his new hobby.

Identifying Relationship Possibilities

Your genetic cousin matches are listed from closest to most distant. Each company has a different algorithm for classification, but typically matches are sorted from highest to lowest number of centimorgans shared. Centimorgans (cM) are the unit of measurement used to describe the amount of DNA shared between two people. The more you share with a match, the more likely you are to be closely related. For each match on your match list, the total number of cM is given, along with a predicted relationship.

AncestryDNA has 7 different groups for predicted relationships: Parent/Child, Immediate Family (full siblings), Close Family (half siblings, grandparent/grandchild, aunt/uncle/niece/nephew), 1^st cousin, 2^nd cousin, 3^rd cousin, 4^th cousin, Distant Cousin.

The number of centimorgans shared with any given genetic cousin is now displayed on the main match page. Previously this information was found by clicking on that person’s username, and then on the lower case “i” in the Predicted Relationship section.

The relationship “buckets” from the testing companies are broad and don’t specify all of the possibilities, such as half or “once removed” relationships. The number of centimorgans shared is more informative than the predicted relationship from the testing company; however, the number itself is not enough to determine a specific relationship. For nearly all relationships, there are other relationships that have the same expected amount of DNA shared. A 1^st cousin relationship has the same expected amount of DNA shared as a half aunt, or a great-grandparent.

Once the number of centimorgans shared between two individuals is known, many resources are available to suggest possible relationships. One resource that we particularly like is the DNA Painter Shared cM Project 3.0 tool v4.[1] This tool is a collaborative effort that combines user-submitted data with statistical probabilities. Many DNA relationship charts show cM ranges for various relationships, but there is always overlap between the ranges. This tool gives the probability of each relationship group.

For our 165 cM example above, the Shared cM Project probability tool shows that Half 2^nd cousin or its equivalents (2^nd cousin once removed, Half 1^st cousin twice removed, 1^st cousin three times removed) are the most likely options, although 2^nd cousin or equivalent is a definite possibility. Other options below are less likely, but still in the realm of possibility.

Use of Family Trees to Narrow Relationship Possibilities

Many new DNA testers will find that their matches are all 2^nd cousins or more distant. Most of us are not acquainted with the majority of our 2^nd, 3^rd, or 4^th cousins, so it’s typical to not recognize the names. Ancestry.com is well known for its large collection of family trees, and many AncestryDNA users have family trees that they have associated with their DNA results. These trees greatly simplify the process of understanding who a genetic cousin is, and how you are related to them. An Ancestry.com subscription is required to view other users’ family trees; however, you will be able to view ‘Common Ancestors' listed even without a subscription if you upload your family tree.

The following examples are the top seven matches from my match list. These matches have a variety of family tree options, including large and small family trees, unattached family trees, private trees, and no tree. The right side of the match list (next to the green “View Match” button) shows family tree information for each user.

Public Family Trees

Match #1 and I share 263 cM. Clicking on either his username or the green “View Match” button shows more information about him. His family tree shows that his mother was my dad's 1^st cousin. We share great-grandparents, so the amount of DNA shared and the 2^nd cousin designation make perfect sense.

Match #2 bears my maternal grandmother’s maiden name, and sure enough, his tree shows his great-grandparents, who are my great-great-grandparents, meaning a 2^nd cousin once removed between us, and the 209 cM shared is consistent with that relationship. Likewise, Match #4 is the only name originally recognized – a known 3^rd cousin with a large family tree.

Unattached Family Trees

The main match list page shows no family tree for Match #5, but clicking through to her page reveals a dropdown box containing a family tree:

The family trees for Matches #1, 2 and 4 were linked to their DNA results. Match #5 has not performed that step, and her tree is what is known as an unattached family tree. Caution is required with unattached family trees, as the tree may not include the tester, but many times they do, and may contain valuable information. In this case, the tree was informative, and showed that Match #5 is a 1^st cousin of Match #2.

Match #3 had a small unattached tree which only listed herself and the name of her deceased husband.  We’ll get back to her in a few paragraphs …

Private Family Trees

When the match page shows the number of people in a family tree followed by a padlock, the user’s family tree is private, and as such, no information is viewable by other users. Some users will grant access to their tree if you ask politely, but oftentimes, the private trees are still searchable. A blue “Search Matches” box appears at the top of the match list. Clicking on that box will display surname and location filters which can be used for searching. Match #7's test was administered by another person, so the only information initially available about her was her initials. A search for my maiden name included Match #7 in the list of results. Further searching of other known names in that family line helped me determine the identity of Match #7.

Shared Matches

Of course, many of the users with a “No family tree” designation actually have no family tree. This was true of Match #6. The Shared Matches feature was key to understanding how both Match #6 and Match #3 were related to me.

The Shared Matches feature is located near the top of the screen when viewing details about an individual match, in a black box in three parts (Pedigree and Surnames, Shared Matches, and Map and Locations). The headings for these boxes are Surnames, Matches, and Locations on Ancestry’s mobile app. Pedigree and Surnames is the default, and will always display initially.

Clicking on Shared Matches will display a new match list, containing genetic cousins who share DNA with the tester and the currently viewed match. Most of the time, when three people share DNA with each other, the three all have a common ancestor. Occasionally the relationships between the three parties will involve three different sets of ancestors, so caution should be utilized in analyzing these results. This situation is rare but should be kept in mind.

Shared Matches functionality can be used to categorize one’s test results into groups. Tracking those groups deserves its own post, but typical strategies include use of the Notes functionality on a user’s match page, word processing or spreadsheet software, or even handwritten notes.

The list of shared matches for Match #6 contained Match #2 and Match #5, who already were identified as relatives of my maternal grandmother. This branch of the family was from Minnesota, which has excellent online birth, marriage, and death records available. Match #6 has an unusual surname, so marriage records involving her first name and surname were searched, resulting in a handful of candidates. These candidates were researched, and one of them was the daughter of my grandmother’s 1^st cousin.

Using These Techniques for Unknown Family

Our discussion thus far has assumed some knowledge of one’s biological family, but the same tools and techniques are used in search for unknown family, whether that be unknown parents, grandparents, or more distant ancestors. Family trees for groups of shared matches can be studied to find common surnames and ancestral couples.

Match #3 shares DNA with a large group of testers who are known descendants of one of my sets of great-great-grandparents. Investigation of Match #3’s small unattached tree showed no connection to the expected family. Contact with Match #3 revealed that she was a late-discovery adoptee, with limited information about her biological mother and no information about her biological father. The techniques discussed in this post, along with extensive family tree building, were used to identify Match #3’s biological parents.

Contact with Other Users

Other AncestryDNA users may be contacted by clicking on the match’s username, and then pressing the green “Send Message” near the top right of the match’s page. (Some users have disabled this feature, in which case no green button is displayed). We recommend keeping initial contact short, while sharing only those details that are pertinent. If the identity of the genetic cousin is a complete mystery, asking them to share names of their great-grandparents is often a good approach.

Studying one’s AncestryDNA results can yield a wealth of information and clues about one’s family. We recommend initial focus on the closest matches with family trees, followed by use of Shared Matches functionality to group matches by line of descent. Understanding the ancestral connections to your genetic cousin matches may take a bit of time, but is a worthwhile pursuit to help further your genealogical research.

Legacy Tree Genealogists is not affiliated with any of the companies mentioned in this blog post. However, our resident DNA experts would be happy to help you analyze and make sense of the results you get from any of them. Contact us and let us know how we can help you.

[1]«The Shared cM Project 3.0 tool v4,» DNA Painter, https://dnapainter.com/tools/sharedcmv4, accessed August 2018.