When I collected the data for my ThruLines Summary post, I noticed an irregularity in the number of matches for Orinda Miles. My ancestor, Phillip Mentzer, has 21 ThruLInes.
Since I am unaware of a second marriage, I expect 21 DNA matches for both Phillip and Orinda Miles. However, I only have THREE matches for Orinda Miles.
When I compiled the data, I colored the text for Orinda’s results in RED so that I would remember this discrepency.
Thinking I had somehow recorded the wrong wife for Phillip Mentzer, I decided to compare my tree with the trees of my DNA matches for Phillip Mentzer. Most of those trees also list Orinda Miles as the wife of Phillip Mentzer. Curious, I looked at the ThruLines for Orinda’s father, Oliver Miles.
That’s when I discovered that I have an additional 15 DNA matches for Orinda Miles, but they are listed as a separate child from my Orinda Miles lines.
Since some records spell her name as Orenda or Orindia, I have added Alternate Names in hopes that I can ‘force’ the ThruLines computers to merge these two Orinda Miles.
Most of those speaking out against ThruLines are either expressing doubt about all of the ‘Evaluate’ suggestions or have found actual errors in their ThruLines. Their point is valid, especially if one doesn’t take the time to document the suggested lineage.
I think the fact that I have tried to document descendants of my second great grandfathers has affected my opinion of ThruLines. Since I already have identified and documented the children, grandchildren and most of the great grandchildren of these ancestors, ThruLines is pulling information from MY tree to draw the connections between my ancestor and the parents of a DNA match – or – often my DNA match. Thus, I don’t have much to evaluate.
However, when I move back a generation, I’m finding that I have been lax about researching the descendants of my 3rd great grandfathers. Thus, I have a lot to ‘Evaluate’ on my ThruLines.
To complete this evaluation, I start at the top and look at what documentation I have and what I might be missing. In particular, I’m concentrating on the following types of sources:
Find a Grave records
Basically, I’m looking for enough documentation to support the family relationships, one generation at a time until reaching my DNA match.
So in my example, I will first check for missing documentation for the Phillip Mentzer, William Andrew Mentzer, and Moses Henry Mentzer.
Based on the need to ‘Evaluate’ Grace and Francis Mentzer, I obviously don’t have the family of Francis A. G. Mentzer in my RootsMagic database. To start the ‘Evaluation’ process, I use the ability of RootsMagic to connect with the FamilySearch tree. I know that the FamilySearch tree is controversial, but I view it as a starting point based on the consensus of other researchers. Thus, I will use RootsMagic to pull down a spouse and children for Francis A. G. Mentzer.
Once I have the family in my RootsMagic database, I will then use RootsMagic’s TreeShare to upload that family to my tree on Ancestry. Shortly after the upload is completed, I will start evaluating Ancestry hints for Francis A. G. Mentzer and his children. As I work with each hint, I am adding events and documentation to my RootsMagic database. After working thru the hints, I again do an upload to my Ancestry tree via RootsMagic’s TreeShare.
Then I start the entire process over again with the next generation down. In this example, that would mean pulling down the spouse and any identified children for Grace Mentzer and her brother, Francis G. Mentzer. Since this generation likely has living children, the FamilySearch tree will probably be incomplete. My next step is to upload to Ancestry so I can use their hinting system to locate records. Since many of their children are likely living, I need to locate records that identify those children. In some cases, I have been able to find birth and/or marriage records for the children. Most of the time, however, I need an obituary that identifies the children. As I find information identifying the children, I add them to my database as living individuals.
Once I have data to support the relationship suggested by ThruLines, I then document the DNA connection. I have created two PRIVATE facts for this documentation: DNAMatch and DNAThruLines:
For my DNA match, I add a DNAMatch fact. I enter 2020 as the SORT DATE and make sure PRIVATE is checked. I then add a source. I use the DNA-Ancestry ThruLines source that I created in RootsMagic. For the ‘Item of Interest’ I enter information about the match. At first, I was only entering the match’s name in this field. However, I have started being more descriptive here to help me identify which of the tests I manage are being matched. Thus, I’m trying to use the following pattern for the ‘Item of Interest’
Initials of tester shares __ cM __ segments with DNAMatch
On the Detail text screen, I add the ThruLines information starting with our common ancestor and working down to the match. I add information about our relationship and about the quantity of DNA shared.
The identity of my match has been whited out in the above image to protect their privacy. Once I have completed filling out the source, I the the MEMORIZE button to copy it.
I then move back a generation and create a DNAThruLines fact again making sure it is marked private and paste the recently copied source. As I document additional cousins, I add their source to this one fact.
I work my way back creating DNAThruLines facts and adding sources until I reach the common ancestor. Instead of continuing that practice back further generations, I SHARE the DNAThruLines fact with the ancestors of that common ancestor. In cases where there is only one wife, I also share the fact with the spouse of the common ancestor.
Judson Foster Crawford is the common ancestor for the DNAThruLines sources shown above. His DNAThruLines fact is shared with his wife, Mary Foster, and their ancestors going back about 3 generations.
By sharing the DNAThruLines fact in this way, I have already documented the DNA relationships thru that one child when I move back a generation. Thus, I only have to document the DNA relationships thru the other children.
By working thru the ThruLines suggestions in this way, I am improving my tree by adding descendants and documentation for those descendants. I am also able to add documentation for my DNA match. I am currently working on DNA matches for my 3rd great grandparents. This is a slow process, but is allowing me to validate those ThruLines suggestions.
Are you overwhelmed with DNA data? Have you ever tried looking at the data in a different way to see what you can learn from it?
The Leeds method of looking at DNA matches is often used for this purpose. Today, I read the blog post, 7-gen-1-sheet, by Ann Raymont. In this post, the author explains how to set up a spreadsheet to display 7 generations of ancestors. Once the spreadsheet is created, color coding can be used to identify patterns such as European roots, lineage society lines or whether a specific source has been used.
Intrigued by how this spreadsheet could be used, I decided to create the page of ancestors. As I was creating the spreadsheet, I decided to use it to look at my ThruLines data. Since I’ve tested myself, my two brothers and my mother, I have four sets of ThruLines. Even though I’ve looked thru this data for each match, I’ve never compared the results.
By adding columns for each of my DNA tests, I was able to record the number of matches for each ancestor from the 4 DNA tests.
Having this data all in one place will help me evaluate my tree in relation to my DNA results. For example, does it indicate an error in my tree if I only have a few matches for that ancestor? Having this data side by side has also allowed me to see that even though I might only have a few matches with descendants of a particular ancestor, my brothers or mother could have quite a few more matches. In those cases, the probability that my tree is accurate increases when I look at all four tests versus looking at just my results.
Now that I have 9 generations of ancestors on my spreadsheet, there are several other ways that I hope to utilize this sheet.
Color code states of residence in 1850
Color code ancestors whom I have found an obituary
Color code ancestors whom I have a Find a Grave source for
Color code potential DAR ancestor lines
Thank you Ann Raymont for sharing your 7-gen-1-sheet method of looking at our ancestors.
As with many other genealogists, I’m struggling with the upcoming loss of small matches from my Ancestry DNA match list. I have no idea how many of these matches I have but I’m guessing it is in the thousands.
Since I will never have the time to work thru thousands upon thousands of matches to figure out how we connect, I’m hoping to use Ancestry’s computer technology to help me. Thus, I’m concentrating on my ThruLines matches, including the “potential ancestors”.
I have done a lot of descendancy research which I believe is helping Ancestry’s ThruLines technology connect me to DNA matches who have very small trees. Seeing the words ‘No Tree’ or ‘Unlinked Tree’ in my list of matches means I will scroll right past the match and will never take the time to figure out our relationship.
Thus, I need my over 100,000 matches to have a searchable tree attached to their DNA test(s).
Recently Jason Lee posted a Facebook poll asking whether readers agree or disagree with the statement, “All ThruLines errors are because of errors in trees.” As of today, 802 people agree with that statement, while I am in the minority of 120 people disagreeing.
The reason, I disagree with the statement is that the ThruLines for my 2nd great grandfather, James Crawford. This ThruLines suggests that William Monroe as a child of James Crawford based on a DNA match with a known 3rd cousin once removed.
Even though this match has a very small tree, she has enough in her tree that our trees should connect. Eugene Beggs, son of Walter Beggs and Ethel Anita Lighter is in both of our trees. My match’s tree includes Eugene Beggs’ father, Walter, but not his mother, Ethel Lighter.
Thus, our common ancestors are Washington Marion Crawford and his wife, Mary Foster. Not only is ThruLines suggesting an incorrect connection on the Crawford line but also on the Foster line. This time, it is suggesting Margaret E. Jordan as a daughter of Zebulon Foster.
In hopes of getting the algorithm to correct these ThruLines, I added this cousin to my tree several months ago. Unfortunately, I haven’t observed any change in the ThruLines. Not only do I have Eugene Beggs and his link to Washington Marion Crawford and Mary Foster in my tree, I have also researched these families and attached sources.
I keep hoping that the computer algorithm for ThruLines will discover this connection and show the correct way our lines connect.
Although I haven’t researched all of the suggested connections thru ThruLines, I haven’t found other errors. For the most part, I have been able to document the suggested connections. Because the ThruLines tool helps me see connections between myself and my DNA matches, I appreciate the information provided. However, I wish there was a way to report this ThruLines error to Ancestry. Unfortunately, I haven’t found such ability.
About halfway thru the video (28:48), Crista explains why I should use actual names and birthdates for living people in my tree. I hadn’t considered that my ‘privitization’ of the names would prevent the computers from being able to match data in my tree with data in a match’s tree. Since I’ve had issues in the past with shared matches not working as expected, I have followed Crista’s advice and used actual names.
Watching this video, I also discovered that the method and terminology for shared matches has changed. Instead of being called ‘shared matches’, it is now called ‘common ancestors’. The terminology change is likely due to the change in how the ‘sharing’ is determined. I believe the shared matches was based on sharing Ancestry hints. With common ancestors, it is determined by comparing tree data.
That’s why Ancestry is recommending that everyone have a tree with at least parents and grandparents. Ancestry’s computers can take that small tree and compare it to all of the other trees on Ancestry. Thus, that small tree might have a parent or grandparent in a larger tree, like mine, that contains a lot of descendants. This is the same technology that is behind the new feature, Ancestry Thrulines.
I decided to test this on my own tree. I had a DNA match that I hadn’t looked at identified as having a common ancestor.
When I clicked on the ‘Common Ancestor’ link, it took me to a comparison page. This screen informed me that my match’s tree was private. However, along the left side of the screen was the suggested common ancestors: Albert Hutchinson and Julia Harding.
When I clicked on Albert Hutchinson, our two lines leading back to Albert Hutchinson were shown.
All of the white boxes on both lines were from my tree! Since I have researched the descendants of Albert Hutchinson, I had enough info in my tree to connect with my match – who only had 11 people in her tree.
Thus, all of my work over the years to research descendants is helping me identify my DNA matches!
Even though I was already familiar with Ancestry’s ThruLines – and appreciate how they are helping me with my DNA matches, I learned several things about how Ancestry’s ‘Common Ancestors’ and ‘ThruLines’ work from this video.
The major tip is to CREATE a tree and ATTACH it to the DNA test.
Make the tree Public OR searchable Private – as long as the tree is searchable, it will help generate ThruLines clues
When the tree is ‘searchable private’ the various generations are shown as PRIVATE in a matches’ view of ThruLines
If possible, add your parents and grandparents to the tree
Use Genealogy Standards when adding information to your tree
Only use maiden names for women in the tree. Using a married name will make it difficult for the computer to match the woman in a tree to the same woman in someone else’s tree.
Don’t use any special characters such as quotation marks, nicknames or symbols in the name fields. Again, this will make it difficult if not impossible for the computer to make the match.
City, County, State, Country format
No abbreviations (For example, spell out the state Kansas instead of using KS)
dd mmm yyyy format. For example: 28 May 2019
I know that place names are a stumbling block for me. In the past, I have abbreviated the state. I’ve also abbreviated the word COUNTY as Co. More recently, I’ve been trying to use the standardized version of a place name. (Thankfully, my genealogy software helps me with this.) However, I still have some of my older, nonstandard, place names in my file. Thus, those place names may be preventing Ancestry’s computers from finding a common ancestor with another match.
As I’m working to meet these standards in my Heartland Genealogy tree, I hope all of my matches that currently don’t have a tree will create and attach a tree to their DNA test.
While trying to figure out why I can’t find 4th cousin matches for descendants of Nelson G. Crawford, I ran across my first obviously incorrect ThruLines screen. One of my known third cousin once removed showed up on the ThruLines for my 4th great-grandfather, James Crawford thru the Walter Beggs and the Monroe line.
Instead of matching thru Walter Beggs and the Monroe line, our common ancestor is thru Ethel Anita Lighter and the Lida A Crawford line.
I found this odd since other close relatives of this match show up on the ThruLines for my 2nd Great Grandfather Washington Marion Crawford.
When I looked at my match’s tree, it was obvious why ThruLines was confused
we are a close DNA match
shared matches suggest a match on my Crawford line
my match’s tree doesn’t contain any names lining up with Crawford ancestry
Thus, the computer guessed. Unfortunately, the computer needs some more information so it can ‘guess’ correctly. Thus, I’m going to message my match and see if she would be willing to add a mother for Eugene Beggs to her tree. Then it will become a waiting game to see if the ThruLines change.
Unfortunately, it is pointing out one of my frustrations with my DNA results. I have several verified second and third cousin DNA matches on my Crawford line but we never got a DNA circle for my 3rd great grandfather, Nelson G. Crawford. When Ancestry released ThruLines, I was hoping to locate some fourth cousins thru one of Nelson’s other children.
However, the ThruLines for Nelson G Crawford only includes DNA matches going thru my 2nd great grandfather, Washington Marion Crawford. It does not include any lines going thru the siblings of Washington Marion Crawford.
When I looked at the ThruLines for my 4th great-grandfather, James Crawford, matches are shown for descendants of Nelson’s sister, Polly Crawford and for all of the matches thru my 2nd great-grandfather, Washington Marion Crawford.
Since Washington Marion Crawford had five siblings that lived to adulthood and had families, I expected to find ThruLines thru at least some of those siblings. Thus, I am wondering why ThruLines hasn’t helped identify any fourth cousin Crawford matches.
I have several theories:
My fourth cousin descendants of Nelson G. Crawford haven’t done a DNA test thru Ancestry
These 4th cousins have done a DNA test but either don’t have a tree or haven’t connected their DNA test to their tree.
The trees for my 4th cousins Crawford matches don’t go back far enough to connect to my research — in other words, I haven’t done enough descendancy research
Nelson’s middle name is different on various trees causing trees to not ‘match’
My tree is wrong
I’m hoping that I have enough evidence to disprove the theory of an incorrect tree. But, I will continue my search for documentation of Nelson G. Crawford and his family. To help other researchers, I have researched many of Nelson G. Crawford’s descendants. My information regarding Nelson can be found