In Elmore Dark Matter I talked about the regular old paper trail steps I followed to uncover Maurice Elmore’s history and his connection to the family of Halsey Orton Elmore. In that search, the key to the entire thing was actually his mother’s family. Deciphering their connections led me down the right path and nicely all that detective work was later confirmed with documentation.
Having discovered Imogene Carr, her father Daniel Carr and mother Mary Smith, I’m going to jump back to DNA for a bit and talk about the segment triangulation I was attempting a year ago with the Carr family.
Segment triangulation?! what the heck is that? Here is my ham handed explanation…of course I use something near and dear to me, food, to explain myself. Here is a nice concise and more professional definition from Blaine Bettinger at the genetic genealogist.
Before I talk about the Carrs, I need to talk about the Boltons.
You’ve seen from my other posts that finding genetic matches has been easier than getting family tree information from those matches. Once you have a collection of people who match on a segment and the (usually much smaller) group of people who will share family tree information, then you can start looking for common ancestors.
The Boltons were my first triangulated DNA ancestors on my grandfather’s side and they are special because I don’t have a particular bias towards them. They do not appear in my family tree. I noticed as I was accumulating family trees for various matches that I started to see a lot of Boltons and that they seemed to be related to the same Bolton family.
This table shows various genetic matches for my dad with people who have this Bolton family in their tree. All of them proved to be paternal matches (so related to my grandfather). Although most of these matches are in triangulated groups, I can only say that one of them is a Bolton triangulated group.
The match on chromosome 1 is related to Thomas Bolton and Jemima Hammack through their daughter Nancy. Thomas Bolton is the son of Robert Bolton and Mary Hubbard. His match is very small (prone to be identical by accident) and although there are four other people on that same segment, I haven’t been able to get any family trees for them. The match on Chromosome 4 is really rough because it was only a surname listing for Bolton. No tree was available…so that doesn’t count for much.
The matches on chromosome 3 all match each other and are related to Thomas Bolton and Elizabeth James. Thomas is the son of Thomas Bolton and Jemima Hammack. Unfortunately, they are all very closely related. My dad matches everyone from a young man to his aunts on to his great aunt in that same group and his great aunt is the daughter of a Bolton who is related to Thomas and Elizabeth, but they don’t count because they are too close to each other. They form more of a straight line than a triangle. I do not have trees for the three other kits in this triangulated group.
| Chr | Start | End | Length cM | SNPs |
|---|---|---|---|---|
| 1 | 199928328 | 204786830 | 8.1 | 1475 |
| 3 | 11000000 | 32000000 | 28.6 cM | 4862 |
| 3 | 11000000 | 30000000 | 25.4 | 4369 |
| 3 | 11000000 | 30000000 | 25.4 cM | 4360 |
| 3 | 11000000 | 21000000 | 11.1 cM | 2001 |
| 4 | 10000000 | 23000000 | 14.3 | 2585 |
| 5 | 4686409 | 32463775 | 37.34 | 6927 |
| 5 | 40322578 | 72536852 | 18.8 | |
| 5 | 40381478 | 65221177 | 14.28 | 4200 |
| 5 | 52272205 | 73523708 | 17.4 | |
| 6 | 8478783 | 23509261 | 23.16 | 4496 |
Chromosome 5 represents the actual triangulated group. None of them are very closely related, although they do meet up at various levels of the same Bolton family. Two share only single segments with my dad. They are related to Rebecca Bolton the daughter of Robert Bolton and Mary Hubbard. One match has two segments on 5. Each of his segments is in it’s own triangulated group, one matching with the others here on chromosome 5 (in the group of three) and one (the larger match) that is in a monster triangulated group of people who appear to generally be from Kentucky and share multiple families there. The two segment person on chromosome 5 is related to Thomas Bolton and Jemima Hammack.
So there is my triangle on chromosome 5, the end result from multiple years of work and contacting matches at three different companies, and they don’t even exist in my tree. All I know is they are somehow related to my grandfather and sometimes appear adjacent to matches to his mother Orvetta Finks.
The last man standing there on chromosome 6 is related to Thomas Bolton and Elizabeth James. He’s all alone there on the paternal side so there is no one to triangulate with.
You can see me working out these connections in 2014 and at the time I’m thinking the Boltons are related to the Williamson family (when I still held out hope that we were at least related to the Williamsons in some way if not the Thompsons).
I reference the Boltons here as a bright spot when I write about being tired of spinning my wheels in autosomal DNA in 2015.
The really scary/interesting bit happened just this year in 2016, when my paternal first cousin had his DNA tested at ancestryDNA. Ancestry picked out the Long and Bolton family as possible DNA only matches for him (but not me). This is the same family connected to that small Bolton match on chromosome 1 for my father. My level of confidence in a Bolton ancestor is 99.99% and it seems ancestryDNA might have picked up on it too.
Before I talk about the Carrs, I need to talk about the Proto-Carrs.
I started working on the Carrs in January of 2015 only I didn’t know I was working on the Carrs. I very rarely know what I’m working on when I start forming groups of related DNA segments. I’ve gotten enough relatives tested that I might know what side of the family I’m working on which is nice, but beyond that I have been in the dark for years on what the family connection would end up being.
The story of the Carrs starts with me contacting matches about what looked like the start of a triangulation group. Here are the players in this story.
| match | chr | start | end | length cM | SNPs |
|---|---|---|---|---|---|
| Evelyn | 1 | 166344260 | 182502103 | 15.04 | 4097 |
| David | 8 | 12770091 | 25799345 | 22.22 | 5269 |
| Bruce | 10 | 36728420 | 77118441 | 34.38 | 8945 |
| Evelyn | 10 | 36728420 | 77118441 | 34.38 | 8945 |
Evelyn and Bruce match my dad on chromosome 10 and they match each other there as well. This is a paternal match because I am lucky enough to have maternal matches for my dad in the same area and these guys don’t match them.
Evelyn also has a segment on chromosome 1. She is all alone there as the only paternal match I’ve confirmed in that area.
As I would come to find out, David on chromosome 8, is the third cousin of Bruce. They share the Carr family in common, but I do not have a tree for David except for his path to the Carr family and a list of surnames he knows. Bruce also shared that David did not match him on chromosome 8. The Carrs are in Boone County Indiana in the mid 1800s before moving on the Howard County Indiana and then Story County Iowa. David and Bruce meet up at William Carr the son of Nathan Carr and Sarah Wiley.
Of course at the time, Indiana was bitter sweet for me because I know something is up (although not exactly what) with my Thompsons who are also from Indiana.
Evelyn and Bruce on chromosome 10 were my targets for the start of a triangulated group, just like the Boltons, there appeared to be a pattern of related people. As it turns out Evelyn also has a Carr from Boone County Indiana in her family tree. Her 3rd great grandmother is Rachel Carr, the poorly documented wife of Martin Brouhard who dies there in Boone and whose children are scattered to other households by 1850. One researcher suggests that she is the sister of Margaret Carr Smith which would make her a daughter of Nathan Carr and Sarah Wiley. Rachel Carr is a brick wall ancestor and I cannot pin her to the rest of the Carr family.
Evelyn’s larger match would make you think that she is somehow more closely related to my dad, but I cannot find it in her tree.
This is a failure to triangulate a common ancestor. I’ve got two people in the triangle, no match in my tree and no real connection between them other than their DNA and the whisper of possible common Carr ancestry. The triangle when paired with David’s information is compelling, but David is in his own triangulated group. Other paternal matches on the same segment are half the size of his (in the 10cM range where his is 22cM) but they all seem to share the same Crawford and Harrell family and are not closely related.
When I first contacted David it was about the Crawfords and Harrells. He has none of those in his known family. Does David’s match represent the Carr family or an earlier tie to one of the ancestors in the Crawford and Harrell families shared by the three other matches on his segment?
In any case, at the time that I put these people together, there were no Carrs in my tree and also at the time I thought it might be likely that we were related to a “Ward” family as the surname Ward was common to the two people on chromosome 10. It remained an interesting group, but I couldn’t get any further.
Now, I can talk about the Carrs.
When I imagined the glorious and amazing process I would follow to find my great grandfather and solve the recently, close to home, mystery of my Thompson family; I thought one of the sign posts pointing the way would be that I might run head long into the Boltons. With so many possible connections and one triangle I felt good about, I was sure they would reveal themselves somehow.
As you know from Elmore Dark Matter, I did not run into the Boltons. I ran into the Carrs. Now it turns out we are related to Daniel Brimmerman Carr who is the son of Annanias Nile Carr and Jane Franklin. Annanias is the son of Nathan Carr and Sarah Wiley. So I see them again, only this time they are in my family tree.
I still don’t have a Carr triangle. I’d feel really accomplished if I could tie up Rachel Carr to Nathan Carr and Sarah Wiley, but that may just be out of reach. As it sits, I’m still waiting for a third leg in that Carr stool. It is possible I could wait a long, long time for this to work itself out.
I failed to work it out a year ago, but I’m not that worried about it. Don’t get me wrong. I believe that segment triangulation is the best way to nail down a distant ancestor and I intend to keep at it, but I also feel like I’ve reached some level of confidence with this one. I’m 85% confident that I am seeing a real genetic match to our Carr ancestors.
Why 85% confident in the Carrs?
Okay. The hope and dream is to stack up those pancake segments, sort the strawberry pancakes from the blueberries and then hunt down those common ancestors. If you have both grandparents to test, then it’s easy to figure out what they gave your parent as DNA but if you don’t then you’re down to sorting segments that match common relatives who are available to test.
What you’ll notice if you test closer relatives like I’ve been able to do with my paternal grandmother’s family, is that not everyone matches everyone else everywhere. At the same time, segments drop off the map and disappear over generations. That means the more generations there are between you, a relative and your shared ancestor, the less likely you are to triangulate on a single segment.
As an example I’m going to turn to our cousin Phillip again. Second cousin one time removed for my dad and his sister. He is a close enough relative that I don’t really question his place in our family tree. For this comparison I’ll use my dad, my aunt and myself and we’ll look at things from Phillip’s perspective.
Notice the top table. My aunt Cheryl has more segments in common with their second cousin one time removed Phillip than my father Thomas does. One of those segments (on chromosome 22) is about 20cM, so it’s a non-trivial segment difference between brother and sister. You can see that even between siblings things are not always equal.
I’m the purple one there, notice in the chromosome matching chart that I do not match Phillip on chromosome 4 (circled in red) even though it’s a nice big segment shared by my aunt and father.
This is where I would note that any or all of my siblings might carry that segment on chromosome 4 that I do not have. Any or all of Cheryl’s children might carry that segment on chromosome 4 too and they would in turn pass it on but I will not. My children have no hope of matching there on chromosome 4.
I have lost a significant segment match that was there just one generation before.
Here is another example. Again using Phillip, but instead of Cheryl I’ll substitute my dad’s uncle Ray (also related to Phillip as his second cousin) and myself.
Notice from the table at the top that Ray has more segments than my dad Thomas does but less centimorgans in common with his second cousin. These are differences between Ray and his sister who handed my dad his matching segments. I’ve circled these segments that Ray shares with Phillip, but Thomas does not. Ray may have handed those segments to any or all of his children and they could have passed them on. My dad does not have those to pass on to his children.
On chromosome 5, Ray appears to have a segment that is just about to wink out of existence while my dad handed a larger one right over to me. Ray’s children and grandchildren may not see that segment…or it may be too small to count.
Again, I’m highlighting chromosome 4 where great uncle Ray shares the same segment with Phillip that my dad does. Ray’s children and grandchildren could carry that segment. My aunt carries it and her children could too. My father carries it and my siblings might, but I do not have that very big segment to pass on. I am just out of the loop on chromosome 4.
This represents one second cousin level match (at different distances of removal) compared to three generations in my family. In those three generations there are differences between siblings and differences between generations. Many segments are common but many are not and some are just lost.
Thinking about this snapshot of chromosomes; see how different Ray is from my dad. He is my dad’s uncle, but he has four segments in that little snapshot that my dad does not. Considering that Ray and I only match on two segments in this snapshot (Ray and I only match Phillip on three segments total), the chances that Ray’s grandson and I would match Phillip on any one shared segment seem more remote. It’s possible, but it becomes less likely.
Ray’s grandson and I would both be Phillip’s second cousins two times removed, neither would be less related and yet we might share totally different segments with him than we do with each other.
As a point of interest, I ran over to Gedmatch.com to check out how my first cousin Tim matches with Phillip. Tim and I share our two paternal grandparents, his dad and my dad are brothers. Tim, my dad and I only overlap each other and match Phillip on one 20cM segment. That’s my first cousin compared to me and our shared 2nd cousin 2x removed and we have one 20cM triangulated segment with a known relative even though we each share multiple segments with that same relative…we just don’t share them with each other.
The genetic lottery makes triangulating a segment with a known relative more difficult than you would think, especially as that relative becomes more distant.
Back to the Carrs then with these things in mind.
The fact that my dad has at a minimum a 30cM match and a separate 22cM match with two members of the Carr family related to his 4th great grandparents (their second great grandparents) is pretty neat. Now add in a possible triangulated segment with another Carr from the same town at the same time as the first two Carrs..and my confidence in seeing a real pattern of ancestry goes up.
I also feel like there is a boost because when I put this group together last year, there was no bias toward the Carr family. They didn’t exist in my tree. I had no expectation of being related to them that might drive me to fabricate evidence. Finding out in the last few months that I am related to them is frosting on the cake I already baked.
Judgement Calls
To me, segment triangulation is a defense against “genetic astrology”. Genetic astrology is, among other things, the parlor trick of DNA matching through “In Common With” lists. For example, you find 10 genetic relatives who are in common with each other (through FTDNA or AncestryDNA) and find that they are all related to George Washington or Robin Hood or Odin and then use that as evidence to back your claim in your own family tree, that you are related to Odin as well.
It’s all about running off half cocked, seeing what you want to see and claiming evidence without doing any of the work. AncestryDNA makes it easiest to do that by providing no way to look at segment data but making “common matches” readily available. It makes it easy to make a claim without real evidence to back it up. It can be dangerous, especially the farther back you go. Here in the U.S. a lot of people are related to early colonial families. How do you know if your DNA Circle for the Dyers was really a match to the Dyers or if some of those people are related to you through the Comstock and Savage family you also share?
By having some standard, like a triangulated segment, you hope to avoid the pitfalls of wishful thinking. WE ALL struggle with wishful thinking. You’re working to eliminate other ancestral possibilities when you form a triangulated group. You want the group to help you rule in or rule out ancestors.
There are always going to be uncertainties. Triangulated segments are not foolproof. I have lots of triangulation groups that don’t point to any common ancestor I can find. I imagine there is one, but those segments could be questionable or their trees could have veered away from a genetic ancestor.
In the end a triangulated segment is really just an attempt to eliminate as much uncertainty as you can. That is why a lot of genetic genealogists consider them a standard of proof for a common ancestor.
Without a triangulated segment then, what makes me think the Carrs are not genetic astrology and wishful thinking?
The truest answer is that they really could be. Any pattern might end up being a false pattern. These relatives are not close enough to be a sure bet like some of my other genetic cousins. I have not, and really cannot, eliminate the possibility that they are all related to some other known or unknown relative of my grandfather. There are plenty of unknowns in the Finks family and even in other branches of the Elmore family after all.
I am trying to have my cake and eating it too in a sense. I believe that a triangulated segment is a good foundation for hunting down ancestors and offers better proof than a bunch of scattered segments around our genome. On the other hand, I can also show within my own family how triangulated segments may not be available even when there is no question of the family relationship.
You get down to having done your homework as best you can and coming up with some level of confidence in the patterns and matches you’re seeing.
For now, this evidence for the Carr family is the best evidence I have. It is not the best evidence there could be. It’s just the best I have. In the end, it’s a judgement call. I’m going to bet on these Carrs and see how things play out.
Seeing this likely Carr connection, in the light of discovering my Elmore family and their Carrs, made everything seem more solid. I’ve got people on both sides of Maurice Elmore’s family. That raises my confidence level in the entire enterprise.