A Segment-ology TIDBIT
There has been some recent discussion about how far back autosomal DNA is useful. Some indicate the “practical” limit to be around 2xGreat grandparents (the 3rd cousin (3C) level). I put the word practical in quotes because I don’t believe the 3C level was intended to be a rock-solid/absolute limit. I think it was intended as recommendation for many genealogists – perhaps for most genealogists just getting started.
I’ve been a genealogist for almost 50 years. I have long since researched most of the available paper records. In the late 1970s, I worked for the Smithsonian, and spent my lunch hours scrolling microfilms at the nearby National Archives; or a weekly drive over to the DAR library to roam through their stacks. I look at DNA as a new tool to add more evidence to my existing Tree and extend it even farther.
I use segment Triangulation to group Matches and to build a Chromosome Map, which also informs about the contribution to my DNA from each generation of my Ancestors. But I also value what’s called genealogy Triangulation. This is when at least three of us (me and two widely separated cousins) agree on the same Common Ancestor (CA). For almost all of my known Ancestors, I have genealogy Triangulation well beyond two other Matches.
To document, and learn from, these CAs, I developed a CA Spreadsheet. See my Common Ancestor Spreadsheet blogpost for a description, a sample and a table of the columns. My CA Spreadsheet includes thousands of DNA Matches and their CA with me. For each Ancestor Couple, this spreadsheet documents way more than two Matches for genealogy Triangulation. It usually has many DNA Match cousins, all in a large genealogy “Triangulation” for each Ancestor Couple.
This spreadsheet also includes the Shared cM amount for each DNA Match. So, it is now easy to sum up the number of DNA Matches and the average cM amount for each generation:
The takeaways here include:
1. I have many DNA Matches in 5C, 6C, 7C relationships. These stats are from my three grandparent’s Ancestors, almost all of them from Colonial Virginia – my maternal grandmother was a recent immigrant, and I get very few Matches on her line. [My parents were not related per GEDcom.]
2. At this distance, the cM relationships trend downward, as expected.
3. The averages are below 20cM.
4. For genealogy beyond the 4C level, I agree with the general concept to ignore the segment size. This is an analysis of genealogy agreement (Triangulation), that happens to be among DNA Matches. I am not claiming that the DNA segments, individually, “prove” each relationship. However, on average, some of the segments will be Identical By Descent, and when included in such large genealogy Triangulations, they increase the confidence that the genealogy is right.
5. Where known (or imputed), I also track the Clusters and Triangulated Groups (TGs) in this spreadsheet. There is usually only one or two Clusters indicated for each Ancestral Couple; and there are usually multiple TGs for each Ancestral Couple.
6. Disclaimer: Is my CA Spreadsheet 100% accurate? NO! Is it over 90% accurate? YES, IMO! Certainly the “story” in the table above is valid.
My MAIN OBSERVATION is that atDNA “works” beyond 4C.
How far back you want to take your genealogy is a very personal decision. You get to set your objectives. This post is to let you know the ability of atDNA to help you with Ancestors back at least to the 7C level.
[22BN] Segment-ology: Distant Common Ancestor Couples TIDBIT by Jim Bartlett 20230307
segment-ology 
“Statistically 2% illegitimate.” That’s a helpful benchmark, but needs context to be useful. People who lived in a small community in a long settled area where everybody knew each other and a church presence was strong usually had a low illegitimacy rate. There is much evidence that this rose when 1) soldiers were posted in town (Napoleonic Britain), or later 2) when the young moved away to mill towns for work. I suspect itinerant labourers helping with the harvest may also have sometimes contributed, but all of the examples I know of married my ancestor and settled down in her/his region.
Later in the late 1900s when most of the land had been taken and industrial wages were declining so that many men could not support a family and so could not marry, I have also found increasing rates of illegitimacy and one of the few provable illegitimacies in a collateral line. The big city is the other place. Sometimes cases get into the newspaper, but while something may have occurred, the exact offender is not necessarily the one fingers point at.
I also find that something about illegitimacy tends to follow in families.
It’s not inevitable, but I often find another illegitimacy in an adjacent generation.
(I think this is cultural, because some siblings have stable families. But who knows?)
LikeLike
Sorry, above should read “late 1800s”
LikeLike
Jim,
This is a bit of a summary and echo of the preceding.
I currently work mostly at the 5C-8C level. A little closer if helping newbies to get a start. My immigrant ancestors arrived in the mid 1800s serving as CAs to matches of my generation at 3C-4C level. The new country was mostly good and they had big families here with lots of matches for me. Some few had problems of genetic compatibility with spouses or lived in areas prone to disease carrying off the young and their lines are very thin, often with only one descendant line. And that often tends to follow into the next generation or two, so some lines can be quite thin.
And I seldom have cousin marriages, but I do have matches’ trees containing two, three and sometimes four of my own lines – mostly on my father’s side, but sometimes with one of my mother’s thrown in for good measure.
But my delight and current focus is on the generations back in the old countries.
In England that’s no problem, generally, so back to 8C is fine. Elsewhere there can be problems finding the records, but the DNA link is shown by heaps of matches and there is someone in colonial records who comes from the same tiny village and could easily be a sibling or close cousin. I just put in my tree a putative link of “unk” BROCK or whatever, to get them into my tree. Sometimes a link has later emerged.
One parish actually had two more generations of records – back to 7C/8C level.
Sure the percentage of matches drops as you go back, But the expected number of cousin matches rises as far back as calculated (8C) – https://isogg.org/wiki/Cousin_statistics
LikeLike
Christopher, There are lots of estimates of NPE (and illigetimate is only one part of this discussion) all the percentages are in the single digits, ususally under 5%. The point is that there is some probability (and it’s per generation – so about 7 generations back with 126 Ancestors, maybe 5 or 6 are not from the usual parents. There are many variants, and of course there are many different situations. I’ve found several, and suspect there may be more. Just be on the lookout – the atDNA may provide some indications.
I, too, am focused mainly on the 6C to 8C level, and I think our atDNA Matches can really help us. At that level the records sometimes get pretty sparce, and/or a lot of people have copied the wrong relationships from each other. When the records run out, some just grab a name that fits – even if it’s some born and raise in Virginia to two parents who were born, married and lived their lives in New England. Drives me crazy. But massing DNA Matches in Clusters who all agree on the same line is powerful. Jim
LikeLike
Yes! This is the argument I was making in my commentary. Jim, I hope it’s okay for me to provide a link to it here. If this violates your policies, forgive me. If it’s okay, I would love to get your response. https://onlinelibrary.wiley.com/share/author/IQV5NGHAPSIDWIRAUNQS?target=10.1002/ajpa.24705
LikeLike
LaKisha – I ok’d your link, with out reading it – I accept any honest feedback, – argee with me or not… It helps when we all listen and evaluate. Jim
LikeLike
LaKisha – I’ve now read your commentary, and agree with your premise!. I think DNA Tests from the 4 major Direct To Consumer DNA testing companies have results that can be used as you describe. African Americans have tough time finding their roots – largely due to the lack of good records before 1870. In general, I don’t advocate or use Shared DNA Segments below about 7cM, but in the specific case you bring out of the TransAtlantic Slave Trade, the Atlantic provides a significant wall between lineages in the US and lineages in Africa. This separation provides some advantage in that there are no intermediate relationships. It seems to me that, if the Shared DNA Segments can be shown to be IBD, a lower cM threshold (below 7cM) can be used. In general, I have used 7cM as a lower threshold for segment Triangulating (and many Shared DNA Segments in the 7-12cM range fail segment Triangulation and are culled out as False (IBS)). However, many of these small segments remain. My contention is that segment Triangulation results in DNA segments (for me, or the base person) which are equivalent to phased data. That is all the SNPs are on one chromosome. We don’t know the ACGT values, but we don’t need to know them – the segments are IBD! As pointed out by many, so far, there has been no scientific study or proof of this (just my logic). But, because of your linked commentary, I am now interesting in segment Triangulion down to 5cM and how that data would help your concept. Thanks again for posting. Is it OK to post your link to a facebook page that is linked to my Advanced DNA SIG (hosted by the Washington, DC FamilySearch Center)? Jim
LikeLike
Feel free to share the link. My following short paper will use simulated data, triangulation groups, and previously published equations estimating time to the most recent common ancestor. It will be a demonstration of the commentary using simulated data, which I need to do to get academic support for continuing to do the same with real data. Jim, thank you for all the work you have been putting out there.
LikeLiked by 1 person
Thanks for you input and feedback, LaKisha – I hope you’ll let us know when the paper is done. Jim
LikeLike
In reply to Douglas.
Douglas, I have one Scottish 3GGF who I have not yet confirmed genetically with a sufficient number of DNA matches – just a couple of very distant matches. I decided to trace the descendants of HIS 3GGF in an attempt to identify living people who I could ask to test.
Well, I was shocked to realise that in every generation, right down to my 3GGF, almost the only surviving descendant who had children was my own ancestor. I calculated that there were only 16 known living descendants of this line, apart from my own first and second cousins who were not helpful to this exercise. These 16 people, at best, are my 5c and 6c. Only one had an unusual enough name that enabled me to make contact, and she kindly agreed to do a test, but sadly she did not match me or my cousins – not unexpected for a 6c.
So it is a valid thing that sometimes there are simply not a sufficient number of living descendants to make the connections we are looking for. I live in hope that one day, a previously unknown branch will appear in my matches, and I will find them among my most distant matches. Obviously I could not trace every single person as some people simply disappeared from the Scottish records, and they may have emigrated if they did not die unrecorded.
LikeLike
Yeah I think this is the problem – some branches are very weak. But with potential NPEs mixed in, it gets very complicated!
LikeLiked by 1 person
Jim, I agree with you, and thanks for writing this. What you say reflects my own experience. I have been able to add several generations to my pedigree by analysing my DNA matches, despite having no paper records to support my hypotheses. However, given the number of DNA matches for me and my relevant second cousins, I strongly believe my hypotheses are valid. If I had limited my DNA analysis to the 3c level, I would have achieved nothing of value in my genetic genealogy research. All my discoveries and Eureka moments have been much more distant than that.
LikeLike
Thanks for your feedback. Before DNA, I had a fairly robust Tree of Ancestors back to Colonial Virginia. My discoveries have been mostly through DNA and all are on distant Ancestors.
LikeLike
Jim, Thanks for the info. You are much more organized than I am but I have had apparent success with finding common French Canadian ancestors for my wife many generations back using atDNA matches. There was a lot of close cousin marriages.
LikeLike
Jay, The cousin marriages do complicate it – I know of 14 cases where my Ancestors married each other. I think segment analaysis helps in those cases. Jim
LikeLike
Thanks, Jim. This is helpful. I’ve only gone through my AncestryDNA matches down to the 30 cM level but will keep working through them.
Lou
LikeLike
Lou, I have to smile… one because you are taking on this project, and two because I’ve gotten down to 20cM (it took a few months), andI’m now working down my maternal side and I’m at 15cM – with a loooong way to go. Jim
LikeLike
Jim, on my way down to 30 cM, I discovered some matches on our Metzger (Butcher) line that are surprisingly large considering the distance. Many share a single segment around 30 to 35 cM with me. They’re at Ancestry so I don’t know know if it’s the same segment.
Lou
LikeLiked by 1 person
Lou, If a person tests at Ancestry and also at one of the other companies, he/she will (virtually always) share the same DNA segment(s) with you. In both cases the Match’s data is being compared to your data – so the same segments should result.
LikeLike
Lou – the further back we go, the more likely it is that we are sharing the same segment. Being in the same Shared Match Cluster would add to that evidence.
LikeLike
Thanks Jim for the statistical summary of your sizeable genealogical project. I have been working on the assumption that the 4C to 8C range can be evidenced through genetic genealogy. However, anything past 6C becomes very difficult due to a number of factors. All of my “paper” ancestors for my mother were living on the continent at the start of the 19th century. However, I notice that select branches of mom’s tree are somewhat obscure wrt genetic matching. I find it very difficult to determine whether that is naturally occurring (ie certain ancestors simply have very few living descendants) or whether those dark areas are due to NPEs. Would like to hear your experience on that matter…
LikeLike
Douglas, I think there are several factors:
1. NPE – if a 2% NPE average is accepted, that mean one NPE within the past 6 generations; another in the 7th generation, 2 in the 8th generation, etc.
2. Family size – In America there were many farms and large families, As I recall, 90% of the men were farmers before the Industrial Rev. Many children begat many more and resulted in many more DNA test takers;>j I don’t know how it was on the continent…
3. Science – at about 7 generation back, some of our Ancestors did not contribute to our DNA – perhaps their DNA got to our mother, but she passed her “other side” DNA to us and we didn’t get any from that Ancestor. Going back, more and more Ancestors drop out from being a DNA-Ancestor. At 1,000 years ago, there was a DNA Ancestor for each piece of our DNA, but a lot whose DNA never reached us.
In my case Ancestors born in 1800 were 3XG grandparents (I’m old) – 4C range – so I expect to find all 32 of them to have contributed DNA to me…. – as well as all of their parents and grandparents.
I like grouping – Triangulated Groups (DNA) or Clusters (Shared Matches) – each one comes from some Ancestor. So look at the Groups you don’t yet have an Ancestor for . Jim
LikeLike
Douglas, I have one Scottish 3GGF who I have not yet confirmed genetically with a sufficient number of DNA matches – just a couple of very distant matches. I decided to trace the descendants of HIS 3GGF in an attempt to identify living people who I could ask to test.
Well, I was shocked to realise that in every generation, right down to my 3GGF, almost the only surviving descendant who had children was my own ancestor. I calculated that there were only 16 known living descendants of this line, apart from my own first and second cousins who were not helpful to this exercise. These 16 people, at best, are my 5c and 6c. Only one had an unusual enough name that enabled me to make contact, and she kindly agreed to do a test, but sadly she did not match me or my cousins – not unexpected for a 6c.
So it is a valid thing that sometimes there are simply not a sufficient number of living descendants to make the connections we are looking for. I live in hope that one day, a previously unknown branch will appear in my matches, and I will find them among my most distant matches. Obviously I could not trace every single person as some people simply disappeared from the Scottish records, and they may have emigrated if they did not die unrecorded.
LikeLike