This Part 3 will look at conclusions (what can we learn from all of this), and propose new spreadsheets to track all of your TGs or track TGs/Clusters through other children of our Ancestors (what we can do!).
Before I review the “Rules”, I want to focus set the stage on the Big Picture: Our autosomal DNA consists of 22 Chromosomes from each of our parents. And each of these Chromosomes has a mosaic of segments from our Ancestors. One analogy is an archeological dig with layers of artifacts – the deeper we go, the more distant (further back in time) the artifacts. Our DNA has a similar pattern. Each of our Chromosomes is composed of segments from our grandparents, passed to us from a parent. Going back another generation, each of our Chromosomes is composed of segments from our Great grandparents; and so on. Even if we went back 100 generations, we’d find that our Chromosomes were made up (completely) with DNA from that generation. This is not to say that every Ancestor in a generation contributed to our DNA; it is to say that some of the Ancestors in a generation contributed all of our DNA. And all of the crossover points are there if we can dig deep enough. But our DNA does not include “signposts” or markers that identify crossover points. Segment Triangulation is the only method I know to determine these crossover points and define specific segments from Ancestors (beyond Visual Phasing of grandparent crossovers). One drawback of Triangulated Groups (TGs), is that they are formed from available Shared Segments with Matches, and they are not formed at any particular generational level. However, we do know that each segment of our DNA came from an Ancestor – each segment represented by a TG came from an Ancestor. When we find a number of (widely separated) Matches in a TG agree on the MRCA, this is powerful evidence that the TG-segment came from that Ancestor. We can then map these TGs and MRCAs. In this respect, we learn that segment Triangulation (TGs) is essential to confirming our biological ancestry. We learn that the DNA is not scattered willy-nilly over our DNA. There are genetic guidelines – which I call “Rules” in this blogpost series. Our Chromosome Map should be in general agreement with these rules. And, perhaps in Part 4, I’ll try to outline how we can use these rules to predict much more.
So what do we learn/do with all this musing?
1. Learn: We can understand how TG segments originate in Ancestors – maybe 5 to 8 generations back – and pass them down to us, and our Matches. At each generation – coming down/getting closer to us – our Ancestors have more TGs – until our parents each pass down about 150-200 TGs to us.
2. Do: Use spreadsheets to track and analyze this growing amount of data… Examine closely areas that deviate from the rules.
Here is a summary of our “Rules”.
-Rule #1: We can expect roughly the calculated numbers of TGs from each generation – an order of magnitude [See Table in Part 1]
-Rule #2: We can expect about 34 crossovers to occur per generation on each side.
-Rule #3: Shared DNA (with Matches) reduces by roughly 1/4 with each generation.
-Rule #4: We should not see Matches beyond 3C, with the same TG, descending from more than two children of a CA.
-Rule #5: The amount of DNA, and number of segments, in each generation, are not affected by external factors. [The number of TGs may be affected by the Matches you have.]
-Rule #6: The sum of the DNA contributions of all Ancestors at each generation will be 100%. This is a hard rule that is true at every generation.
-Rule #7: Each of our Ancestors will have all of the TG-segments both their parents had.
-Rule #8: A TG that subdivides going back, separates into a two smaller segments – one from each parent.
–Observation: Based on my 372 TGs, I estimate, using an 8cM threshold, most of you should get 150 to 200 TGs per side. They will range from small to large and span almost all of your DNA (there may be a few gaps – it all depends on the “coverage” provided by the shared DNA segments with your Matches.)
I think we should track and analyze our TGs. My preferred method is with spreadsheets:
1. Master atDNA Spreadsheet – this is an “everything but the kitchen sink” spreadsheet for me. I include every IBD segment (over 7cM) of every Match – with segment, genealogy, and other information. I still plan, someday, blogposts about spreadsheets. This is a cursory overview. A spreadsheet is highly personal – I recommend you start with downloads from your testing company and add columns to suit yourself – mine is continually evolving.
Here are some of the columns in my spreadsheet:
a. Match info: full name, last name, company “name”, email, POC, company, Notes
b. Segment info: Chr, Start, End, cM, SNPs [from the companies} and TG ID (from me)
c. Genealogy info: Tree URL hyperlink; cousinship (e.g. 3C1R); MRCA couple surnames, side
d. Other info: GEDmatch ID, dates of communication, etc., etc.
e. Most of this info is from company downloads, the rest is typed in as I get it.
My spreadsheet has over 20,000 rows… including the following “Header” rows
f. 22 paternal and 23 maternal Chromosomes – Header/dividers
g. TG summary Header/dividers – uses earliest start location of TG; MRCA (per my judgment)
h. Alternate TG Header – to records alternate MRCAs for some TGs
Spreadsheet sorts – each of these sorts is a valuable tool for me
i. Alphabetical by name
j. Chr & Start – used to Triangulate segments (as they are added)
k. Side & Chr & Start – clearly groups all Matches in a TG – should have same MRCA
l. TG summary Headers, only, by Side & Chr & Start – to analyze MRCAs by generation [these headers have Ahnentafel numbers (in columns for each generation) to the MRCA].
m. This TG summary sort also allows an analysis with respect to Rules #1, #2 and #3 by generation.
2. Common Ancestors Spreadsheet – this is a list of all Matches with Common Ancestors with me (now about 6,000 rows)
a. Match name, shared cM, cousinship, Tree URL; name/birth of MRCA descendants; Ahnentafel
b Any known TG ID or Cluster
c. A header row for each of my Ancestors (usually a couple)
This started with columns for Children of Ancestors; their birth year; and same for grandchildren – this info was typed in for each Match’s line of descent from the MRCA. I have since expanded it to include columns for remaining line of descent (and birth, down to the Match. For females I add married name; for example: Nancy m FLEMING to indicate succeeding descendants have a different surname (saves a little time and space).
Most of this info was from AncestryDNA ThruLines, but I have a number of rows for Matches from the other companies (and the TG info for them, which is like gold).
When this spreadsheet is sorted by MRCA Ahnentafel & Child birth, it looks like a series of Family Group Sheets for each MRCA family.
d. This makes it very easy to check against my Family Groups Sheets researched and developed over the past 45 years. I highlight conflicts in a mud color for further research and analysis.
e. This spreadsheet also has a column for for Potential Ancestors [POT ANC] (usually from Ancestry or MyHeritage]
f. For each Family Group, it’s also easy to check the range and average of Shared cMs for that family
g. Maybe most importantly for me, it allows me to see if there is a TG and/or Cluster thread in each family
h. I can also check for violations of Rule #4. This has already highlighted a few such violations – almost all of which led to an alternate MRCA with a much better “fit”.
i. All instances of a Match with multiple TGs and/or multiple MRCAs, must be adjudicated. A TG can only link with one ancestral line. This spreadsheet is a good tool for that analysis.
With this Common Ancestors spreadsheet I’m finding three things: a) each Ancestor does tend to have a consensus of Clusters and/or TGs; and there is the occasional outlier (this illustrates that just because you have shared DNA and an MRCA with a Match, it doesn’t necessarily mean the shared DNA came from *that* MRCA – there could be other MRCAs); b) there tends to be only a few groups (Clusters and/or TGs) in each Family Group – roughly in line with the table at the beginning of this blog post; and an occasional outlier of more than 2 children with the same TG (which indicates to me there is probably an issue – probably, also, where the shared DNA didn’t come from *that* MRCA).
3. TG Summary Quick Sheet – taken from my Master atDNA Spreadsheet, this just has a few columns, and I can fit all 372 TGs into 2 pages (maternal and paternal).
a. Columns: Chr; Start; End; TG ID; side; 8 columns for 8 generations of Ahnentafel numbers; MRCA Ahnentafel; MRCA cousinship; MRCA surnames
This Quick Sheet (1-page; front/back) is a handy reference – linking TGs to MRCAs and finding TG IDs for any segment. I include the Chr Headers (solid black) which highlight the TGs in each Chromosome.
I cannot over emphasize that spreadsheets are a tool, and you should adopt any spreadsheet to your own objectives and methodology. Don’t be afraid to add or delete columns or header rows.
Final thoughts for this Part 3 – think through the guidelines (“Rules”); and set up spreadsheets to help track and analyze your data. However, building and maintaining a spreadsheet to manage your Matches and Segments or Common Ancestors takes time – it’s not for everyone. Or… use whatever system you want – genetic genealogy is your hobby, and you are free to enjoy it however you want.
I might have a Part 4 with some more thoughts …
[15J] Segment-ology: Distribution of TGs – Part 3 by Jim Bartlett 20211008
Another possible method beyond TG is ancestor DNA reconstruction. It is naturally not absolute, but with a big data approach, you can learn much more than just studying things from a perspective of one individual or one family.
Kuba, I like that idea, but remember that ourselves and our Ancestors only pass part of our DNA down to our children – the rest is lost. And the best thing we can do toward ancestor DNA reconstruction, is to form all of our own Triangulated Groups, and link them to the correct Ancestors (Chromosome Mapping). We would need many descendants to do Chromosome Mapping, in order to start an ancestor DNA reconstruction. Jim
I agree, that we will not have 100% reconstructions. I tried to hint at that in the original comment.
Yet using this method/concept, we can (theoretically) leverage all tested people, thus learn about TGs and recombination points that are completely unattainable using only our own DNA.
Here, naturally, the larger families have a big plus. 🙂
OTOH, DNA from an ancestor than no-one inherited does not have to be reconstructed, as it will not help us with current research.
I know it is not a completely practical comment (at large scale, we can borland tools or HAPI to reconstruct a single generation or two) until the DNA companies start to build tools like this (as you need the raw data of millions of individuals to succeed). Still, I strongly believe the concept would be able to add insights on top of just person TG research.
Thank you for your articles, Jim, they are much appreciated!
Jim, thanks so much for the thought-provoking articles. I almost always tweak my data and method after reading. I’d be most interested in seeing a sample of the Common Ancestors worksheet, maybe in a future post. I imagine this to be columns showing each descendant down to the match, meaning 5 columns for a 3C, 6 for a 3C1R or 4C, etc. Correct? Are you also tracking location? I’d be most interested in knowing how you track trees for matches whose MRCA is not yet known. Research trees in Ancestry? I find the biggest challenge is comparing surnames and locations on these trees to zero in on a common ancestry. Any details appreciated. Marc
Marc, See my other reply. I think using geography/location is a good insight. Clearly our shared DNA segment can only come from one location… Jim
Jim, thanks for yet another great article. They are thought-provoking and usually lead me to some refinements in my own data. I’d be very interested in seeing examples of the Children of Ancestor columns you mentioned above. As you described it, I envision for a 3C (an an example) 4 columns off to the left of a 2xG parent MRCA couple, 5 columns for a 3C1R or 4C, etc. Is this correct? On a related item, I’m wondering how you track trees of matches with potential ancestors. Are you doing this with research trees on Ancestry? I find the most difficult part of determining MRCAs is identifying common surnames or places within a group of matches (either TG or Cluster), especially when the MRCA is more than 3 or 4 generations out. Any hints here would be greatly appreciated.
Marc, Thanks for your kind words, feedback and questions. I’ll try to get a sample of the Common Ancestor Spreadsheet into a blog post soon. I actually have 3 pairs (name & birth year) of columns for the first 3 descendants of the MRCA, and then one column for whatever is left (example of the left over descentants: James 1878 > Mary 1903 m DAVIS > Frank 1930 > * (where * is the Match). This way I can sort the first 3 generations down (they are like Family Group Sheets). The whole spreadsheet has these 7 columns… Since the 1980s I’ve been recording my generalogy in a Text document (now in Word) – one doc for each surname, in outline format – each generation looks like a Family Group Sheet. And, yes, in order to use the Ancestry tools, I sometimes create Private, non-searchable, Trees for the Potential Ancestors – I haven’t found any of them that I agree with yet…
I haven’t developed any firm hints yet to determine the correct distant MRCA, except for genealogy Triangulation – In the cases of my HIGGINBOTHAMs and CUMMINGs line the Cluster/TG is filled with Matches who all agree with the MRCA. And in other cases I use Walk The Ancestor Back – again, the large majority of Matches with MRCAs are on the same line of my Ancestry. Jim
Thanks, Jim! Very helpful. I’m looking forward to your even more detailed and illustrated spreadsheet post.
Fascinating! You are a fanatic. I am hoping to find a greatly simplified way to start this. I was wondering what overall results you have obtained from this? The “rules” learnings are quite helpful.
James, The simple way to start a spreadsheet is to download from FTDNA, MyHeritage, or 23andMe. Add columns for other info as you find it – like an MRCA column. When you Triangulate, you can add a TG ID [a previous post outlined how to do that].
Overall I have mapped all my segments at least to a parent; and some back to the 7xG grandparent level. I’ve found two brick-wall Ancestors so far; and now I’m on a roll to find out more about a few of my Ancestors I’ve neglected for a long time. I’ve pretty much confirmed almost all my known Ancestors to the 5xG grandparent level, but haven’t yet linked them to TGs… Jim
Thank you Jim!
Julie, You are welcome. I’m having fun sorting this all out. Glad that I can share it too. Jim
Another great article!
LikeLiked by 1 person