Branch Groups
I’m methodically working my way through my Ancestors and Matches using Pro Tools. My main tool is my Common Ancestor Spreadsheet, which is now growing very rapidly. I’m not really in it for the bulk, but for the advantages of Branch Groups. What I call Branch Groups are groups of my DNA Matches under one child or grandchild of one of my Ancestors – these Matches are on the same Tree “branch”. Such Matches are closer to each other (than to me) and tend to share more DNA with each other. They stand out with DNA shares over 90cM; and I take notice. I can often “fit” them into a Branch Group. On the other hand, I’ve found some Matches that have the right genealogy for a Branch Group, but they don’t share much DNA with others in the Group – more on this below.
Here are some thoughts and observations:
SMOM – Shared Matches of Matches aka “Rabbit Holes” – haha. When you select a Match and click on the Shared Matches button – you get a list of all the Matches you both have in your respective Match lists. These are your Shared Matches (SMs) with that Match. Each of these SMs shares some DNA with you that you both got from the same Common Ancestor (CA). And, with Pro Tools, you know how much DNA each of these SMs also shares with the “base” Match that *they* got from some CA. Often these two CAs are the same (or one is ancestral to the other); but sometimes the CAs are completely different – *their* CA could be unrelated to you or related to you on a different line – see Outliers below). When we’ve done our homework and entered Notes for many Matches, we can usually look down the SM list and easily see if there is a consensus, or not – see Birds of a Feather below. Like with auto-Clustering, a consensus indicates a group of Matches that mostly match each other, indicating a Common Ancestor among them. Usually, their CA is also one of your Ancestors – BINGO! This is a Branch Group. Sometimes their CA is unknown to you – this could be a random happenstance. Or it could be a Floating Branch Group – see below.
Branch Group aka Cluster. When you find SMOMs who share high levels of shared DNA (cMs) with each other they usually form a Branch Group. By “high levels” I mean at least 90cM; but I often drop down to around 50cM as the group grows larger. I consider 20-25cM as “in the noise”, and usually not worth the trip down a rabbit hole. [For your own situation, experiment to find a threshold that usually gives you efficient results.] Sometimes you can get 5-10 (or more) of these SMOMs which link under a child or grandchild or Great grandchild of one of your Ancestors. And then it’s easier to find other SMOMs that fit into the Branch Group. Use an SMOM in a Branch Group to make a new Shared Match list, invariably with new SMOMs… the clues (or rabbit holes) are everywhere! As it turns out in a Branch Group, not all Match descendants will Match all of the other Matches in the group. Remember: at the 4C level, roughly 50% of true 4C won’t show up as matches to each other.
Birds of a Feather. On many Shared Match lists, a scan of the Notes indicates a clear consensus – most SMs have Notes indicating the same CA; and some are from the same line (up or down a generation). These are birds of a feather – they cluster together. And Pro Tools shows them to be close relatives – these are a Branch Group. In these cases, I’m much more likely to review Matches not yet linked in, and to build their Trees back to find the link. As a quick check, click on a Match and see *their* SMs with you – are they indeed Birds of a Feather? Or not?. For some Shared Match lists, a quick scan of existing Notes may indicate they are all over the place – on both sides; on different branches – so, it’s difficult to determine a consensus. Move on…
Outliers – linked by genealogy, but not linked by shared DNA. I’ve now run into a very few cases of DNA Matches who are clearly genealogy relatives (in my Common Ancestors spreadsheet) under Ancestor XYZ, but they do not share DNA with other close cousin Matches under XYZ. In each case, so far, they are also related to me in another way, and they do share DNA with their other cousins. Thinking about multiple segments and/or multiple relationships leads me to Triangulated segments, but I’ll put that discussion off for a future blog post. Just be aware that a Match with one shared segment can only be genetically related one way. Pro Tools may help determine which one.
Collateral SURNAMES in Branch Groups. Less than 1% of my Matches have the same SURNAME as the CA we share [Y and mt lines are pretty rare]. This means my Common Ancestor spreadsheet (tracking the lines of descent down to Matches) includes Collateral SURNAMEs. As I’m working on an MCRA Branch Group in my spreadsheet, I’m reviewing each of my Match cousins, and reviewing all of the SMOM shared cMs, and checking the Trees of those over 90cM (and glancing at some down to 50cM). Often there is enough to tie those Matches to my Tree (even some with no Trees). It really helps to review the Collateral SURNAMEs already recorded in my spreadsheet for that Branch – that’s usually where I’m going to find a link. And it means I don’t have to build a tree back for each Match – I can usually copy the line of descent of an existing Match in the spreadsheet, and just change the last few generations. A big time saver – in searching and typing… Recognizing a Collateral SURNAME in a Match’s scrawny Tree is helpful. Sometimes I’ll filter a long Shared Match list by a Collateral SURNAME…
Floating Branch Groups. A few times I’ve found a Branch Group that I cannot link to my Tree. They usually include parent/children, siblings, aunt/uncle/niece/nephew, and maybe some 1C or 2C, all in a tight family group. All the interrelationship cMs are on target. But, other than being on a Shared Match list with some known Matches in a Branch Group, I cannot find a link. In most cases this has happened “near” a Brick Wall (or “iffy”) Ancestor of mine. So I’ve created a Floating Branch in my Tree, so I can link other Matches to it. I need to do a study of closest known Matches to see where this Floating Branch is headed – another rabbit hole. Such a Floating Branch could just be a mirage (not really linked to me), or I might find some “tendril” Matches (maybe through a Collateral SURNAME filter) that help find the link. I operate under the belief that ALL Matches over 15cM (and many under 15cM) are true cousins, and many are within a genealogy timeframe and should fit in my Tree somewhere.
I am now convinced of two things: A) A lot more of our under-20cM Matches are well within our genealogy timeframe than I originally thought; and B) our Brick Walls (out to at least 8C level) have plenty of Matches forming Branch Groups. With each generation going back, it’s harder and harder to figure them out, but Pro Tools can often provide new insights. This helps offset the fact that many Matches have NO Trees or very scrawny Trees. There is hope! But it takes work!
[22CR] Segment-ology: Pro Tools Part 10 Branch Groups by Jim Bartlett 20240812
segment-ology 
Jim , good morning , can you help me with these matches , i am mostly iberian but in my isolated region had a italian colony +- 300 to 340 years ago , my chromosome 20 shows 20 italians in the same segment but i can found only 6 of them triangulated with 8,8 cm size , another ones is between 7,5 to 8 cm , these matches could be false? If not what probability of be true? My family came from isolated regions and one branch is unknown, 20 italians matches but i can triangulate with only 14 and only 6 of them have 8,8 cm of triangulated size , each one of these matches have +- 8 to 9,7 cm size, triangulated only 8,8 cm and only 6 of them
LikeLike
See other replies – we must be in different time zones… Jim
LikeLike
Thanks Jim, I chanced on this today but I have been doing a similar spreadsheet exercise for a month now. I agree with your conclusions. There are obviously a number of small segments which are legitimately assignable to Branch Groups. I have many Branch Groups with cMs 8-10 who have cM’s of 100+ on the SMOM side. I have been colour coding branch groups, which is good visually but I am running out of meaningful colour scales so I will look at your coding method for delineation of branch groups.
I am doing an intensive comparison between two groups, on of recognised branch assigned relatives and otherwise unrecognised relatives who don’t know who their true common ancestor is but do know how they are related to each other. The aim is to ‘nail down the culprit’. This has helped me no end with the assignment of these common relatives to branch’s on both the known and unknown side. All this arising from a time before sperm banks!!! I like your terminology for organising thoughts and may adopt it. Please keep up the posts.
LikeLike
Pat, Thanks for the very kind feedback – glad you found this site. The “Outline” Tab in the header bar is your friend – it’s like a table of contents with hyperlinks to the posts.
I have no 1C; so 2C to 4C are my foundation – and with Pro Tools I’ve found a number of new cousins that can tie in, and they range down to 7cM. Sometimes they have no Tree, but are a sibling, child/grandchild, even avuncular – they are easy to tie in. I don’t really care about the names of each link to my Tree, just that the spider-web of interrelationships works (building a solid sub branch). Actually this works, for me, out to 8C – I just need to work on finding the intermediate cousins that will strengthen the links back to my Ancestor.
And, like you, I’ve found some floating branches. In general they appear to be where I expected them, but I still cannot find the link (need to focus more on the geography – although I have at least one example of a Civil War soldier in my Tree, who probably fathered a child far away from his home.) Sometimes it’s a tough jigsaw puzzle to put together – we have a lot of the pieces, it’s just fitting them together the right way… Jim
LikeLike
Ciao jim ho 201 matches sul cromosoma 2 da persone finlandesi I segmenti sono tra 7,5cM fino ha 15cM e una regione pileup ?che vuol dire?non ho nessun collegamento con persone della finlandia.
LikeLike
Kevin, That means most of them are probably false, or from very, very, distant Ancestors. Jim
LikeLike
Ciao jim quel gruppo sul cabale 16 quanto faccio uno contro 2 ha volte ilsegmento e 9cM ha volte uno contro 2 e 8,7 cm ha volte e 7,5cM e a volte e 7,1cM.adesso e rimasto solo il mio test e non riesco ha capire e questa triangolazione del segmenti e tra il 4 5 ho e piu lontano e dovrebbe rssere dal lati paterno dove hi 2 tg stesso canale 16 stesso allineamento ma 2 tg separate secondo myheritage hanche se il primo tg qualcuno ha stessi numeri inizio e finali con me e con alcuni del secondo tg sempre sul canale 16.ma ognuno di loro e separato nelle loro corrispondenza sono io che sto nelle corrispondenze del primo tg e sto nelle corrispondenze nel srcondo tg che significa?e un muro di mattoni.
LikeLike
Kevin.quel gruppo che ti parlo sempre sul canale 16 ci sono tra loro chi e piu vicino e anche tra di loro sono cugini tra 4 e 5 grado qualcuno tra loro e mezzo cugino 1 grado 2 grado tra loro pou anche tra di loro sono cugini come me tra il 4 e 5 grado con la triangolazione.
LikeLike
Kevin – yes – that is as it should be. A Triangulated Group is two things: 1. a group of Matches who share the same, DNA; 2. your DNA segment (represented by the full amount of all the shared segments in the TG. Your segment came to you from a parent who got it from one of their parents, etc. all the way back to the first of your Ancestors who had that segment. Say that first Ancestors was your 5XGreat grandparent, 8 generations back. That Ancestor had the full TG segment included in whatever they passed down the line to your parent to you. So you may have some 6C Matches to that Ancestor. You may also have some 5C Matches to his/her child who is your Ancestor; and some 4C, 3C, 2C and 1C Matches to other Ancestors *on this same line* – each one of these Ancestors along this line had *at least* (maybe larger) the segment in you TG. Congratulations on the new Matches. Jim
LikeLike
“Just be aware that a Match with one shared segment can only be genetically related one way.”
How on Earth has this escaped me?!
Thanks
LikeLike
Ciao jim kevin.adesso ho il test di mia madre e il test mio e abbiamo molti match ho da italia del sud come nou ho italiani in srati uniti ho match da paesi balcanici.ma mia madre sul chromosome 5 ha molti match con persone da albania e altre sempre balcani e I match hanno icona della triangolazione e I segmenti ci sono certi che e 14,6cM altri 12 cM ecc.puo essere una regione di pileup?grazie .
LikeLike
Jim: I am learning a lot from the approaches you are describing regarding shared matches, particularly with some of the practical cut-off points with respect to “amount” of DNA shared… My question is how much emphasis do you put on “number of segments shared” when comparing DNA matches among individuals at this point or does that wash-out, since it may correlate with “amount” of DNA ? Also, am I correct in assuming that if Ancestry indicates that two people share “two segments” that both segments at least have to be >7-8 cM… Thanks George
LikeLike
George, Thanks for the kind words. Re: # segments – I usually don’t pay any attention. IMO, it’s like trying to sharpen a marshmallow. I’ve seen Ancestry with 9cM total in 2 segs, but when you look at the actual – pre-Timber – the total is somewhat larger. However, I do expect such segments to be from Ancestors beyond my genealogy horizon – if I don’t see a link quickly I move on. Jim
LikeLike
Ciao jim.kevin sto imparanto molto dai tuoi post.allora quel gruppo su myheritage hanno tra loro una parentala piy vicino cone mezso cugini tra loro ma anche tra loro sono tra il 2 3 4 grado.mentre con me questi match sono tra il figli di cugini di 4 grado ho cugini di 5.ce una triangolazione cone ti ho detto tante volte loro non anno alberi importanti e poi ce il problema del chr 16che ho due Tg sul 16 e sono sul mio lato del test ormai mia madre non fa parte e non riesco ha capire doce potra esserci una relazione.
LikeLike
I enjoy reading your approaches, Jim, because this tool is so new there are so many directions to go in.
My favorite approach lately has been to try to identify the two common ancestors for ALL of the 20 cM+ matches that I have in a particular triangulation group. I identify many sets of MRCA’s, perhaps dozens, among my many matches and try to see where they line up. I’ve done about 30 of these deep dives and have solved about 15 of them – and all 15 couples were born between 1590 and 1700.
This works for British Isles – not so much for other geographies (someday…). And FamilySearch Family Tree is a terrific free tool to look back in time quickly, particularly for colonial New England.
LikeLike
Rich, Thanks for your feedback and approaches. The data is there to solve many questions – it’s about finding the most efficient way… Jim
LikeLike