This is another blog post that gives you some idea of what to expect with autosomal DNA and your segments. In this post we’ll look at the formation of a TG (Triangulated Group). We’ll walk through the steps:
- Start with overlapping segment data
- Simplify the data by rounding
- Sort by Chromosome and Start location
- Then Triangulate the segments (no genealogy required)
- Highlight one of the two resulting TGs
- Show this data graphically – like you’d see in a chromosome browser
- Overlay the total TG
- Then use our imagination and x-ray vision (or GEDmatch) to show what the ancestral segments of the Matches might look like
- Do some analysis…
Figure 1. Some overlapping segment data
Letters represent Match names – data is taken from my spreadsheet.
Figure 2 – divide the Start/End locations by 1000000
It’s much easier to read the Start/End locations in Mbp; and it’s just as accurate for genealogy.
Figure 3 – the data is sorted by Chromosome and Start location
This makes it much easier to see overlapping segments.
Figure 4 – this shows the results of Triangulation into groups 16A and 16B.
No genealogy was involved in this process – it’s purely a matter of comparing segments at 23andMe or GEDmatch; or looking for ICW Matches in this list and each ICW list at FTDNA. Again, this is real data from my spreadsheet. Often there is more mixture between the two TGs, but I hope you get the idea.
Figure 5 – Here is only TG 16B data
It’s still arranged by Chromosome and Start location.
Figure 6 – Same data and the shared segments displayed graphically
This is how you’d see the data in a chromosome browser. Note the top 11 bars will all match each other. The bottom bars will usually all match each other too, and they’ll usually also match the top 8 bars, but maybe R and N will not match at the 7cM level at GEDmatch. Just lower the level to 500 SNPs and 5cM and you’ll find there is enough for a Match. Let’s see what the TG for this data looks like…
Figure 7 – Now the fun begins…
Usually the TG is pretty clear cut, but I’ve intentionally selected one with two kinds of ambiguity. In almost all cases the ends of the segments are fuzzy. You can read about Fuzzy Data in my blog post here.
Judgment is needed at this point. I’ve shown the “guaranteed” TG in red, with orange tips where the data looks fuzzy. I want to emphasize that this is NOT a problem for genealogy – the TG (wherever the true crossover points are that define the TG Start and End locations – somewhere in the orange areas) represents an ancestral segment from one of your ancestors. The fuzzy ends are not an issue. Your Matches will share a Common Ancestor with you – and that’s where the focus should be. The crossovers defining the real TG will be somewhere in the fuzzy orange tips.
You can also see that this data indicates a probable more distant crossover point – around say 51Mbp. In this case the top 8 Matches and Match S are probably closer cousins sharing a larger, closer segment with you. At this point you might want to review Crossovers by Generation here. Going back one or more generations we may see the large red TG being subdivided into two smaller ancestral segments – each with its own Common Ancestor each one of which is ancestral to the Common Ancestor for the red segment. In this case the last 8 Matches (T, J, Q, L, M, I, A and K) will have a different, more distant CA than Matches G and H. The main, red, TG may be from a 5G grandparent, and the smaller, green and purple segments may be from a 6G or 7G grandparent. Actually the purple segment, as an example, may pass intact through several generations, and you could share this same segment with 7C and 8C…
Again, most of your TGs will be tighter and from a single CA, but I wanted to take this opportunity to show what sometimes happens. You can avoid any conflict by watching for this situation and just declaring two TGs in this case – see the green and purple bars. Then it’s like the case where a close cousin spans more than one TG – the close cousin will help you define a larger segment from a closer ancestor, and the close cousin, along with different groups of Matches in different TGs will share more distant Common Ancestors with those TGs, but those more distant CAs will be ancestral to the MRCA the close cousin shares with you.
So now let’s use our imagination a little (or we could actually Triangulate this area from the perspective of some of our Matches. In this next Figure 8, I’ve guessed at what the ancestral segments might be the Common Ancestor down to the different Matches – as shown in green.
Figure 8 – showing ancestral segments for all Matches
In some cases our Matches have somewhat larger ancestral segments than we have, or they might have segments that extend over one end of our ancestral segment, or the other. In all cases the blue represents the overlap between each Match’s ancestral segment and our own ancestral segment. And the data is not exact, so the ends often don’t line up vertically. The long green bar at the bottom of Figure 8 is a segment an ancestor passed down to living people – you got part of it, the red part.
At GEDmatch it’s often fun, and instructive, to compare two Matches to each other. Sometimes they turn out to be parent/child, or siblings. This is the exercise you’ll want to do if you are trying to map an ancestor.
So, again, if you have collected a lot of shared segments from FTDNA, 23andMe and GEDmatch, they have to go somewhere. It’s not hard to compare them to each other and see where they Triangulate. If they are IBD, they have to go on one chromosome or the other. When you do this you’ll find there are natural break points where the crossover points are located (often the precise location is a little fuzzy). Just look at the data above.
05E Segment-ology: Anatomy of a TG by Jim Bartlett 20160204