Hi Jim, I was waiting for you to react to this, as I know we've discussed the point of how many TG's there are. It's obviously determined by the size of these groups (meaning how far one goes back). But when I saw that I immediately thought of your magic 400 number. Now Ancestry uses a minimum of 20cM to show them as shared matches, as I wrote I can only imagine that this is such a large number due to colonial ancestry on her BF's side. Btw, we are still reading the data, she has even more shared matches groups now and it will be interesting to see how many turn out to be TG's. Will surely feedback when we do have a much larger base to calculate on, as you know we're still beta testing with a handful (literally) of users only. Btw, when you say 400 is that with splitting them into the smallest possible groups? Many times we do see TG's where we have smaller TG's at each end which are only connected by some DNA cousins who share the whole segment. That would probably go above 400. Andreas > On 1 Oct 2017, at 20:44, Jim Bartlett <[email protected]> wrote: > > would

(Perhaps echoing Andreas) I don't quite get the value of "counting TGs," because Mom has K TGs one generation back, but then N (>K) TGs another generation back, etc. Saying "Mom has 273 TGs" doesn't seem hugely useful if some of them are TGs traced to a 3-generation-ago MRCA pair but others are TGs traced to an 8-generation-ago MRCA pair--i.e. they're not "equivalent" TGs. Wouldn't a TG be usefully defined to be "final" only when the size of the HIR has reached the "lower limit" below which we can't reliably call matches (in the TG) IBD? Else any TG around an HIR larger than that might be (probably is) subject to being subdivided ... Best, Eric > -----Original Message----- > From: GENEALOGY-DNA [mailto:[email protected]] On > Behalf Of Andreas West > Sent: October 2, 2017 03.20 > To: [email protected] > Subject: Re: [DNA] Difficulty understand Ancestry's shared matches > I was waiting for you to react to this, as I know we've discussed the point of > how many TG's there are. It's obviously determined by the size of these

Eric Just as we use cM to roughly estimate the cousinship, I use total TGs as a quality assurance factor - as in: not too many, not too few. If I would up with 8 TGs on Chr 5 at the grandparent level; I'd be looking hard at flipping at least one, maybe two, of them. In general, at the grandparent level, we don't expect to see more than 4 crossovers (5 TGs) on one Chr. Not that it cannot happen, it generally does not. The point is - at each generation, does it look reasonable. I also throw out the 400 TG number to indicate, roughly, what you're getting into when you decide to map. And, yes, for a given shared segment threshold, each person's map will wind up with one set of crossover points, and ancestral segments between them (represented by TGs). Jim Bartlett - atDNA blog: www.segmentology.org > On Oct 2, 2017, at 5:41 AM, Eric S Johnson <[email protected]> wrote: > > (Perhaps echoing Andreas) I don't quite get the value of "counting TGs," > because Mom has K TGs one generation back, but then N (>K) TGs another > generation back, etc. > > Saying "Mom has 273 TGs" doesn't seem hugely useful if some of them are TGs > traced to a 3-generation-ago MRCA pair but others are TGs traced to an > 8-generation-ago MRCA pair--i.e. they're not "equivalent" TGs. > > Wouldn't a TG be usefully defined to be "final" only when the size of the > HIR has reached the "lower limit" below which we can't reliably call matches > (in the TG) IBD? > Else any TG around an HIR larger than that might be (probably is) > subject to being subdivided ... > > Best, > Eric