GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-11 > 1288895494


From: Ann Turner <>
Subject: Re: [DNA] Where does the 7cM block come from?
Date: Thu, 4 Nov 2010 11:31:34 -0700
References: <27088983.314105.1288816730337.JavaMail.root@vznit170132>
In-Reply-To: <27088983.314105.1288816730337.JavaMail.root@vznit170132>


There are several issues here.

Yes, 7.7 cM was selected for the threshold in the hopes that it would
identify regions that are Identical by Descent (IBD). Selecting a threshold
is a balancing act between false positives and false negatives. A false
positive would be a segment that wasn't inherited through the pedigree. We
could call that Identical by State, or IBS. A false negative would be a
small segment that was inherited through the pedigree, but it was so chopped
up that it doesn't meet the threshold.

However, reasonably large segments can make "quite a trip" down through the
generations. Each time it survives one more generation in the 50-50 lottery,
it's as if you've rebooted the clock. See the section called "Inheritance
Patterns" in http://isogg.org/wiki/Relative_Finder. It's certainly quite
possible for a single large segment to come from a single ancestor. We are
seeing what did survive, regardless of how improbable it might have seemed
in the beginning.

I may need to rethink what constitutes a large segment for Family Finder,
however. In a previous thread, I showed an example of how a smaller segments
within a child's match did not occur in either parent. One such paradox was
resolved when I looked at the child's two haplotypes. It wasn't really a
segment at all.

More recently, I looked at how many of the *individuals* in the child's
whole match list did not show up in either parent. It was 41/108, quite a
high percentage. All of them were in the "speculative" category, but some of
the longest blocks were sufficiently long that I don't think "fuzzy
boundaries" could account for them. That's when a few alleles from the other
parent *appear* to continue the long consecutive run of half-identical SNPs.
I often invoke that explanation for similar cases at 23andMe where the
child's match is close to the threshold, but the percentage of such cases is
quite a bit lower there.

This is just one instance, and it might not be a typical case. I'd be
interested in hearing some more stats from father/mother/child trios about
how many of a child's matches do not occur in either parent, and the lengths
of the longest blocks. I used an Excel spreadsheet with the COUNTIF function
-- I can send a template to anyone who's interested.

The big question is what it means if you succeed in tracing a common
ancestor for your DNA matches in the speculative category.

Ann Turner

On Wed, Nov 3, 2010 at 1:38 PM, Jim Bartlett <>wrote:

> Anne, et al,
>
> I am now at 90 Family Finder Matches. Each one has at least one
> 7.70cM block (in one case there are 5 such blocks and in 5 other cases there
> are 2 such blocks). The corresponding SNPs range from 500 to 5623. Clearly,
> my results are not rich in close matches, but I've been persistent and
> doggedly traced every lead in looking for common ancestors. Ten of my FF
> matches and I have now determined paper trails to common ancestors,
> indicating we range from 3rd to 10th cousins. All but one of these are from
> common ancestors in Colonial Virginia, some back to the 1600s.
>
> Fresh from the FTDNA Conference in Houston, I'm now convinced that any
> block below 5.0cM is almost certainly in the Colonial "DNA soup", and
> probably not really traceable to one line. In fact, upon close examination,
> I find many of these small blocks are present in several different matches
> who have common ancestors across more than one of my grandparents.
>
> So I'm now focused on the large, 7.7cM or larger, block that each of my
> matches has (and 84 of them only have one this size). It appears the 7.7
> cutoff size was selected, inter alia, to give a high probablility of
> "Identical By Descent". So if this is the case, then can I assume that this
> "large" block was present in the common ancestor, and was thus included in
> the DNA passed down through every generation to me and my FF match? Easy to
> visualize for a third cousin; quite a trip for a tenth cousin. Or was this
> large block probaby from the "general gene pool" in Anne's email below?
>
> Or is this just the case if I can find a close 3rd to 5th cousin, and that
> same "large" block is just a happenstance, a false positive, in the case of
> a 10th cousin?
>
> And then back my question of several days ago - would that large 7.7cM
> block from say a 5th cousin, be more likely to come from a 3.85cM block from
> each of the parents (making up the common ancestor couple) that just happen
> to fit together in this way; or is it more likely that this 7.7cM block came
> from one or the other of the two parents that were common to me and the FF
> match?
>
> As a genealogist of over 35 years, I'm very happy with Family Finder - so
> far, I've found new cousins in over ten percent of the names and emails
> given to me under the FF program. No knowledge of DNA required, only faith
> that it works. Compare that with zero common ancestors found through FMS of
> H3* (Dad) or T2b6 (me).
>
> Jim Bartlett
>
> On Nov 1, 2010, *Ann Turner* <> wrote:
>
> In another thread, I wrote:
>
> On Fri, Oct 29, 2010 at 6:47 AM, Ann Turner <> wrote:
> > I suspect many if not most of the segments at the 5th cousin level come
> > from the general gene pool, not the one ancestral couple that has been
> > identified.
>
>


This thread: