GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2002-11 > 1037855629


From: "Richard Hite" <>
Subject: RE: [DNA] MRCA calculations
Date: Thu, 21 Nov 2002 00:15:31 -0500
In-Reply-To: <5.1.0.14.0.20021120124346.047a3ab8@incoming.verizon.net>


John,

Thanks for your comments. I just noticed a typo in my initial message.

The 464 readings should be as follows:
Subject ASubject B
464a15 17
464b17 17
464c17 18
464d18* 18

*Mistakenly given as 17 before.

The other two differences again
38811 12
45816 18

Your comment about three mutations on 464 really had me puzzled because I
saw two. Then I realized my error.

With this, and the other two mutations, FTDNA had cited a genetic distance
of six between these two subjects. When you said seven I was really
confused and then I saw my typo.

This may not alter your perception of this comparison, but I wanted to be
sure you (and everyone else) had the correct numbers.

The only comment of yours that I questioned was this one "Note that two
unrelated men will typically have a mismatch of 5 to 10 steps." Is that for
a 12-marker test or a 25-marker test? It has been my experience with the
study I am involved in (17 samples so far), unrelated men have been 20 or
more steps away at the 25-marker level. I am one of the subjects and the
one relative I have in the study is one step away from me at the 25-marker
level. The next closest to me is 23 steps away. This is pretty typical.
In the study, we have five pairs of supposed relatives (one a trio), all
exact matches or within one or two steps of each other at the 25-marker
level (except the pair above). Typically, though, none of them are within
20 steps of any non-relatives. It may be noteworthy that in this sample, we
have no fewer than five haplogroups represented (HG1, HG2, HG3, HG4/21, and
HG35). That much diversity in haplogroups within a single surname study
with only 17 total samples so far may be atypical, I don't know.

Richard Hite


-----Original Message-----
From: John S Walden [mailto:]
Sent: Wednesday, November 20, 2002 1:25 PM
To:
Subject: Re: [DNA] MRCA calculations


Richard
Below is my take on your testing issues

At 10:15 PM 11/19/2002 -0500, you wrote:
>I see a consensus emerging here that on a 25-marker test (such as that
>offered by Family Tree DNA) any less that 22 matches is considered
>unrelated.

There are no YES vs NO exact point in the Y-DNA test because of the
random nature of the mutations.

> In a study I am involved in, we have one case of a 21/25 match
>between two individuals with a solid paper trail back to a common ancestor
>(a total of 15 transmissions).

For a 25 marker test and 15 transmissions and .002 rate of change
the expected mutations are:
0 47%
1 35 %
2 13%
3 3%
4 0.6%

The reason FTDNA say 3 or more mismatches is that statisticians like the
95% interval
for their accept/reject point. In your case a 22 marker match has only a
3% chance of being
right.

> Two of the differences, however, are on DYS
>464. These are the readings:
>
> Subject A Subject B
>464a 15 17
>464b 17 17
>464c 17 18
>464d 17 18

DYS464 are 4 different markers but they all come out on the same
test so what is a,b,c, or d cannot be determined. They are just
reported in ascending order.
But in your we can "theorize" a haplotype and count the mutations to
your A and B
Theoretcal
ancestor Subject A Subject B
464a 17 15 17
464b 17 17 17
464c 17 17 18
464d 17 17 18

Steps 2 2

No matter how you look at it and which starting point you pick there
are at least 3 mutations and 4 steps involved.

>Just for the record, the other mismatches are at DYS 388 (A=11, B=12) and
>DYS 458 (A=16, B=18). They match on all the other markers.

Now adding in the above DYS388 an 458 we have another 3 steps for a total
of 7
The odds of two men being related with 6 mutations in 15 births is 1 in
100,000

Thus the Y-DNA results to not support the proposition that these men as have
a common male ancestor in the last 400 years or so.

Note that two unrelated men will typically have a mismatch of 5 to 10 steps.

Also note I did not say it cannot happen - just that the odds against it is
100,000 to 1
It is much more likely there is an undocumented adoption or other event.

<snip>

>Richard Hite

John W

______________________________

This thread: