GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2003-11 > 1069444164


From: "Earl Beaty" <>
Subject: Re: [DNA] Genetic Distance calculation method -- which method is most correct?
Date: Fri, 21 Nov 2003 12:49:24 -0700
References: <OF1A2B197A.A0F0DD49-ON85256DE3.007525EA-85256DE3.0078EBB1@downstate.edu> <REME20031119183221@alum.mit.edu> <3FBD7168.7000706@kerchner.com> <REME20031120224215@alum.mit.edu> <3FBE63E0.6030000@kerchner.com>


Charles,

Much of the problem with these things is posing the right question, and
looking at the problem from the right point of view. Perhaps I can offer a
useful alternate point of view.

The situation is like a board game (say Monopoly with a little modification)
in which a flip of a coin determines movement. Heads-one step forward;
tails-one step backward. If you find that a player has advanced 3 steps you
can ask how many flips did it take to get that far. It is possible to get
there in 3, but there is the real possibility that in doing those 3 a tails
result appeared. So on the average it take more than 3 flips to move a
distance of 3 in either direction. John's formula says that on the average
you would expect to flip 9 times to gain a net movement of 3. Nine seems
intuitively a little high but one can analyze the math or just spend a
little time flipping coins. The case of a distance of 2 which John cites is
a little easier.

--Earl Beaty

----- Original Message -----
From: "Charles" <>
To: <>
Sent: Friday, November 21, 2003 12:13 PM
Subject: Re: [DNA] Genetic Distance calculation method -- which method is
most correct?


> John:
>
> Re. Genetic Distance calculation/estimation. Using your "sum of the
> squares of the differences" method vs. "simply adding up the
> differences" method used by FamilyTreeDNA.com in their Y-DNA surname
> group project tables of calculating/estimating Genetic Distance between
> two people.
>
> Again, for these examples let's ignore the markers which need special
> attention such as multi-copy markers, etc., for counting differences. My
> math method question is how does one add up the marker differences once
> they have been determined for each marker.
>
> Here is an example for discussion purposes:
>
> Example 1:
> If two people sharing the same surname in a surname project group differ
> at three markers such that they differ at the three markers by 1 step at
> each marker, then:
>
> FamilyTreeDNA.com method says the Genetic distance is: 1+1+1 = 3
>
> John Chandler method says the Genetic Distance is: (1*1)+(1*1)+(1*1) = 3
>
> Example 2:
> If two people sharing the same surname in a surname project group differ
> at only one marker by 3 steps, then:
>
> FamilyTreeDNA.com method says the Genetic distance is: 3 = 3
>
> John Chandler method says the Genetic Distance is: (3*3) = 9
>
> It doesn't make sense to me to emphasize/weight a 3 step difference at
> one marker (whether or not they were three 1 step mutations over time or
> a single one 3 step jump) much more than three 1 step differences at
> three markers. The sum of the squares of the differences method makes a
> 3 step difference at one marker imply that the two people are three
> times further apart in Genetic Distance as compared to two people who
> have 1 step differences at three markers? John, you're a math and
> statistics expert. I am not. Can you clarify more why your method for
> calculating Genetic Distance is the correct one, using my above
> examples? Can others please comment too. Thanks.
>
> Charles
>
>
>
>
>
>
> ==============================
> To join Ancestry.com and access our 1.2 billion online genealogy records,
go to:
> http://www.ancestry.com/rd/redir.asp?targetid=571&sourceid=1237
>


This thread: