Archiver > GENEALOGY-DNA > 2010-02 > 1266998057

From: "Tim Janzen" <>
Subject: Re: [DNA] Family Finder question
Date: Tue, 23 Feb 2010 23:54:17 -0800
In-Reply-To: <>

Dear Connie,
The blog at covers
some of the issues you are interested in. Family Finder and Relative Finder
are going to be better at locating people who share a common ancestor who
lived in the early 1800s than they will at locating people who share a
common ancestor who lived in the early 1700s. Fifth cousins theoretically
only share about .0488% of their DNA in common. Whether people who are 6th,
7th, and 8th cousins will share matching DNA segments that meet the
threshold for a match that Family Finder and Relative Finder have
established is to a large extent a matter of luck. As with a lottery, you
have a higher chance of winning if you buy more tickets. The way to
increase your chances that you will find people who have matching segments
with you or your relatives is to have more of your relatives tested. Thus,
if you have a genealogical brick wall on a particular line in the early
1700s, the best way to attempt to break through that brick wall from a
genetic genealogy standpoint is to test as many descendents of the ancestor
whose parentage at this point is unclear as feasible. I posted the section
below on the ISOGG list recently and it might be worth repeating it on this
list for you.
Tim Janzen

"You have hit on one of the core issues I have struggled with as
well as I have been trying to analyze the Low German Mennonite data for
about 30 people that we have back from 23andMe. See for background. I know
my dad shares segments with people who are his 5th and 6th cousins, but I
can’t at this point be certain that those segments were necessarily from the
ancestors I know they share in common. Instead the segments could be from
other ancestors that they share in common that I have not yet been able to
identify due the fact that some of the pedigree charts of the participants
aren’t complete back to the 6th to 8th generation level. The bottom line is
that you have to be very careful how you interpret autosomal DNA results,
particularly when you are trying to prove connections at the 5th-7th cousin
level of relationship. Testing multiple people from multiple lines of
descent is probably the best way to try to tackle this problem. As a
general rule, the more endogamous the population you are dealing with (such
as Low German Mennonites or Jews), the more challenging it is to correctly
interpret the results as to the precise ancestor or couple from whom the
shared segment or segments were inherited.
For those who are trying to overcome a specific brick wall, this
is the approach I am currently taking:
1. Identify an ancestor whose ancestry is unknown, preferably one that is
not more than about 6-7 generations back and preferably one that you have a
full name for.
2. Test the oldest generation of people in the family and as many of their
other close relatives (1st and 2nd cousins) who descend from this ancestor
whose ancestry is unknown as feasible. This helps increase the amount of
autosomal DNA that you will have that came from the ancestor whose ancestry
is unknown.
3. Identify candidate parents for the ancestor whose ancestry is unknown.
4. Test as many of the oldest generation of descendents of each set of
candidate parents as feasible. If descendents of all of the candidate
parents are not available, then test descendents of the parents or
grandparents of the candidate parents.
5. Analyze the data (looking at total number of shared segments, total cMs,
etc.) looking to see if there is a reasonable possibility of a connection
with descendents of any of the candidate parents.
6. Once you have zeroed in on a probable set of parents for the ancestor
whose ancestry is unknown, then do additional testing on descendents of
close relatives of the probable parents to help confirm your conclusions
about the genealogical relationship."

-----Original Message-----
[mailto:] On Behalf Of Connie
Sent: Tuesday, February 23, 2010 9:02 PM
Subject: [DNA] Family Finder question

If I am understanding the 23andMe and the new Family Finder
testing, it works best at finding 3rd - 5th cousins.  I located
cousins at that level and exchanged Gedcoms for years and have little
desire to do that again.  Other researchers in my projects also feel
that way as we are all trying to push back beyond 1700.

So my questions is - am I understanding the limitations of Family Finder
correctly - will it be ineffective at finding relationships in the time
period that I need? 

Connie McKenzie

This thread: