|
Anything above this line is advertising and is not a link to information on my web site DNA Project Notes |
Grouping Participants
When I posted
FTDNA Genetic Distance & Family Group Assignments Anne W. Nelson requested it be expanded to explain more about family group assignments. For now I'm gathering remarks from group admins and placing them on this page; this project/page will probably soon be turned over to ISOGG.Unless noted otherwise, the remarks were posted at one of the below:
GENEALOGY-DNA-L at RootsWeb - to join (free):
Emails addresses have been removed from the below posts.
=================
Anne W. Nelson - 31 Jul 2005
http://archiver.rootsweb.com/th/read/GENEALOGY-DNA/2005-07/1122838368
1. Sharing of an unusual allele for their haplogroup combined with the same surname and a match within the probability distribution for the number of markers. That is, while 12-50 markers have a 90% or better probability of having NO mutations, 1-3 mutations also carries a non-zero probability of occurring, on any ONE man. Therefore, as John Chandler and others have pointed out, we can have documented cases in which the comparison of two men who happen, by chance, to each have 1-3 mutations can result in these two men showing 1-6 mutations. And the more generations they are separated the larger that "extreme number" of mutations can grow.
Therefore unless there is a BETTER match on the number of tested matches, Group Administrators may put a person in a group that is the "best available match at present." It may be that a future testee will be found whose DNA is midway between the two and "splits the difference." Or a future testee may come in who is a better match and the original person will be reassigned to a different group. This science is still evolving and we are still learning.
2. Membership in a shared sub-sub clade with the same surname and match within the realm of probabilities. For example, three men who share the characteristics marker alleles usually found in I1a "Anglo-Saxon variety" are more likely to be related to one another than to a man with the characteristic marker alleles of I1a "Norse" variety, even if they have more than expected the number of mutations between/among them.
3. Paper trail clues -- all born in NC in one decade, moving in similar migration paths, marriages to the same related surnames, etc. or whatever paper trail evidence suggests a possible MRCA, including unusual names or shared middle names for which there is no obvious explanation (i.e., all three show the middle name Day, but there is no known marriage in any to a Day family but all have LNU wives once you get back to 1790s.
4. Some combination of these factors.
=====================
Bill Bailey - 2005 - post to GENEALOGY-DNA-L
"Lowe DNA"
"....I've grown accustomed to thinking of "genetic distance" as a description of how many mutations have occurred..."
When looking at genetic distance, my first thought what was the original, ancestral family NRY DNA STR marker pattern at 25 or 37 markers...?
And then comparing the genetic distance of each of the family members....how much they differ from that original marker pattern...are they transient mutations that will mutate back (be corrected) the next generation; or, a mutation that will continue to increase the genetic distance even further beyond the 21/25 or 32/37 limits.
======================
Nora J. Probasco
2005 post to ISOGG
. . . . . there are many factors that must be considered. And it is hard to compare your surname project with another project. There are some who have few if any mutations and others, like my project, that have more.
A Project Administrator can use the FTDNA tools, but must also factor in the genealogy research info. The better the genealogy research the more helpful it is in determining possible matches. Another factor is "in-betweeners" who can prove relationships too between those participants that seemingly have a large Genetic Distance gap. One also needs to study the markers to see if there are any unique situations with their participants' markers. In my case, I found a unique situation with my participants having a value of 12 in DYS#393 and being in Haplogroup R1b as proved by SNP testing.
==============
John A. Blair - ISOGG
In some ways it actually gets easier as you get more participants and have established groups.
Here is the rationale I use to group participants:
a. The Blair surname dates back to about 1200 (about 25-30 generations)
b. I use the 25 marker test as the basis for grouping participants
c. I use FTDNATiP, and look at probability of a MRCA within 24 generations (600 years according to FTDNATiP).
d. Anyone who have a 50% probability or better of sharing a common ancestor within 24 generation is considered in the initial group.
Based on these criteria at those that match 21/25 or better are considered for the group. I realize that I'm probably more "generous" than most when it come to including someone in a group, but this is just my initial cut.
Using the above selection process will probably result in some participants ending up in more than one group. If this happens I look a which group has the better match.
I also look at the 37 matches if available. These are often the tie breaker and overrule the initial screening process. Normally my cutoff on 37 markers is a 30/37 match.
There are other factors I also consider. If you have a large group you can often calculate a most likely "ancestral haplotype" (the most common values for each marker). This becomes the haplotype or DNA signature you assume the common ancestor had. If this is possible then this becomes the set of values that I use to compare test results. For example in my Group 03 I have 11 participants who have tested 25 markers. When you compare each participant against all of the others in this group we have some group members that mismatch on 4 or 5 markers. But if you calculate the "ancestral haplotype" (which is very evident when you look at the test results) you find that 1 participant is an exact match, 5 mismatch by 1, 4 mismatch by 2, and 1 mismatches by 3.
The final factor I consider is rare marker values shared by participants. Ten of the participants in group 03 have a value of 26 on DYS#390. This is an extremely rare value, occurring in only 1.11% of all test results I have examined.
One other suggestion I would highly recommend. If you intend to try to connect participants to common ancestors, please insist that participants provide you with pedigree charts. Knowing the oldest known ancestor is good but a complete pedigree chart is much more valuable. I create simple line pedigree charts on all members of a group which makes it so much easier to visualize what you are dealing with.
If anyone wants to see how I display my test results and grouping here are some links:
Blair DNA Home Page:
To Do:
|
GKBopp |