GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2010-03 > 1269930411


From:
Subject: Re: [DNA] Family Finder Test
Date: Tue, 30 Mar 2010 02:32:18 -0400


> It should be possible to gather crossover statistics
> as more families partake in the FF tests
____

From what I have seen, the FF FTDNA tests won't do as good a job of
estimating crossover statistics as the RF tests at 23andMe. The FF
tests lump the full matching segments in with the half matching
segments. RF does this as well in the Advanced Labs section but in the
Family Inheritance section, you can actually count the transitions
between the dark full-match with the blue half-match and the white
no-match among siblings. I don't have my results from FF yet so please
let me know if I am mistaken.

I have a Google Doc spreadsheet started to look at the 23andMe RF
transitions at
http://docs.google.com/Doc?docid=0AR3qUyFYAhKuZGQ2M2NrcDZfMGM4NXc4a2hz&hl=en

When comparing two siblings, you actually are taking into account four
chromosomes, each is a combination of two grandparents in terms of
haploblocks.
If you assume that there will not be exactly the same crossover
location, you can take the crossover points and divide by 4. So if
there are 7 crossover
points within a chromosome in the graphic in the Family Inheritance
section, then you would divide by 4 and come up with an average of 1.75
crossovers for that chromosome for that particular sibling pair. There
are likely to be more crossovers than our estimates because some of the
chromosomes have missing information. I am also counting the faintest
blocks that are barely visible for my family, so there probably will be
observer error involved as well.

Does this make sense?

Kathy J.



This thread: