Archiver > GENEALOGY-DNA > 2010-03 > 1268427222

Subject: Re: [DNA] Question about matches on Chromo #7 at 23andme RF
Date: Fri, 12 Mar 2010 15:53:42 EST

Yes, the assumption is that no-calls would have matched. They're sprinkled
here and there, each one surrounded by dozens to hundreds or even thousands
of good calls. It's certainly possible that ignoring a no-call would
artificially extend the long run of consecutive half-identical SNPs by one more
SNP. However, the exact boundaries are already fuzzy, since a half-identical
region could be extended by inheriting an allele from the other parent that
makes it *appear* that the run is continuing. The only pattern that
unequivocally terminates the run is an "opposite homozygote" (e.g. AA in one person
and GG in the other person).

Ann Turner

In a message dated 3/12/2010 12:17:35 PM Pacific Standard Time,

> First, I don't see how no-calls can simply be ignored. A length of a
> segment is calculated from results. If the program reaches the end of a
> string of matching values, and then there is a gap because the next
> values are no-calls, does it extend the calculated segment anyway
> assuming that the no-calls would have matches, or does it end the
> segment length as the last match, assuming that the no-calls would not
> have matched? Either way a value for the segment length or end point has
> to be reported, so it is not just a matter of ignoring.

This thread: