Archiver > GENEALOGY-DNA > 2005-09 > 1127932305

From: "Ken Nordtvedt" <>
Subject: Re: [DNA] Re: Another set of extended DYS464, DYF385S1, DYF399S1 results
Date: Wed, 28 Sep 2005 12:31:45 -0600
References: <>

Yes, because this is not testing for a cancer gene, but rather is just an
offshoot of a hobby like genealogy, I think there has been some casualness
in the labs reporting their error rate, or even trying to determine it.
They are mortal, and if we are reasonable we can not expect perfection, but
it would be nice to know the rates, both for STRs and SNPs.

I bet if these genetic profiles are used in court to convict someone of
something or establish innocence, the counsel for the other side will demand
some kind of measure of accuracy and error rates for the numbers. Since I
do population studies mainly and am looking for statistical properties of a
population, I can tolerate occasional errors more than those wanting the
strongest possible proof that Jack the Ripper or Charlemagne were or were
not in their ancestry.

> List,
> I am a complete novice but if different labs get different results on
> the
> same person.. how often does this happen? This would seem to be a serious
> problem if folks are paying for tests that cost a pretty good chunk of
> money and
> then may receive a false or partially inaccurate 12,25,37, 43 or whatever
> marker result. If the labs are over booked, does that make the ones
> testing the
> samples work too fast and not have proper controls over results? Too much
> pressure? Do some labs have better testing methods and history of
> accuracy? Are
> samples prone to some inaccuracy depending on how and when you take them
> and mail
> them in and handle them? If so, what approximate percent of error exists
> in
> the total folks that have been tested at the labs?

This thread: