GENEALOGY-DNA-L Archives

Archiver > GENEALOGY-DNA > 2006-07 > 1153800927


From: Vincent Vizachero <>
Subject: What confidence level?
Date: Mon, 24 Jul 2006 23:15:27 -0500


If DNA testing is used properly (in my opinion, that means that the
testing is in support of the evaluation of a hypothesized
relationship), what level of confidence should we use? Or asked
another way, what rate of false negatives are we willing to tolerate.

If two men suspect that they are seventh cousins (i.e. have a common
ancestor eight generations ago) and undertake a pair of 37 marker Y-
DNA tests, at what point do we suggest that the results do not
support their hypothesis?
If their results are a 30/37 match should we suggest that the DNA
does not support the hypothesis, or rather that the DNA is
inconclusive and they should undertake more testing with additional
cousins if possible?

Assuming that we test many such pairs of men who ARE (in fact)
related, about 95% of the time they will have a genetic distance of
three or less. But about five percent of the time they will have a
GD of four or more, and maybe a GD as much as eight. So clearly a GD
of seven IS consistent with the stated relationship, albeit at a low
confidence level.

Vince

This thread: