Archiver > GENEALOGY-DNA > 2011-10 > 1319485776

From: "Lancaster-Boon" <>
Subject: [DNA] How should ISOGG define "private SNP" and "provisionalSNP"?
Date: Mon, 24 Oct 2011 21:49:38 +0200

Hello Jared

A few quick remarks:-

First, an operational definition will typically tend to approximate a
theoretical one in some way and I am quite confident that this was the aim
of ISOGG. So if this is no longer true, then it is simply a matter of an
operational definition which no longer achieves what it set out to do.

Second, a critical problem with many operational definitions of private SNPs
is that they do not consider the importance of negative tests. There is a
big difference between an SNP found in 1 person and tested in 1 person, and
an SNP which has been carefully cross-checked by testing 100 people largely
chosen for their haplotypes and SNP status in order to cover all

The 99 negatives make an SNP appear to be securely private, not the one
positive test.

Best Regards


It is useful to separate theoretical definitions and operational
definitions. For example, the theoretical definition of a private SNP might
be: "a SNV that exists in only one family in the world, having arisen from a
de novo mutation no more than four generations ago." A operational
definition for a private SNP might be: "a SNV that is not found in the most
recent release of [insert your favorite database name, e.g., Kaviar or the
interval FTDNA database]." One can never be sure that someone somewhere does
not have a particular SNV, so there will never be certainty that a
particular variant meets the theoretical definition. Therefore, one needs to
have clear operational definitions.

Jared Roach, M.D., Ph.D.
Senior Research Scientist
Institute for Systems Biology
401 Terry Avenue North
Seattle, WA 98109
phone: (206) 732-2108

This thread: