Archive for July 21, 2008

DNA helix

The biggest story of this past weekend is this L.A. Times piece on the accuracy of the DNA identifications and the statistics used to come up with the probabilities of a random match. Much has been said in the blawgosphere (all of it worth reading) by the Greenfield, the Underdog Mark Katz, the public defender dude and an engrossing lesson in statistics by the Windypundit. They’ve done a wonderful job, so I won’t dwell on it much.

The gist is this: A DNA “match” is when the two pieces of biological evidence share the same markers (loci) at 9 random locations (or 11 or 13 or 15…) The odds of that happening, the State usually contends, is one in 6 billion (and they use that number arbitrarily because it is the population of the world). So in essence, they argue that the only person in the world who could be the source of that piece of DNA is the defendant.

The problem should be evident. Of the millions and billions of loci in the human DNA, 9 or 11 or 15 is an infinitesimal fraction. So, it is possible that if we start looking for matches at 20 or 30 or 100 or 1000 loci, the probabilities will drop. Then we might be talking about 1 in 100,000 or 1 in 20,000. Not bad odds, given the populations of cities and countries.

Still, if these “Arizona searches” (as some are trying to call them in an effort to get the name to stick) haven’t reached your State yet, a number like 1 in 300 billion trillion gazillion quadrasesquecentenellian shouldn’t faze you.

DNA cannot tell a jury when it was deposited or whether the donor actually committed a crime during the deposit. This will not change, even if the odds that it was your guy that deposited the DNA get better.

photo credit: beckita115