Ask A Scientist Molecular Biology Archive

name        James B.
status       educator
age       60s

Question -   Your experts say it is not possible to identify race from DNA, yet two articles 
in the New England Journal of Medicine in March 2003 and an article in the New York Times 
relating to DNA tests used to narrow suspects in a criminal case in Louisiana 
(NY Times, 6/3/2003) seem to call this into question. Any response from your experts?
-----------------
It was not possible before the draft of the Human Genome was released.  We have much more 
information than before and the databases are being updated daily.  It may be possible in 
the future to identify a number of attributes via DNA.

vanhoeck
=====================================================
I am not the genetics expert; however, I am a student of "scientific methodology" and the 
sources cited here are excellent examples of how it is (not) applied. Consider the references 
in NEJM.
The two articles are not peer reviewed research papers, but appear in the "Perspectives" section 
of the journal, which is a give-and-take section where different positions and contrasting views 
on controversial topics are put forth -- These do not have to meet the same criteria required of 
research papers. The journal points out that the topic is "fraught with sensitivities". So 
neither article "proves" anything.
The article in the N.Y. Times (with all due respect it is a newspaper, not a scientific journal) 
reports an assertion by a Mr. Tony Frudakis, chief scietific officer of a company DNAPrint 
Genomics Inc. But an assertion is not only not proof, it is not even evidence.
The article describes the arrest of one Derrick Todd Lee on May 27 as the alleged killer. He 
has been charged, but no jury has convicted him. He has not even been brought to trial. 
Nonetheless, Mr. Frudakis claims in the N.Y. Times article "...that is why the case ws solved 
two months after we ran the test for them." I think the claim of "solved" is premature, 
especially in the face of recent events where a number of inmates on death row in several 
states have been found innocent on the basis of DNA analysis.
DNA can make a strong negative case that a person's DNA does not match evidence; it is very 
much more uncertain in making a positive correlation. Careful researchers in this area state 
that certain DNA analyses are not inconsistent with this or that person -- the double negative 
states possibility, not identity.
DNAGenomics' "test" involves determining "how much of a person's ancestry comes from each of 
four groups: sub-Saharan African, East Asian, Indo-European, and Native American." from which 
Frudakis claims to be able to estimate the individual's skin
tone. Since all of these groups, roughly speaking, have dark complexions, of course any "match" 
will fall within that ethnic sample space. If people of Nordic ancestry were included, would 
there have been any differences in the DNA profiles? This is a beautiful example of the argument 
I call: THERE ARE NO TIGERS IN MY BACKYARD. It goes like this. I sprinkle salt in my backyard
everyday to ward off invasion by man-eating tigers. You ask, "Have you ever seen any man-eating 
tigers, or any tigers at all, in your backyard.?" To which I respond, "No! See how well my 
sprinkling salt works!!"
The partial definition of the human genome is a tremendous advance in science that no doubt will 
provide enormous advances in many fields, but the methodology can also be misused. FYI the N.Y. 
Times article cites two academic DNA "experts" who doubt Frudakis' claims. The example of "cold 
fusion" reminds us to be very careful about claims made for any invention, technology, or 
discovery by those who have a financial interest in the success or validity of the claim.
 
Vince Calder
=====================================================