PCR vs. DNA Fingerprinting?

Tom Thatcher ttha at troi.cc.rochester.edu
Wed Jul 20 13:32:39 EST 1994


In <53123.rin0mxw at bumed30.med.navy.mil> "M. L. Huang" <rin0mxw at bumed30.med.navy.mil> writes:

>On Thu, 14 Jul 1994 20:49:45 CDT, 
>U27111 at uicvm.uic.edu  <U27111 at uicvm.uic.edu> wrote:

>>I just read the Associated Press article on the DNA fingerprinting
>>done on the glove found in the O.J. Simpson case.
>>
>>They make a distinction between DNA fingerprinting and PCR and this
>>confused me.  I thought all DNA testing was as a result of PCR,
>>otherwise, how do they have enough of a sample to work with?
>>
>>As a matter of fact, the way the article described it, it almost
>>sounded like a southern blot.

>Kathy,

>I think the process is like this:
>     first, PCR to get enough DNA
>     cut with specific enzymes
>     Southern blot
>     look for unique/signature bands, thus polymorphism or fingerprints

Yes and no...
There is a PCR-based kit available for use by (otherwise untrained) police
crime labs that produces combinations of colored dots on a small number
of spots (like 6) on a test strip.  I assume that a small sample is
amplified with six sets of primers and the result is simply +/-
amplification.  I assume the primers are built to polymorphic areas.
If the blood in OJs car is 1,2,6 and the victims are 2,3,4 and 1,2,5,
then the victims are excluded as possible sources [totally hypothetical
example]

RFLP analysis involves probing a Southern blot after digestion with
restriction enzymes.  I suppose PCR can be used to generate the starting 
material, although I don't know.  RFLP takes many weeks and requires a
trained lab that follows FBI guidelines.

Either test is useful for _excluding_ identity--any mismatch, assuming the
test was performed accurately and the sample was clean, proves that the
samples came from different people.

If all the tested results match (bands in RFLP or colored dots in PCR)
then its a numbers game--what are the odds that the actual criminal and
the suspect would match if they were unrelated persons.  In this case,
RFLP is much more rigorous and dependable.  In fact, the PCR-based kit
is admissable in court in very few states.

Tom Thatcher
University of Rochester Cancer Center



More information about the Methods mailing list