Science de jure.
The recent arrest of the BTK serial killer is an example of the tremendous impact science has had in criminology. Here's the skinny on how they caught the bastard -- apparently police had a blood sample from BTK's daughter. A genetic search found that portions of her genome matched up with DNA samples they had from BTK. The authorities then knew that the BTK killer had to be in her family. Mix these scientific results with a little bit of the Five- O action and voila -- they found the killer. Of interest is the reliability of these tests. Current estimates demonstrate that such testing is 99.9% accurate. Yet, does the .1% constitute reasonable doubt? Don't think so? Me neither. However, many attorneys claim sample contamination when DNA evidence is used against their clients. I argue that any technique with 99.9% reliability is pretty damning and that human error is the cause of most genetic evidence mishaps. I read a great interview of Bruce Weir, an expert witness in the O.J. Simpson case. It looks like it's the lack of peer review that makes DNA testing suspect. Remove human bias and the technique is bulletproof, methinks.
Above is a spectral karyotype of the 23 paired human chromosomes.