DNA testing reliability

[A_Feeble_Mind]

How reliable is DNA testing when it comes to identifying a murderer? I just watched the 48 Hours Mystery episode entitled "Deadly Ride" and it makes me question several aspects of using DNA for identifying purposes. You can read the material from the episode here.

To summarize, a murder occurred in 1969, but no one was tried. In 2005, there was a trial in which the only evidence (or at least the only evidence mentioned in the show) was DNA evidence on several places on the victim's pantyhose. The DNA could not be identified as blood or semen, but was from some other source, presumably sweat.

There was also some blood on her sleeve. This is the cause of my question regarding DNA testing. The blood matched someone else, a convicted killer. However, it is unlikely that he could have committed the murder as he was 4 at the time. The blood of the convicted killer was, apparently, being processed at the same time as these other samples, indicating that there is a chance of contamination.

Even with no explanation for how the boy's blood could have ended up on the woman's sleeve, the jury finds the defendent guilty. To me, it would seem that it would be essential to explain how the boy (who was over a hundred miles from the crime scene) was involved.

(For a better, complete explanation of all the details, please see the above link.)

The person found guilty was Gary Leiterman, someone who had no criminal history and was a normal, married with grown children person who didn't seem like he was trying to hide something. I know looks can be deceiving, but with my questions about DNA, I really suspect that he is innocent.

So, this leads to my questions: how likely is it that there was some sort of contamination? Doesn't the placement of the boy's DNA at the crime scene at least hint at some contamination? And, even if there is no contamination, is a DNA test 100% reliable?

 

[Ducky]

Well as far as I see it, DNA testing will tell you who the host of that DNA was. As to how it got at the scene is up to debate. Like fingerprints, it only proves someone was at the scene. I could have visited your house as a repair man and shed some hair from my head onto your rug fixing your cable hookup. Someone could have shown up hours later and beaten you to death in that room. My dna is in that room, but it doesn't prove I am a murderer.

Or do you mean how reliable is the dna matching process? That is fallable to human error like all things. Which is something for the defense to persue when dealing with it in a case.

 

[athon]

DNA analysis is incredibly reliable. However, it pays to think of it a little like blood testing. It can falsify a murderer; if the DNA does not match, it is NOT him who left the sample. If it does, it is PROBABLY him.

The probability increases with the amount of sample tested. If X+1 base pairs are matched, it is better than just testing X, as it increases the likelihood that nobody else shares that sequence. Only small parts of the most variable sections of your genome are sequenced, afterall. It is unlikely that somebody would share that sequence, but not impossible.

I'd look up the chances of a match, but I'm being lazy. I know it varies depending on the severity of the crime.

The chances of a stuff up are not scientific, but human. Forensic labs all over the world can and do often make mistakes. I know the Queensland John Tongue Lab's had some massive administrative changes a few years ago due to mistakes being made.

To answer your question, the science is good. The administration of the science is open to question.

Athon