Dear Science,

With the acquittal of Amanda Knox for the murder of Meredith Kercher, supposedly because of bad DNA evidence used in the original trial, I'm wondering again about DNA evidence. Wasn't the truth of DNA testing a part of the O. J. Simpson trial? When I read about DNA evidence proving something in a trial, what exactly is going on? How can DNA be used to prove if someone is innocent or guilty? How can it be wrong?

Seattle Sleuth

DNA testing at a crime scene is a bit like fingerprint testing: It's a way of placing a person at a particular location by detecting a pattern that is (more or less) unique to the individual.

The DNA in each of our cells contains about three billion base pairs (A C T G) arranged into a sequence. For some stretches, the sequence is quite similar from person to person. For example, the stretches of DNA that describe how to make proteins need to be similar from one person to another; too many differences would result in proteins that don't work. Other parts of DNA are radically different from person to person. These parts of DNA are often not doing much—repetitive sequences of junk. Therefore, there is no selection pressure to keep them the same from one person to another. They can, and do, drift. These are the stretches best suited for DNA fingerprinting an individual person.

Let's think of a different problem: trying to tell if two brothers are identical twins. If the twins are really identical, even these rapidly changing stretches of DNA should be identical in each twin. If the twins are fraternal, some (but not all) of these fingerprinting sequences of DNA will be different. Since some of these fingerprint sequences will be the same in the fraternal twins, it's only when we find a difference that we can definitively say the twins aren't identical. For each sequence we find that is the same, the chance that the brothers are identical twins gets a little bit higher. After looking at thousands of fingerprinting stretches of DNA, the chance becomes overwhelming.

The doubt with DNA testing doesn't come from the DNA fingerprinting technique. The doubt comes from how sensitive the test is—making it easier to fake results. The tiniest smudge of DNA can be detected with modern techniques. Believing DNA test results requires one to believe in the honesty and competency of the people collecting and processing the samples. The tests are so sensitive, it's easy to intentionally or unintentionally find a person's DNA fingerprint where it wasn't. Hence the doubt in the Amanda Knox case. Once you show incompetence in the collecting and testing of evidence, the results can be contested.

Sequentially yours,

Science

Send your science questions to dearscience@thestranger.com.