DNA is present in nearly every cell of our bodies, and we leave cells behind everywhere we go without even realizing it. Flakes of skin, drops of blood, hair, and saliva all contain DNA that can be used to identify us. In fact, the study of forensics, commonly used by police departments and prosecutors around the world, frequently relies upon these small bits of shed DNA to link criminals to the crimes they commit. This fascinating science is often portrayed on popular television shows as a simple, exact, and infallible method of finding a perpetrator and bringing him or her to justice. In truth, however, teasing out a DNA fingerprint and determining the likelihood of a match between a suspect and a crime scene is a complicated process that relies upon probability to a greater extent than most people realize. Government-administered DNA databases, such as the Combined DNA Index System (CODIS), do help speed the process, but they also bring to light complex ethical issues involving the rights of victims and suspects alike. Thus, understanding the ways in which DNA evidence is obtained and analyzed, what this evidence can tell investigators, and how this evidence is used within the legal system is critical to appreciating the true ethical and legal impact of forensic genetics.

Within the U.S., the 13-STR profile is a widely used means of identification, and this technology is now routinely employed to identify human remains, to establish or exclude paternity, or to match a suspect to a crime scene sample.

In order to utilize STR information as a means of human identification, the FBI established the frequency with which each allele of each of the 13 core STRs naturally occurs in people of different ethnic backgrounds. To this end, the FBI analyzed DNA samples from hundreds of unrelated Caucasian, African American, Hispanic, and Asian individuals. Assuming that all 13 STRs follow the principle of independent assortment (and they should, as they are scattered widely across the genome) and that the population randomly mates, a statistical calculation based upon the FBI-determined STR allele frequencies reveals that the probability of two unrelated Caucasians having identical STR profiles, or so-called "DNA fingerprints," is approximately 1 in 575 trillion (Reilly, 2001).

This very small number needs to be put into perspective. Note that this figure refers to pairs of people, and there are many pairs of people in the world. Indeed, for the 100 million Caucasians in the world, there are 5,000 trillion pairs of people, so roughly eight or nine pairs would be expected to match at the 13 STR loci. This predicted matching does not specify which profile is shared by two people, and the chance that anyone matches the particular profile associated with a crime is still very small. The distinction between two people sharing a profile and one person having a particular profile is an example of the so-called "birthday problem." Here, the probability that a person has a particular birthday is 1 in 365, ignoring February 29, but there is a 50% chance that two people in a random group of 23 people have the same unspecified birthday (Weir, 2007).

To perform a forensic DNA analysis, DNA is first extracted from a sample. Just one nanogram of DNA is usually a sufficient quantity to provide good data. The region containing each STR is then PCR amplified and resolved according to size, giving an overall profile of STR sizes (alleles). The 13 core STRs vary in length from 100 to 300 bases, allowing even partially degraded DNA samples to be successfully analyzed. The costs of analysis, both in time and reagents, are significantly reduced by the amplification of all 13 STRs in just two multiplex PCR reactions.

Depending on the complexity of the repeat unit, the different alleles of an STR can vary by as little as a single nucleotide. For instance, the aforementioned D7S820 STR is relatively simple and contains between 5 and 16 repeats of GATA. Another STR, D21S11, has a more complex repeat pattern consisting of a mixture of tetra- and trinucleotides, and it therefore has alleles that vary in size by a single base pair. Because of the need to differentiate single-base differences, PCR products are typically resolved using automated DNA sequencing technologies with software that recognizes allele patterns by comparison to a known "ladder."