What in the world is DNA profiling?

On 31st July 1986 Dawn Ashworth left a friend's house and disappeared. Two days later she was found in a nearby field. Although it was clear that she had been raped and strangled, further investigation uncovered something more sinister-- Dawn had died at the hands of a serial killer who had committed a similar murder two and a half years earlier. Under pressure, the police convicted a local boy Richard Buckland, who seemed to know details of the crime, some of which were not made public. The catch? Buckland was innocent, the real killer was Colin Pitchfork. All it took was a blood sample and the police figured out the actual culprit. How did they do it?

There are many intrinsic characteristics that are unique to each of us. While that sounds like its coming from a life coach, there is a fundamental reason why you're unique: your genome. The genome is a collection of all the genetic information you carry in your DNA. DNA stands for deoxyribonucleic acid, which is a fancy way of referring to the string of characters that determine everything about you; your appearance, your personality, potential health disorders, everything is spelt out in your DNA. Therefore, getting a DNA sample is a great way to identify someone. 

Figure 1: A sign on the public transportation system in Perth. A great way to keep the disruptors at bay. Source. 

Using DNA to identify the killer is a convenient plot line for most contemporary TV detectives. Indeed the technique of using DNA profiling was developed only in 1984 by Sir Alec Jeffreys, a British geneticist who was knighted for this reason. Jefferys had discovered that DNA sequences could be used to precisely identify individuals and determine their family relations. Initially, he used DNA profiling to help children who were denied citizenship because immigration officials did not believe that these children had British parents. When Jefferys went on to propose that DNA profiling could be used to solve crime, he was laughed at. He was vindicated when the local police involved in the Ashworth case approached him to prove that Buckland was the killer. On inspecting the blood and semen samples from the girls, Jefferys determined that they were raped by the same man and Buckland was innocent. A large scale DNA investigation followed this startling discovery. It spanned eight months and involved testing blood samples from 5,511 men and finally implicated Colin Pitchfork.

All humans share 99.9% of their DNA sequences. However, the remaining 0.1% is different enough to make DNA profiling possible. Cells from inside the cheek, known as a buccal sample, body fluids, such as blood, saliva, semen, or tissue samples can all serve as samples for DNA testing. Contrary to what is seen on crime shows, hair is not the best DNA sample. Hair fibers contain dead cells, which contain degraded DNA. Although hair roots contain DNA that can be analyzed, the probability of getting a good sequence is quite low, around 60-70%.

DNA profiling uses parts of the genome that are highly variable. These sequences are known as variable number tandem repeats or VNTRs. These are short sequences that are repeated multiple times and are positioned next to each other, like a string of identical pearls. These sequences are similar in closely related individuals but highly variable in people who are not. Since the pattern of VNTRs in an individual is inheritable, it can also be used for parental identification. An extension of this principle is DNA ancestry where your ancestors can be traced based on the VNTR patterns. Although there are several methods that can be used to detect VNTRs, the underlying principal is the same: sequence the VNTRs and compare the test samples to the reference. In a crime scene, for example, the reference would be the DNA found at the site and the test would be the DNA of the suspect.  

Figure 2: Differing lengths of the VNTRs in 6 different individuals. If the pattern of the reference and test samples are identical, it's a match. The two extreme lanes are DNA ladders that indicate the size of the fragments. Source.

DNA sequencing has helped solve an extensive list of crimes. In August 2018 it was used to finally identify the Golden State Killer. The accused, Joseph DeAngelo, had committed at least 13 murders and more than 50 rapes. DeAngelo was caught because the officials uploaded the DNA profile of the killer. The website GEDmatch, which maps family trees, found 10-20 distant relatives of the Golden State Killer which identified DeAngelo as the main suspect. A sample of his DNA was secretly collected from his garbage can and confirmed his connection to the crime. 

As with all scientific techniques, there are significant drawbacks to DNA profiling. The machines that are used to detect DNA are extremely sensitive and can pick up trace amounts of DNA, which is a problem because most crime scenes can be contaminated with DNA from several sources. Suppose 10 innocent people walked through a room prior to a crime and left small bits of their DNA behind, such as old skin flaking off. They would still be considered suspects. Although DNA evidence is seldom enough to convict someone, there is a rising worry that the mistakes in DNA profiling are overshadowed by how convincing their results are. To that end I'm reminded of a tenet in science: one piece of evidence is like a one legged stool, too shaky; two pieces is like a two legged stool, better but still wobbly; three independent pieces of evidence makes a milk stool, sturdy and reliable. Although DNA profiling alone is not enough, it helps build a stronger case in an investigation.