Physical evidence left at a crime scene is often used to match a suspect’s DNA to the evidence, placing that suspect at the crime scene. But what if there is no suspect, and the DNA evidence yields no “hits” or matches in CODIS (Combined DNA Index System)? Investigators usually catalog the DNA evidence and explore other clues to the crime. However, a radical new study of genes responsible for skin pigmentation disorders offers new possibilities for predicting physical characteristics from crime scene evidence that may help criminal investigations.
The US Justice Department funded a 5-year, $680,000 study at the University of Arizona with such forensic science applications in mind. The researchers took their work on the genetic disorders that affect people’s eye, hair and skin colors (such as albinism) a step further to devise a way to predict these physical characteristics. Based on variations in genetic sequences in DNA samples alone, law enforcement agencies could outline a broad physical description of a suspect based on the DNA evidence left behind at a crime scene.
The UA research team, lead by Murray Brilliant, UA professor at the Steele Children’s Research Center, explained the methodology. They measured the skin tone, hair color and eye color of 1,000 UA students, a diverse sample in terms of race and ethnicity. They collected the students’ DNA with a buccal swab, and then examined the genes known to affect a person’s appearance, in terms of pigment. Using statistical analyses, the researchers noted connections and observed patterns between the students’ genetic data and appearances.
They found that three or four out of the ~30,000 total human genes control much of a person’s skin, hair, and eye pigmentation. The team was able to identify matches to these pigment-related genes by comparing the students’ features to these three main characteristics, comprising hundreds of eye colors and shades, the 11 identified dermatological skin tones, and precise melanin levels in hair. This methodology allowed researchers to predict someone’s hair color with about 80% accuracy, eye color to about 75% accuracy, and skin color to about 50% accuracy—percentages they say are relatively accurate in light of the broad range of color for each of these attributes.
Consequently, if an unknown sample found at a crime scene is analyzed but a DNA match is not identified, the UA team’s methods could be used as an investigative tool to narrow the pool of suspects. According to UA’s Murray Brilliant, “it’s not 100 percent, but 80% is more accurate than an eyewitness.”