Heather Dewey-Hagborg, a doctoral student at Rensselaer Polytechnic Institute has created 3D software that produces images of people's faces from DNA samples. Then she takes it a step further and uses a 3D printer to create a sculpture of what that person probably looks like.
Dewey-Hagborg came up with the idea during a therapy session. She was staring at the art around the room when she noticed a single hair trapped within a piece of cracked glass.
"I just became obsessed with thinking about whose hair that was and what they might look like, and what they might be like," she says.
The thought stayed with her on the subway ride home. She noticed cigarette butts, discarded chewing gum and other DNA laden objects strewn around. In a move that will make many people cringe, she started collecting them and bringing them back to her lab for analysis.
Once in the lab she amplified regions of the DNA using a technique called Polymerase Chain Reaction (PCR).
"I send the results of my PCR reactions off to a lab for sequencing and what I get back are basically text files filled with sequences of As, Ts, Cs, and Gs, the nucleotides that comprise DNA. I align these using a bioinformatics program and determine what allele is present for a particular SNP on each sample," Dewey-Hagborg says. "Then I feed this information into a custom computer program I wrote that takes all these values which code for physical genetic traits and parameterize a 3d model of a face to represent them."
Dewey-Hagborg has used her process to capture traits like gender, ancestry, eye color, hair color, freckles, lighter or darker skin, and certain facial features like nose width and the distance between a person's eyes.
Just how accurate are the DNA-based sculptures? Dewey-Hagborg says they have a passing "family resemblance" to the person. "It's important to note that this is a work in progress. [The sculptures] will have similar traits and ancestry, but might look more like a possible cousin than a spitting image of the person themselves. The reason for this is multifold, but the primary reason is the research on facial morphology, the way human faces differ, is still in very early stages."