Faces in a Crowd: New DNA Technique Can Pick One Person’s DNA Signature Out of Hundreds

The opening bars of “Who are You” crash through the speakers, the credits roll and the lead actors of “CSI” pull up at the crime scene to investigate yet another murder.

Since CSI” premiered in 2000, the show, its spin-offs and imitators have hammered home the idea that a person’s genetic material can readily establish his or her presence at a crime scene. The programs’ popularity has even led to the “CSI effect,” where real juries have come to expect that such information is part of the evidence provided at all trials.

But reality is rarely as simple as TV. According to the President’s DNA Initiative, DNA labs across the country saw a 73% increase in their casework from 1997 to 2000, while their backlog was nearly double that figure. One contributor to lab backlogs: samples containing more than one source of DNA. Picking out each individual from such mixed samples can be extraordinarily costly and time-consuming.

Researchers from the Translational Genomics Research Institute (TGen) and the University of California at Los Angeles have developed a new way to detect whether a specific individual’s DNA is part of a mixed sample, even when that person’s genetic material makes up as little as one one-thousandth of the total. As detailed in PLoS Genetics, this method developed by Nils Homer and his colleagues could help forensic investigators determine just who was at a crime scene.

“By employing the powers of genomic technology, it is now possible to know with near certainty that a particular individual was at a particular location,” TGen researcher and study author David Craig said in a statement. “Even with only trace amounts of DNA and even if dozens or even hundreds of others were there, too.”

The researchers used high-density SNP arrays, much like those used by 23andMe, to genotype a specific individual, a DNA mixture, and a reference DNA sample. They then used statistical methods to compare the individual to both the DNA mixture under investigation and the reference sample. These statistics allowed them to determine if a person had in fact contributed to the DNA mixture.

The researchers demonstrated the utility of their method by testing it on DNA mixtures composed of anywhere from two to 200 people of Caucasian ethnicity, with each single person contributing as little as one-tenth of a percent of the total sample material.

“It opens up a whole new can of worms of what’s possible to do forensically,” said Stanley Nelson, a UCLA-based co-author of the study (and 23andMe advisor), in a statement. (23andMe was not associated with this research.

The ability to distinguish an individual’s DNA from a mixture also has implications for genome-wide association studies. Such projects require large quantities of genetic information, and genotyping the thousands of individuals participating is a costly process. When researchers try to make the data publicly available for use in other studies, they often pool the information in order to maintain each individual participant’s confidentiality. Homer and his colleagues argue that their technique for picking one individual’s DNA out of a crowd could strip away that pretense of anonymity.

“Our findings make it very clear that such an approach realistically does not conceal identity,” they wrote in their study, suggesting that researchers should confidentially share their data in its entirety instead.

Image from the National Forensic Science Technology Center