Forensic test can predict hair and eye colour from DNA

The test can predict both hair and eye colour from samples left at a crime scene

Scientists have developed a forensic test that can predict both the hair and eye colour of a possible suspect using DNA left at a crime scene.

The team that developed the test says it could provide valuable leads in cases where perpetrators cannot be identified through DNA profiling.

The Hirisplex system could allow investigators to narrow down a large group of possible suspects.

Details appear in the journal Forensic Science International: Genetics.

Predicting phenotypes - outward traits such as hair colour or eye colour - from DNA information is an emerging field in forensics.

An important current approach, known as genetic profiling, involves comparing crime scene DNA with that from a suspect or with a profile stored in a database.

But this relies on the person either being among a pool of suspects identified by the police or having their profile in a DNA database.

Tools such as Hirisplex could be useful in those cases where the perpetrator is completely unknown to the authorities, said Prof Manfred Kayser, who led the study.

He said the test "includes the 24 currently best eye and hair colour predictive DNA markers. In its design we took care that the test can cope with the challenges of forensic DNA analysis such as low amounts of material."

Prof Kayser, from Erasmus University Medical Center in Rotterdam, the Netherlands, added: "The test is very sensitive and produces complete results on even smaller DNA amounts than usually used for forensic DNA profiling."

He told BBC News that the journal article described everything needed to establish the test in a forensic lab, but that the team was also in touch with industry regarding their knowledge about hair and eye colour prediction.

The test system includes the six DNA markers previously used in a test for eye colour known as Irisplex, combining them with predictive markers for hair.

In the study, the authors used Hirisplex to predict hair colour phenotypes in a sample drawn from three European populations.

On average, their prediction accuracy was 69.5% for blonde hair, 78.5% for brown, 80% for red and 87.5% for black hair colour.

Analysis on worldwide DNA samples suggested the results were similar regardless of a person's geographic ancestry.

The team was also able to determine, with a prediction accuracy of about 86%, whether a brown-eyed, black haired person was of non-European versus European origin (excluding some nearby areas such as the Middle East).

The findings were also outlined at the sixth European Academy of Forensic Science conference in The Hague this week.

[e-mail]