Scale of tsunami tragedy complicates DNA matching

DNA testing is crucial to identifying many of the victims of the Asian tsunami, but the sheer numbers of dead make the task extremely difficult and risks false matches, say experts.

At least 150,000 people were killed after a magnitude 9.0 earthquake shook south east Asia on Sunday 26 December and spawned a huge tsunami. The enormous wave pounded coastal regions in 12 countries, with Indonesia, Sri Lanka, India and Thailand the worst affected.

With many thousands of people still missing, and many bodies decomposed beyond recognition, DNA analysis will be critical in returning the victims' bodies to their families. Hundreds of forensic scientists from countries including the UK, US and China have flown in to help with the effort.

"It's extremely challenging, and unprecedented in terms of mass disasters and mass fatalities," says detective sergeant Gary Locker, coordinator of the UK's Metropolitan Police's disaster victim identification team.

His team of about 40 officers is mainly in Thailand, recovering the bodies of UK tourists. Their colleagues in the UK are taking DNA samples for matching from relatives. The team will use fingerprinting and dental records, as well as DNA analysis, to identify victims.

Other teams have also been dispatched to reunite families with their dead. Disaster management company, Kenyon International, has sent 37 specialists to Thailand on behalf on an undisclosed client. "We are actively engaged around the clock in processing who has died," says Lowell Briggs, at the company's headquarters in Houston, Texas, US. The company is using special human remains identification software.

Perished together
To make an DNA identification, scientists must match samples from victims to those from relatives. But on this scale, and where several related people may have perished together, such matching could be problematic, says David Balding, a statistical geneticist at Imperial College, London, UK. 

DNA identification relies on matching repetitive sequences known as markers. In the UK, forensic scientists tend to use 10 of these markers, or loci, to reliably link a person to a crime scene sample. A direct match of a tsunami victim could be made with a DNA samples from their toothbrush or hairbrush, but these may not be available and can often be contaminated with DNA from other people.

Balding says 10 markers are also enough to be confident of a paternity test, as DNA is usually also available from the mother. "But if you only have one parent, I guess the standard 10 wouldn't be enough to be convincing," Balding told New Scientist.

However, matching a parent with a child, or vice-versa, will not be as difficult as other relationships. This is because a child inherits two copies of each marker - one from each parent. Therefore, a parent and child will have at least one of the markers in common for each of the 10 loci.

But this need not be the case when matching siblings. And for matching more distantly related relatives, far more markers would need to be tested for, he believes.

Coincidental similarities
Statistical geneticists, working to identify victims of the September 11 attacks in the US, found DNA matching using relatives problematic. Nearly 3000 victims had to be identified from 15,000 body parts. "With so large a set of victims, coincidental similarities between non-relatives abound", write US scientists Charles Brenner and Bruce Weir (Theoretical Population Biology vol 63, p 173). They found DNA samples from several relatives were needed to establish the identity of a victim.

Balding's own work, due to be published in Forensic Science International, suggests that as many as 50 DNA markers need to be tested to distinguish half-siblings from completely unrelated people. His team studied children born by donor insemination who wanted to find out if they had the same fathers.

Alec Jeffreys, at the University of Leicester, UK, who pioneered DNA profiling techniques, agrees that using 10 markers is not robust in this situation where two samples from the same person are not being matched. He believes using 15 to 20 markers might be better.

However, Jeffreys says the DNA tests used in conjunction with other information, such as location, clothing and distinguishing marks, will help identify individuals and reduce "the pool of false matches".

He also notes that there are no "major technical hurdles" in retrieving suitable DNA samples from victims in this disaster. In contrast, victims of the September 11 attacks were frequently very badly burned, while the bodies in mass graves in Bosnia had been buried for years.

"Even with a quite seriously decomposed corpse, you will be able to get DNA from it," he notes. Reasonable quality DNA can still be extracted from the soft tissues, including muscle, of tsunami victims he told New Scientist.