UK scientists build dog DNA database to help police solve crimes

Prof Denise Syndercombe Court, a forensics expert at King's College London who is leading the work, said canine traces are often not fully investigated. She gave the example of a human case such as a kidnapping. 'Lots of us have dogs. We live with dogs all the time. And so if you can identify canine material associated with a crime like that, then that can be very powerful in being able to provide intelligence to help solve a human case where there's no other evidence,' she said according to The Guardian.

The work faces practical limits. Syndercombe Court said forensic samples may contain only a few cells. 'You have a few cells that are left at a crime scene ... So you've got to make something useful out of that,' she said. One method focuses on short tandem repeats (STRs), small DNA segments whose repeat numbers can yield two figures for each STR, one inherited from each parent. Syndercombe Court said the set of numbers provides a virtually unique profile, for humans and dogs. In dogs, some numbers are more common in certain breeds.

By focusing on STRs, the team built a database covering 80 breeds. For the 10 most common breeds, including labradors, cocker spaniels and miniature schnauzers, samples were taken from at least 50 dogs of each breed. Canine DNA from a crime scene can be compared with DNA from a particular dog to confirm or rule out a link, or screened against the database to identify the breed of origin. The latter may be useful when the suspected dog is unknown, when a sample cannot be taken from a canine suspect, or when a sample does not match a dog under suspicion.

Syndercombe Court said the team hoped to expand the database to include breeds banned in the UK, such as pitbull terriers, American XL bullies and Japanese tosas. The team also created a database using mitochondrial DNA sequencing. Because cells contain many mitochondria, this approach can help when little material is available, though it is more limited because mitochondrial DNA is passed only from mother to offspring.

The approach has already provided information in police investigations. In one case, it helped clarify whether serious injuries a man suffered were caused by another person, as he claimed, or occurred when he tripped over his own dogs as they attacked the accused. Syndercombe Court said forensic analysis supported the latter. She said the accused's T-shirt had holes and DNA from two different dogs was found on the garment.

In another case, DNA analysis examined a sheep attack. Police thought the dogs belonged to a particular person, but DNA from those dogs did not match DNA recovered from the crime scene. 'So we were not only able to exclude his dogs, which were of a particular breed, but also say the breed of dog that [did] this attack was very likely to be [another breed],' Syndercombe Court said.

Dr Nicholas Dawnay, a forensics expert at Liverpool John Moores University who was not involved in the work, said research into how best to use canine DNA from crime scenes was hugely important. 'Whether it is a dog attacking a human or livestock or wildlife species, the ability to match the individual dog back to the attack is the primary goal. Obtaining breed information at the early stages of an investigation may help narrow down the pool of possible dogs, which is exciting,' he said. Dawnay said a major challenge remained funding canine forensic casework. 'Police budgets are stretched and without effective funding, the benefits of this research may struggle to be realised,' he said.