Training bees to detect diabetes

LOS ALAMOS, N.M. (KRQE) – A New Mexico researcher is helping train honey bees to detect a deadly disease. It could be a new, low-cost way for developing countries to catch the disease early.

A group of foreign graduate students called Bee Healthy teamed up with a Los Alamos National Lab researcher to test whether bees can in fact detect diabetes.

Dr. Robert Wingo said he started working with bees about 10 years ago, training them like search dogs to detect explosives.

About a month ago, students called him up wondering if the busy bees could also help detect diseases like diabetes.

“At first, I was like, ehhh I don’t know…” Dr. Wingo said.

However, after looking into it further, he agreed to give it a try and help the five students competing for the Hult Prize. The winner gets $1 million to start up their project.

“They lined it up, and I flew out there with several boxes of gear. I landed. We grabbed the gear and went out to a beekeeper’s place, collected bees and started training bees,” he said.

In Boston, they put the bees in harnesses and trained them to stick their tongues out when they smelled high levels of acetone.

People with diabetes often have higher concentrations of acetone in their breath. The bees got the smell, then a reward of sugar water to lick. They did that over and over until eventually the bees just stuck out their tongues at the smell.

It worked with people, too.

“We believe there is sufficient evidence to merit looking at this more,” Dr. Wingo said.

The sweet promise of success, but not without a few stings along the way.

“It hurts every time. It does. It hurts every single time,” Dr. Wingo said.

The team of students he helped is hoping bees will become a self-sustaining tool for detection in third-world countries.

People can sell the honey and wax from the beehives and use that money to finance the diabetes detection program.

The team is one of six finalists. The winner will be announced next month in New York.

blog comments powered by Disqus