Researchers at a Dutch university have trained bees to detect coronavirus using a cheap and easily scalable model that could help fight the pandemic in underdeveloped countries.
Like other diseases, such as cancer, coronavirus causes metabolic changes in the human body that generate a unique smell. The team at Wageningen University – which specialises in life sciences research – has taught the bees to identify that smell.
The project used classic Pavlovian conditioning on an initial swarm of about 150 bees, so that each time they were exposed to the virus they were given a reward of sugar-flavoured water – and put their tongues out to collect the droplets.
By repeating that process, the bees came to associate Sars-CoV-2 – the virus that causes the disease known as Covid-19 – with the sugary reward and stuck out their tongues when the smell was present.
So extending the tongue – or more specifically in insect terms, the proboscis – is confirmation of a positive coronavirus result.
The scientists found that although both dogs and bees are known to share a highly developed sense of smell, the bees learned how to detect the virus after just minutes of training.
Even more remarkably, they were then able to identify the smell of coronavirus within seconds.
The testing is carried out on coronavirus swabs taken from an individual in the same way as currently, from the nose, the throat or in the form of saliva.
Each swab is then exposed in a container to five or six bees that are immobile – "in harness", as Wim Van der Poel, professor of Emerging and Zoonotic Viruses at Wageningen, puts it – and to someone trained to read their response, it is very clear.
“Certainly, this is basic by Western standards, but it works, it’s accurate, and it can handle large throughput in circumstances where options are limited”, Prof Van der Poel, a member of the Dutch government’s coronavirus Outbreak Management Team, told The Irish Times.
Prof Van der Poel was the lead researcher last year examining outbreaks of Covid-19 on 69 Dutch mink farms and possible links to humans, and it was that work that led to testing the detection capacity of the bees on healthy and infected mink first.
Crucially, there were very few false positives or false negatives in those initial tests. Subsequent tests using humans were equally encouraging.
In addition, there’s no danger that the bees will spread the virus. “They are not sensitive to it, and, more than that, they never come into direct contact with it during the training.”
Ultimately, the Wageningen team is focusing on the development of a biochip it hopes will use machine learning combined with genes from insect odour receptors to detect the virus.
