From: julianlzb87@gmail.com   
      
   Using AI, biologists have begun to decipher electrical distress signals   
   that are emitted when pests strike   
      
      
      
   Biologists have worked out how to “listen” to plants’ distress calls,   
   paving the way for farmers to be sent an alert when crops are attacked   
   by pests and diseases.   
      
   Scientists at Syngenta, an agricultural technology company in Basel,   
   Switzerland, for the first time have used machine learning to decode the   
   electrical signals emitted by plants.   
      
   And so far, the message that has come across most clearly is the   
   equivalent of a plant dialling 999. Patrik Hoegger, head of the   
   company’s insect control research group, said: “It’s the plant’s way of   
   saying, ‘Ouch, I’m being hurt.’”   
      
   Botanists have known for more than a century that plants transmit   
   electrophysiological signals between different parts of their structure.   
   But until now these signals, which travel through the ion currents in   
   tissue channels that transport water, minerals and nutrients throughout   
   a plant, have been meaningless to humans.   
      
   The boom in machine learning and artificial intelligence capability,   
   however, has transformed scientists’ ability to decipher and analyse   
   huge amounts of data.   
      
   Anke Buchholz, a plant scientist at Syngenta, said: “Ten years ago we   
   couldn’t have dreamed about doing something like this.”   
      
   By wiring up plants to electrical monitoring equipment in their huge   
   glasshouses outside Basel, Buchholz and her team worked out they could   
   start to process the data. And the plants were not keeping quiet. “A   
   human cannot handle this amount of information,” she said. “We are   
   getting 256 data points per second.”   
      
   When the plants had been wired up, the scientists experimented with   
   exposing them to different stressors. In one experiment, the results of   
   which are published in the Journal of Pest Science, they exposed tomato   
   plants to microscopic nematode worms in the soil. In a second   
   experiment, published in the Scientific Reports journal, they released a   
   swarm of stink bugs among soybean plants.   
      
   By monitoring the electrical output from the plants and running it   
   through their computer models, the team could spot the moment the plants   
   came under attack.   
      
   This could be a vital early warning signal, Hoegger said, because in   
   each of these cases it is almost impossible to detect pest damage until   
   it is too late.   
      
   “Nematodes are pests in the soil so you don’t see them,” he said. “You   
   only see the damage when you realise the yield is far lower than usual.”   
      
   Likewise, the damage caused by stink bugs — a pest which is common in   
   the soybean farms of south and central America — is usually only   
   detected about a week after they have attacked a plant, by which time it   
   is too late.   
      
   Soybeans are a crucial crop, accounting for roughly half of the world’s   
   plant-based protein. Yet 21 per cent of crops are lost each year due to   
   pests and pathogens.   
      
   Buchholz said use of the new technology could help lead to more   
   targeted, gentler pesticides. Current crop protection techniques take a   
   carpet-bombing strategy: farmers blitz their fields with chemicals to   
   combat any threat.   
      
   “At the moment we have a hard target,” Buchholz said. “The insects have   
   to be dead to make sure they are not harming the crops.”   
      
   But if farmers are able to tell from their monitoring equipment when a   
   crop comes under attack — and from what species of insect — it might be   
   sufficient to use far more specialised chemicals simply to deter the   
   pests, or at least only harm the ones which are causing a problem. “This   
   might give us a softer way of preventing insects from harming the   
   crops,” said Buchholz.   
      
   That is a goal for the future. At the moment the scientists can only   
   tell when a plant is under stress — not the particular pest that is   
   attacking it. But Buchholz said it was a “medium term” target to be able   
   to differentiate plants’ cries, telling apart the different insects,   
   diseases or fungal infections that afflict them.   
      
   “We would like to build up a dedicated library,” she said. “That would   
   be a very useful tool to have in our hands.”   
      
   Listening to plants is within scientists’ reach. Now, they just need to   
   work out how to reply.   
      
   https://www.thetimes.com/uk/science/article/scientists-plants-fa   
   ming-pesticides-2zrdm0qc0   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)