Pathogens cost American soybean farmers $14.6 billion a year — about $33 in losses for every acre planted, before the chemicals used to fight them are counted. In an opinion piece for AgFunderNews, Kevin Fries, lead agronomic modeller at California biotech company InnerPlant, and Dr Christopher Seifert, its head of software product, argue that most new disease-detection technology fails for the same reason: it looks for symptoms, and symptoms come too late.

Fungal diseases such as white mould and frogeye leaf spot (FLS) are the main worry. Typical yield loss from FLS is now around 35% and can reach 60%, the authors write. A well-timed fungicide can cut FLS incidence by up to 50%, but only if it lands at the right moment — spraying before the disease is present does nothing, and spraying late may not contain the outbreak.

The economics are tight. A farmer needs as much as 7.5 extra bushels per acre to pay for one fungicide pass, yet the Iowa Soybean Association puts the gain from blanket applications at about one bushel per acre. Most growers, the authors say, still "spray and pray" during key growth windows: one survey found 65% of Missouri's soybean acres were sprayed, mostly for FLS, but only 10% were scouted first.

Scouting on foot rarely helps, they argue, because the black spots that reveal FLS mean the spray window has already closed; the pathogen begins spreading within 48 hours of becoming symptomatic. Drones with cameras give the same answer at greater cost, and most drone services fly only two or three times in the roughly 45-day period when fungicides are useful. AI models trained on photos of healthy and diseased leaves are faster and more accurate than people, but still see only visual cues that appear weeks after infection begins.

InnerPlant's answer — the authors acknowledge it is their own company's — is to turn the soybean plant into the sensor. When the plant's immune system responds to a pathogen, an engineered gene makes it fluoresce; viewed through a special camera, the plants "shine" as soon as infection starts, about ten days before visible symptoms. Sentinel plots of such plants could be monitored continuously and tell a farmer which fields, or parts of fields, actually need spraying.

They cite a USDA soybean-rust surveillance trial in which regular monitoring of sentinel plots, using visual inspection alone, delivered up to $299 million in annual savings for growers. Their own calculation is that pre-symptomatic alerts could lift total yield by at least six percentage points over intermittent visual scouting — a net benefit of nearly $40 per acre at average yields, with some farmers gaining as much as 25 bushels per acre against unmonitored fields.

Spraying too early or too late, they conclude, raises the odds of spraying again, adds input cost, puts more chemicals in the soil and speeds up fungicide resistance. Smarter scouting will be part of the answer, but only if it is paired with sensing that goes beyond what the eye or a camera can see. The views are the authors' own.

Source: AgFunderNews