Under the ground hides something invisible to the naked eye, something so small that it only takes a microscope to see it: nematodes.
Tiny organisms are ubiquitous in Earth’s ecosystems and are found on the highest mountains to the darkest corners of the oceans. Often called roundworms, nematodes vastly outnumber any other organisms on Earth. Some nematode species, however, can cause problems when present in excessive numbers.
Researchers at the Virginia Tech College of Agriculture and Life Sciences are studying how to use artificial intelligence to identify the presence of soybean plant parasitic nematodes that cause more than $1 billion in soybean losses each year in the United States.
As part of the research, a robust microscopic dataset of soybean nematodes will be developed along with artificial intelligence algorithms for identification and geographic information system infestation heatmaps. The result of the research is to bring all this together in an executable tool for producers.
“By implementing an executable tool to control soybean nematodes, we are able to save growers time and money while improving environmental sustainability,” said David Langston, professor at Tidewater Agricultural Research. and Extension Center and researcher on the project. “We want this research to be easy to implement and to be a practical and sustainable way to control nematode infestations.”
Visual symptoms of soybean nematodes can be confused with other crop stressors, leading to significant crop loss when left untreated. This research will allow rapid identification for nematode management decisions while reducing the number of pesticides used on crops.
These pesticides are often non-selective, which means that they kill both good and bad nematodes in the same way that antibiotics target both good and bad bacteria in the human body. These non-selective pesticides kill beneficial nematodes that are essential to the carbon cycle, which is the process by which carbon is exchanged through the Earth’s biosphere, geosphere, pedosphere, hydrosphere and atmosphere.
There is no cure for nematodes during the growing season. Nematode control tactics are implemented after harvesting the current crop and before or during planting of the upcoming crop.
The big problem is that most nematode samples are taken at or after harvest in the fall and the bottleneck created by using available methods to assess nematode populations means that results may be available. too late for growers to order seeds of nematode-resistant varieties for the next season. season. Nematicides are applied just before planting or at planting time.
With fewer people having the expertise to identify nematodes, like the root-knot nematode Meloidogyne incognita and the soybean cyst nematode Heterodera glycines, the two nematodes that researchers are analyzing, systems need to be in place to improve sample processing. , which is already a bottleneck. It takes considerable time to manually scan and identify nematodes in samples sent to the lab, the researchers said.
During the peak growing season, this can mean delays in identification and growers miss the window to make decisions. By removing the human element and training the algorithms, appropriate treatment prescriptions can be assigned almost at the same time samples are received.
“It’s a problem we’re trying to solve on two fronts,” said Abhilash Chandel, assistant professor in the Department of Biological Systems Engineering. “The increased capacity of this manual counting and identification is a task. The other is to resolve ground-level issues, which includes nematode infestation levels using large-scale satellite images or drone imagery.
Using images to discern causal factors will be a challenge, the researchers said. This will require images and algorithms to determine the difference between altitude, soil types, nutrient leaching, pH levels, etc.
Current best practices for nematode identification in the United States still invoke the timeless classical microscope to manually count and identify nematodes in each sample. This process needs to improve, the researchers said.
During the 2022 growing season, nematodes were sampled from farm hot spots in Virginia using 20 geotagged samples per acre of field. Over 40,000 microscopic images were analyzed using a Motic compound microscope and a Canon Vixia HF G20. Christian Pittman treats, identifies and counts these nematodes.
Mychele Batista da Silva, a postdoctoral researcher working with Langston, is a nematologist who performs differential host bioassays, which are analytical methods to measure the concentration or potency of a substance through its effects on a living system, and uses molecular tools to identify nematode species. and breeds that cannot be identified visually.
From these images, artificial intelligence algorithms were created for the identification of soybean nematodes. Convolutional, a class of neural networks commonly applied to imaging, Recurrent, a class of neural networks that typically creates a cycle, and artificial neural networks, inspired by the biological neural networks that make up brains, will be trained, validated and refined throughout the research process.
Once the research project is complete, an executable tool will be made available to producers along with a secure method of sharing results remotely.
This article was written by Max Esterhuizen of Virginia Tech, a communications and marketing specialist who has spent his professional career in higher education and sports journalism.