Breakthrough In Nematode Identification With AI

To effectively and sustainably control harmful nematodes in agriculture, it's first necessary to determine which species you're dealing with. Identifying nematode species is complex, costly, and requires specialized expertise that is in short supply worldwide. Researchers at Wageningen University & Research (WUR) are collaborating on the development of an AI nematode identification system. A first milestone is the successful use of a microscope to independently identify the root knot nematode Meloidogyne chitwoodi . In tests, the system matched the performance of an experienced taxonomic nematologist.

Dig into the ground and chances are you'll encounter dozens, if not hundreds, of nematode species. These thread-like worms, measuring between 0.2 and 3 millimeters in length, are found almost everywhere in the soil – even on agricultural land. While some nematode species have a positive effect on the soil, there are also species that farmers prefer not to have in their fields, says researcher Leendert Molendijk. "When harmful nematodes are present in the soil, such as stem nematodes or root-knot nematodes, it can mean that you're no longer allowed to export bulbs, onions, and seed potatoes, for example. Furthermore, crops can develop deformities, making them difficult or impossible to market." Harmful nematodes are estimated to cause tens of billions of euros in damage to crops worldwide annually. Researchers estimate that 10% of agricultural production is affected.

Importance of identification
With Integrated Nematode Management—controlling harmful nematode species with minimal nematicides—it's crucial to know exactly which species you're dealing with, says Molendijk. "You don't want to control harmless and beneficial nematodes. Moreover, the effectiveness of measures varies by species. Some decrease after rotation with other crops or green manures, but there are also species that thrive on all kinds of crops. The timing of cultivation can also help eliminate certain species. Sometimes drastic measures are necessary, such as inundation, flooding a field."

Very specialized work
Identifying nematode species requires a great deal of expertise, explains researcher Pella Brinkman. "The differences between species are often minimal. It's really a matter of details: the shape of the spines, the length of a transparent part of the tail tip, or the number of head rings. Moreover, you often only know which species you're dealing with when the nematode is fully grown. Identification is usually done manually in specialized laboratories using microscopes, but sometimes additional molecular analysis is necessary. Worldwide, there are only a few dozen labs with the right resources and expertise, including Wageningen University & Research. It's an expensive and specialized field."

Challenging nematode species
One of the most challenging nematode species to identify is Meloidogyne chitwoodi , also known as the corn root knot nematode. Molendijk: “It requires a lot of knowledge and identification work to be able to say with certainty whether it is this nematode species. There are particularly strong similarities with Meloidogyne fallax —also known as the corn root knot nematode. When agritech company Veridi Technologies approached us and asked if we wanted to collaborate on an AI identification system, Meloidogyne chitwoodi seemed like a logical species to start with. If it works for such a complex species, it should also be possible for less difficult to distinguish species.”

AI system for species identification
Brinkman explains the input WUR provided for the AI ​​system. “We provided, among other things, the nematodes from our culture. We also identified the nematodes, the annotation. This is essential for training the model on large numbers of images. Furthermore, we performed validation, a reliability check. We did this using field samples that we knew contained high numbers of root knot nematodes. When the system made errors, the team helped analyze possible causes and advised the AI ​​on the morphological characteristics it should focus on to increase identification accuracy.”

High accuracy
Tests using Veridi's AI-powered microscope (Nemascope) demonstrated 96 percent accuracy for identifying Meloidogyne chitwoodi . This research demonstrates that AI can become a useful tool for nematode identification. Molendijk: "We've been working on technical possibilities for more efficient identification since the 1990s. The fact that this is now possible with AI is a significant milestone. If we can also apply this to other nematode species, it could have a significant impact worldwide, especially in regions lacking taxonomic knowledge and advisory services. If we give farmers worldwide access to an affordable identification system, they will have a better understanding of soil health. And that, in turn, can lead to improved crop yields and less reliance on nematicides."

Soil Monitoring Act
Within the framework of a grant from the European Innovation Council, Veridi Technologies and WUR are building on this initial success to conduct additional tests and research to expand the Nemascope's capabilities to include non-parasitic, free-living nematodes. Besides parasites, nematodes are also good indicators of soil biodiversity. This topic is of growing importance due to the recently adopted European Soil Monitoring Act.