PheNeSens – Phenotyping of Nematodes with Sensors

In the project PheNeSens LemnaTec works together with Institute of Imaging and Computer Vision of the RWTH Aachen University and Institute of Plant Protection in Field Crops and Grass Land of the Julius Kühn Insitut on the development of phenotyping methods for plant pathogenic nematodes.

PheNeSens – Phaenotypisierung von Nematoden mit Sensoren

Im Projekt PheNeSens arbeitet LemnaTec gemeinsam mit dem Lehrstuhl für Bildverarbeitung der RWTH Aachen und dem Institut für Pflanzenschutz in Ackerbau und Grünland des Julius Kühn Insituts an der Entwicklung von Phänotypisierungsmethoden für pflanzenpathogene Nematoden.


Project work and achievements

We are developing methods for phenotyping cysts and free stages of nematodes. Taking the sugar beet cyst nematode – Heterodera schachtii – as model organism, we have established image acquisition methods and developed image processing procedures.

As cysts and free stages substantially differ in size, optical magnification was required in different magnitudes. While cysts can be imaged with a high-magnification lens mounted onto a camera, free stages require microscopic imaging.

Many nematodes cause damage in crops, like our model organism Heterodera schachtii, the sugar beet cyst nematode. Similar pest threaten potato plants, soybean, cereal, and vegetables. Their monitoring is important for the trade of agricultural goods, plant breeding, or plant protection activities.

Nowadays, monitoring still uses visual scoring and other manual practices so that it is time consuming and depends on experts. Thus, the influence of the human factor is still prominent. Developing automatic monitoring tools will speed up the process and makes it more reliable and reproducibel. Therefor our developments are an important contribution in improving nematode monitoring.

Nematode monitoring starts with extracting the nematodes from soil and plants. This includes washing and sieving steps.

At the end of this preparation, cysts together with debris are present on a filter paper. Our cyst imaging method stats with these filters and targets to find the cysts among the debris.

Breaking up the cysts in further working steps gives access to their content, the eggs and juvenils, i.e. the free stages of the nematodes. These stages can be imaged and analyzed with our microscopic imaging and image processing method. The method is applicable for cyst forming species to analyze cyst content, but can be similarly applied to nematodes that do not form cysts and thus always are present as free stages.

Cyst Phenotyping

For cyst imaging, we used a high magnification lens mounted in an LED illuminated imaging cabinet. Filter papers originating from cyst preparation can be photographed as image series so that the single images cover the whole area. They are stiched to a full image and processed with a deep-learning segmenting model.

Result of the segmenting is not only the count of the cysts per sample, but also dimensions of each detected cysts, e.g. length/width or projected area. With these data we can characterize the populations on phenotypic level. With achieving the dimensional data in one step together with the counting, the procedure is superior to traditional visual counting processes.

Egg, Juvenile, and Free Stage Phenotyping

In case of Heterodera schachtii, we apply this technology for the cyst content, i.e. the eggs and juveniles. Broader speaking, it is applicable to any free living stages of nematodes, regardless whether they originate from cysts or not.

For image acquisition of free stages we mounted an industrial camera onto a microscope and equipped this with an automatic sample stage. Thereby, samples are moved along the lens and image series are acquired. These series cover the full sample area on each microscopic slide.

Machine learning based segmentation models allow detecting eggs and juveniles. All detected free stages are counted and phenotyped, i.e. measured for length and width, or area. Similarly to cyst phenotyping, the data enable in-depth characterization of the population going far beyond visual scoring.

The Project Team

Hans Georg Luigs, Senior Engineer, LemnaTec

Dr.  Rüdiger Goetz, Software Specialist, LemnaTec

Dr. Marcus Jansen, Biologist, LemnaTec

Marcus Jansen

Dr. Matthias Daub, Julius Kühn Institut

Prof. Dr. Dorit Merhof, Lehrstuhl für Bildverarbeitung, RWTH Aachen

Long Chen, Research Scientist Electrical Engineering & Information Technology, PhD candidate, LfB RWTH Aachen

Dr. Martin Strauch, Post-Doctoral Researcher Bioinformatics, LfB RWTH Aachen