Elucidation of turnip yellows virus (TuYV) spectral reflectance pattern in Nicotiana benthamiana by non-imaging sensor technology

Turnip yellows virus (TuYV), belonging to the genus Polerovirus of the family Solemoviridae, is an aphid transmissible, pathogenic virus causing considerable yield losses in oilseed rape (Brassica napus subsp. napus) cultivation. Virus detection in infected plants is difficult due to the phloem limitation and the irregular distribution of the virions within the host plant tissue. The spectral reflectance of leaves following TuYV inoculation in the model plant Nicotiana benthamiana was determined with the help of a handheld and portable non-imaging spectrometer under greenhouse conditions. Virus infection and a classification of virus content groups very low, low, medium and high were possible to be predicted with a mean average of 87% accuracy compared with the non-inoculated control plants when referenced to enzyme-linked immunosorbent assay (ELISA) data. Different spectral reflectance patterns (380-2500 nm) according to the specific virus content group of different leaf levels along the plant axis were mapped. Furthermore, machine learning allowed identifying the importance of specific areas in the spectral signature following TuYV infection. Between control and medium a specific peak within the visible area around 700 nm wavelength and between control and high the importance of the spectral area > 750 nm were identified. These data can serve as a basis to characterize metabolic or cytochemical changes after virus infection in host plants. Furthermore, spectral sensing devices enable non-destructive and informative dissection of leaf physiological properties and provide valuable information for monitoring virus spread in plants in greenhouse applications or on the field scale.

Automated summary

This study determined the spectral reflectance of leaves following TuYV inoculation in Nicotiana benthamiana, predicting virus infection and virus content groups using machine learning. Specific areas in the spectral signature were identified as important indicators of virus infection. These findings can help characterize metabolic or cytochemical changes after virus infection and provide valuable information for non-destructive monitoring of virus spread in plants.