Detection and diagnosis of seed-borne diseases utilising machine learning enhanced gas plasma integrated multispectral imaging (DeTecSeeD)

The rapid emergence and global spreading of pathogens, fungi and bacteria causing diseases, with crop seeds as vehicles (seed-borne diseases) are a threat to agricultural sustainability and food security. Food-borne disease is an ever-present threat and often associated with the consumption of fresh food such as horticultural products. A research study conducted by the UK's Food Standards Agency (FSA) entitled 'The Burden of Food-borne Disease in the UK 2018' identified an estimate of 2.4 million foodborne disease cases per year, the total burden from food-borne illness is ~£9bn in the UK. Food-borne diseases due to pathogen contamination can be disastrous and can cause 'pain, grief and suffering' for the consumers due to food-borne related illness, chronic disability and fatalities. An example for this is the 2011 outbreak of thousands of infections (>50 deaths) with enterohemorrhagic Escherichia coli (EHEC) caused by imported horticultural seeds used for sprout cultivation. Vegetable crops are frequently infected with seed-borne fungal pathogens. In such cases the fungi or bacteria are already present inside the seed or on the outer seed surface, which reduces seed quality and can cause seed rot and seedling damping-off during early crop establishment. Seed treatment with fungicides used to control seed-borne fungal pathogens of vegetable crops, but the most used fungicide thiram is banned in the EU since 2019 by regulations. In 2017 the FAO (United Nation Food and Agriculture Organization) issued new UN guidelines and phytosanitary measures to make international trade with plants and seeds - vital to feed the world's population - safer. Seed-borne diseases are of particular concern since horticultural and agricultural crop seeds are used for seedling raising (e.g. lettuce) or direct sowing (e.g. cabbage) and thereby any pests they carry could establish themselves and spread during crop production. A critical issue with regard to crop seed health is that the detection and diagnosis of seed-borne pathogens by conventional methods (e.g. pathogen isolation and growth analysis or molecular identification by DNA barcoding) are laborious and time-consuming processes which require expert knowledge. Novel detection and diagnosis technology platforms for seed-borne pathogens which are rapid, sensitive, reliable and efficient is therefore urgently needed in seed industry, at official seed testing stations, and for the international trade with commercial crop seed at borders. The DeTecSeeD (Detection Technology for Seed-borne Diseases) TRDF project consortium will explore the development of a new technology for the detection and diagnosis of seed-borne diseases. Prof Gerhard Leubner, Head of Royal Holloway University of London's (RHUL) Seed Science and Technology Group, Dr Tina Steinbrecher, RHUL's expert in seed biophysics, and Dr Felipe Iza from Loughborough University (LU) Wolfson School of Mechanical, Electrical and Manufacturing Engineering, have established an interdisciplinary team to explore if machine learning enhanced gas plasma integrated multispectral imaging may provide a novel detection and diagnosis of seed-borne diseases. The approach is interdisciplinary and the expected impact on the research community wide-reaching. Successful proof-of-concept will enable future work into translating the findings for building easy-to-use devices with wide applicability.

lThe DeTecSeeD (Detection Technology for Seed-borne Diseases) TRDF project consortium will explore the development of a novel and innovative transformative technology for the detection and diagnosis of seed-borne diseases. To investigate this, Prof Gerhard Leubner, Head of Royal Holloway University of London's (RHUL) Seed Science and Technology Group, Dr Tina Steinbrecher (RHUL's expert in seed biophysics) and Dr Felipe Iza from Loughborough University (LU) Wolfson School of Mechanical, Electrical and Manufacturing Engineering, will combine Low-Temperature Gas Plasma, MultiSpectral Imaging, Machine Learning and Viability Staining Assays into a rapid, sensitive and effective detection and diagnosis methods for pathogens in crop seeds. The approach is interdisciplinary and the expected impact on the research community wide-reaching; post-project, it will enable future work into translation and building devices. Depending on the Low-Temperature Gas Plasma treatment conditions, the plasma can etch the seed coat, thereby exposing organisms that were not just sitting on the seed surface. Furthermore, Machine Learning algorithms and Viability Staining Assays will be applied to investigate if this combination is suitable not only to detect pathogens but also to differentiate between dead and alive pathogens. Novel rapid, sensitive and effective detection and diagnosis methods of seed-borne pathogens would be beneficial including for International Seed Testing Association-certified laboratories (the standard in the seed industry) and for any research into novel environmental-friendly technologies for crop seed disinfection. Finally, the potential to miniaturise the technology, simplify the wavelengths and algorithms used and provide a concept for an integrated instrument that is easy to use and widely applicable, creating the basis for its implementation and commercialisation in future work.