Evaluation of the effectiveness of antagonistic bacteria against Phytophthora blight disease in pepper with artificial intelligence

Abstract

Phytophthora blight of pepper, caused by Phytophthora capsici, is a disease seriously affects pepper production. The aims of this study were to evaluate the control of P. capsici by antagonistic bacteria in pot trials and to score disease severity by using artificial intelligence (AI). For this purpose, six P. capsici isolates used in the study were identified with PCR, and pathogenicity tests were performed. Biological control pot trials were conducted by testing three antagonistic bacteria. At the end of the pot trials, the disease severity in the plants was rated visually. An AI system was also created for the current study. The scores from the AI system were compared with the visual scale scores with different methods. All pathogen isolates were identified as P. capsici according to the PCR results. PcU5 was found to be the most aggressive isolate of the tested P. capsici isolates according to pathogenicity tests, with 45 mm lesion length. Pca17 was found to be an effective bacterial antagonist, with 77% efficiency in pot trials. The AI system, called PhytAi (https://phytai.online), significantly correctly scored the disease severity when compared with the visual scores (r2 = 0.96). PhytAi can be used not only by phytopathologists but also by any researcher working in plant sciences who needs visual classification. This is a general classification system that supports scale ratings; the determination of varieties, disease severity measurement, and all kinds of visual classification, identification, etc. can be easily performed without changes in the system thanks to AI.