Improving Common Bacterial Blight Phenotyping by Using Rub Inoculation and Machine Learning: Cheaper, Better, Faster,...

Area of Science:

• Plant Pathology
• Bacterial Blight
• Plant-Pathogen Interactions

Background:

• Common bacterial blight, caused by *Xanthomonas* species, significantly threatens common bean production.
• Pathogenicity varies among strains, influenced by type III secretion systems and transcription activator-like effectors (TALEs).
• Detecting subtle phenotype changes from single gene impacts requires precise methodologies.

Purpose of the Study:

• To compare two inoculation and symptom assessment methods for distinguishing *tal* mutants from wild-type strains.
• To identify a sensitive and efficient method for assessing disease progression in common beans.
• To develop a miniaturized pathogenicity test with time-saving benefits.

Main Methods:

• Comparison of rub inoculation on first leaves versus dip inoculation on first-trifoliate leaves.
• Symptom assessment using machine learning-based imaging versus chlorophyll fluorescence imaging.
• Evaluation of two *tal* mutants against their corresponding wild-type strains.

Main Results:

• Rub inoculation combined with machine learning-based imaging significantly distinguished wild-type strains from *tal* mutants.
• Dip inoculation with chlorophyll fluorescence imaging failed to differentiate the strains.
• The developed method enabled miniaturization and significant time savings in pathogenicity testing.

Conclusions:

• Machine learning-based imaging of rub-inoculated leaves is a superior method for distinguishing subtle differences in bacterial blight pathogenicity.
• This novel approach enhances the efficiency and scalability of plant disease assessment.
• The findings contribute to improved strategies for managing common bacterial blight in common beans.