Disease Detection

We work on developing machine learning methods to detect plant diseases before visible symptoms emerge. We use hyperspectral imaging to detect changes in plant reflectance that are indicative of pathogen infection. We also use Nanopore sequencing to detect pathogen DNA in plant samples to determine whether plant has been infected or not.

Culture-free pathogen detection using nanopore sequencing. Nanopore sequencing is an emerging technology and Oxford MinION platform allows sequencing experiments to be performed in field conditions. We use both inoculated samples and field samples to develop methods for real-time pathogen detection using whole genome, meta-genomic sequencing. The advantage is that our method does not require culture or amplification of target pathogen DNA. We have developed methods for detecting tomato pathogens including Pseudomonas and Xanthomonas species. We are also testing our approach for boxwood blight pathogen and grapevine pathogen, Xylella fastidiosa.

Machine learning and feature selection spectral data analysis. Using machine learning and features selection, we can identify characteristic spectral bands related to plant health status. More details coming soon…

DON content screen with hyperspectral imaging. More details coming soon…