Image analysis for the automatic phenotyping of Orobanche cumana tubercles on sunflower roots

Abstract

Background The parasitic plant Orobanche cumana is one of the most important threats to sunflower crops in Europe. Resistant sunflower varieties have been developed, but new O. cumana races have evolved and have overcome introgressed resistance genes, leading to the recurrent need for new resistance methods. Screening for resistance requires the phenotyping of thousands of sunflower plants to various O. cumana races. Most phenotyping experiments have been performed in fields at the later stage of the interaction, requiring time and space. A rapid phenotyping screening method under controlled conditions would need less space and would allow screening for resistance of many sunflower genotypes. Our study proposes a phenotyping tool for the sunflower/O. cumana interaction under controlled conditions through image analysis for broomrape tubercle analysis at early stages of the interaction. Results We optimized the phenotyping of sunflower/O. cumana interactions by using rhizotrons (transparent Plexiglas boxes) in a growth chamber to control culture conditions and Orobanche inoculum. We used a Raspberry Pi computer with a picamera for acquiring images of inoculated sunflower roots 3 weeks post inoculation. We set up a macro using ImageJ free software for the automatic counting of the number of tubercles. This phenotyping tool was named RhizOSun. We evaluated five sunflower genotypes inoculated with two O. cumana races and showed that automatic counting of the number of tubercles using RhizOSun was highly correlated with manual time-consuming counting and could be efficiently used for screening sunflower genotypes at the tubercle stage. Conclusion This method is rapid, accurate and low-cost. It allows rapid imaging of numerous rhizotrons over time, and it enables image tracking of all the data with time kinetics. This paves the way toward automatization of phenotyping in rhizotrons that could be used for other root phenotyping, such as symbiotic nodules on legumes.

Wild Helianthus species: A reservoir of resistance genes for sustainable pyramidal resistance to broomrape in sunflower

“…Plants were grown in rhizotrons as in Le Ru et al. (2021) , with some modifications adapted to the slow growth of wild Helianthus species.…”

“…Compatible and incompatible attachments were counted at 14 dai using a binocular (Leica S6D). At 21 and 28 dai, pictures of the whole root system were taken using a Raspberry/picamera as described in Le Ru et al. (2021) and the final number of tubercles and necrotic tubercles were counted using ImageJ software (version 1.52a; Schneider et al., 2012 ) and the magic wand tool.…”

Development and characterization of a new sunflower source of resistance to race G of Orobanche cumana Wallr. derived from Helianthus anomalus

“…Two separate experiments were conducted, in both cases using six rhizotrons of each of the sunflower genotypes ANOM1 and the susceptible control B117 (Martín-Sanz et al 2016). The rhizotrons consisted of two plates of Plexiglas 12 × 12 cm and were prepared following Louarn et al (2016) and Le Ru et al (2021) procedures with slight modifications. Sunflower seeds were sterilized in 4.8% sodium hypochlorite for 10 min and rinsed with sterile distilled water thrice.…”

Managing Root Parasitic Weeds to Facilitate Legume Reintroduction into Mediterranean Rain-Fed Farming Systems

“…An often neglected limitation is the availability of fast but sufficiently reliable screening techniques that allow us to find what we need [87]. Hyperspectral imaging is being adjusted for early detection of broomrape infection to help with precise herbicide application in terms of time and space [169], with continuous attempts also to automate image phenotyping in seedling responses in rhizotrons [170], which is still too laborious and time-consuming. As a result, sources of resistance to broomrape are limited and poorly characterized.…”