Wheat, one of the most important crops worldwide, is imperative to global food security. Take-all root disease is caused by an ascomycete fungus, Gaeumannomyces tritici (Gt), which penetrates the roots and destroys the vasculature. The resultant black necrotic lesions disrupt nutrient and water uptake and cause yield losses of up to 60%. Other closely related Gaeumannomyces species, such as G. hyphopodioides (Gh), are proposed biological control agents against take-all. G. hyphopodioides is able to invade the outer cortical root layer but, unlike Take-all, it does not extensively penetrate the stele and is characterised by grey runner hyphae on the root surface and sub-epidermal vesicles. My project focussed on assessing the ability of advanced imaging technology to accurately detect and quantify fungal root colonisation and to assess the impact on above ground health by measuring photosynthetic parameters.
A 5-week time-course seedling pot bioassay was set up using 250g of soil from Webb’s Field on the Rothamsted Farm, UK. An artificial inoculum layer derived from PDA plates contained Gt, Gh or an uncolonised control. The pots were baited with 10 seeds of the wheat cultivar ‘Hereward’, and kept in a controlled environment room. Each week, the pots were visually assessed for the presence of necrotic lesions, greying and vesicles. The level of infection was recorded as the proportion of roots showing at least one area of fungal colonisation, making this a semi-quantitative method of assessment. Potentially, imaging could provide a more efficient method to assess root colonisation, allowing the affected root area to be accurately quantified and also removing the subjectivity that can be associated with manual scoring. Consequently, a key question was: can multispectral imaging be used to quantitatively assess and distinguish between Gt and Gh seedling root colonisation?
Multispectral imaging was accomplished using a ‘VideometerLab’ imager, a system that uses 19 different wavelengths ranging from UV to the NIR. Training images were taken of healthy roots, and those colonised with Gt or Gh. These training images were used to build statistical transformations, which were subsequently used to score pixels as corresponding to healthy or colonised root tissue. The threshold pixel score chosen ensured that, in the case of Take-all, only the dark black lesions were scored. The ‘VideometerLab’ and visual assessments of Take-all colonised roots showed a significant correlation (Spearman’s Rank, Rs = 0.670, p = <0.001, n =24). Interestingly, whilst low, the Take-all scores for the uninoculated, control plants, were consistently higher using the image-based scoring. It is possible that the ‘VideometerLab’ was detecting other fungal colonisation phenotypes caused by contamination in the soil which were not visible to the human eye. The ‘VideometerLab’ struggled to distinguish the root greying caused by Gh from mild Take-all symptoms and preliminary analysis has shown that the root greying phenotypes associated with the presence of both fungal species have similar pixel scores across all wavelengths. However, the dark necrotic Take-all lesions had a distinct spectral signature, supporting the validity of our method for scoring Take-all.
Prior to harvesting the roots for the imaging described above, a LemnaTec ‘PhenoCenter’ was used to evaluate the impact of fungal root colonisation on above ground plant health. Specifically, a PAM fluorescence camera was used to evaluate photosynthetic efficiency in the foliage. This technique is based on the observation that light energy absorbed by photosystem II can be dissipated via three routes: photochemical reactions, dissipated heat or as fluorescence – it’s the competition between these processes which is exploited to quantify photosynthetic parameters. While a small reduction in photosynthetic efficiency was observed for the Take-all infected plants, no significant difference was observed between the Gh and control pots. These are encouraging results as it would be undesirable for Gh to impact photosynthetic efficiency if used as a biocontrol agent.
I would like to thank BSPP who have sponsored my summer bursary project which has given me the knowledge and the confidence to continue my studies in food security after graduating. I have had an amazing opportunity to increase my practical laboratory experience whilst developing skills which will be invaluable to me in the future namely; communication, problem-solving and statistical analysis. For all of this, I would like to thank Rothamsted Research and specifically Vanessa McMillan, Tom Ashfield and Tania Chancellor who have given up so much of their time to supervise me. I would also like to acknowledge Crop Health and Protection, an InnovateUK-supported organisation, in whose phenotyping facility the imaging was conducted.
Niamh Kavanagh
University of Leeds
Figure shows a Take-all colonised wheat seedling. On the left is the raw image with red arrows highlighting take-all lesions which would be scored manually. On the right the same image analysed using the ‘VideometerLab’ shows the classification of root tissue into diseased (blue) and healthy (orange/yellow) areas.
