Genome-Wide Association Study for Maize Leaf Cuticular Conductance Identifies Candidate Genes Involved in the Regulation of Cuticle Development.

The cuticle, a hydrophobic layer of cutin and waxes synthesized by plant epidermal cells, is the major barrier to water loss when stomata are closed at night and under water-limited conditions. Elucidating the genetic architecture of natural variation for leaf cuticular conductance (gc) is important for identifying genes relevant to improving crop productivity in drought-prone environments. To this end, we conducted a genome-wide association study of gc of adult leaves in a maize inbred association panel that was evaluated in four environments (Maricopa, AZ, and San Diego, CA, in 2016 and 2017). Five genomic regions significantly associated with gc were resolved to seven plausible candidate genes (ISTL1, two SEC14 homologs, cyclase-associated protein, a CER7 homolog, GDSL lipase, and ß-D-XYLOSIDASE 4). These candidates are potentially involved in cuticle biosynthesis, trafficking and deposition of cuticle lipids, cutin polymerization, and cell wall modification. Laser microdissection RNA sequencing revealed that all these candidate genes, with the exception of the CER7 homolog, were expressed in the zone of the expanding adult maize leaf where cuticle maturation occurs. With direct application to genetic improvement, moderately high average predictive abilities were observed for whole-genome prediction of gc in locations (0.46 and 0.45) and across all environments (0.52). The findings of this study provide novel insights into the genetic control of gc and have the potential to help breeders more effectively develop drought-tolerant maize for target environments.

Image set for deep learning: field images of maize annotated with disease symptoms.

OBJECTIVES: Automated detection and quantification of plant diseases would enable more rapid gains in plant breeding and faster scouting of farmers' fields. However, it is difficult for a simple algorithm to distinguish between the target disease and other sources of dead plant tissue in a typical field, especially given the many variations in lighting and orientation. Training a machine learning algorithm to accurately detect a given disease from images taken in the field requires a massive amount of human-generated training data. DATA DESCRIPTION: This data set contains images of maize (Zea mays L.) leaves taken in three ways: by a hand-held camera, with a camera mounted on a boom, and with a camera mounted on a small unmanned aircraft system (sUAS, commonly known as a drone). Lesions of northern leaf blight (NLB), a common foliar disease of maize, were annotated in each image by one of two human experts. The three data sets together contain 18,222 images annotated with 105,705 NLB lesions, making this the largest publicly available image set annotated for a single plant disease.

Quantitative trait locus mapping reveals complex genetic architecture of quantitative virulence in the wheat pathogen Zymoseptoria tritici.

We conducted a comprehensive analysis of virulence in the fungal wheat pathogen Zymoseptoria tritici using quantitative trait locus (QTL) mapping. High-throughput phenotyping based on automated image analysis allowed the measurement of pathogen virulence on a scale and with a precision that was not previously possible. Across two mapping populations encompassing more than 520 progeny, 540 710 pycnidia were counted and their sizes and grey values were measured. A significant correlation was found between pycnidia size and both spore size and number. Precise measurements of percentage leaf area covered by lesions provided a quantitative measure of host damage. Combining these large and accurate phenotypic datasets with a dense panel of restriction site-associated DNA sequencing (RADseq) genetic markers enabled us to genetically dissect pathogen virulence into components related to host damage and those related to pathogen reproduction. We showed that different components of virulence can be under separate genetic control. Large- and small-effect QTLs were identified for all traits, with some QTLs specific to mapping populations, cultivars and traits and other QTLs shared among traits within the same mapping population. We associated the presence of four accessory chromosomes with small, but significant, increases in several virulence traits, providing the first evidence for a meaningful function associated with accessory chromosomes in this organism. A large-effect QTL involved in host specialization was identified on chromosome 7, leading to the identification of candidate genes having a large effect on virulence.

Automated Identification of Northern Leaf Blight-Infected Maize Plants from Field Imagery Using Deep Learning.

Northern leaf blight (NLB) can cause severe yield loss in maize; however, scouting large areas to accurately diagnose the disease is time consuming and difficult. We demonstrate a system capable of automatically identifying NLB lesions in field-acquired images of maize plants with high reliability. This approach uses a computational pipeline of convolutional neural networks (CNNs) that addresses the challenges of limited data and the myriad irregularities that appear in images of field-grown plants. Several CNNs were trained to classify small regions of images as containing NLB lesions or not; their predictions were combined into separate heat maps, then fed into a final CNN trained to classify the entire image as containing diseased plants or not. The system achieved 96.7% accuracy on test set images not used in training. We suggest that such systems mounted on aerial- or ground-based vehicles can help in automated high-throughput plant phenotyping, precision breeding for disease resistance, and reduced pesticide use through targeted application across a variety of plant and disease categories.

An Improved Method for Measuring Quantitative Resistance to the Wheat Pathogen Zymoseptoria tritici Using High-Throughput Automated Image Analysis.

Zymoseptoria tritici causes Septoria tritici blotch (STB) on wheat. An improved method of quantifying STB symptoms was developed based on automated analysis of diseased leaf images made using a flatbed scanner. Naturally infected leaves (n = 949) sampled from fungicide-treated field plots comprising 39 wheat cultivars grown in Switzerland and 9 recombinant inbred lines (RIL) grown in Oregon were included in these analyses. Measures of quantitative resistance were percent leaf area covered by lesions, pycnidia size and gray value, and pycnidia density per leaf and lesion. These measures were obtained automatically with a batch-processing macro utilizing the image-processing software ImageJ. All phenotypes in both locations showed a continuous distribution, as expected for a quantitative trait. The trait distributions at both sites were largely overlapping even though the field and host environments were quite different. Cultivars and RILs could be assigned to two or more statistically different groups for each measured phenotype. Traditional visual assessments of field resistance were highly correlated with quantitative resistance measures based on image analysis for the Oregon RILs. These results show that automated image analysis provides a promising tool for assessing quantitative resistance to Z. tritici under field conditions.

Quantitative trait locus mapping of melanization in the plant pathogenic fungus Zymoseptoria tritici.

Melanin plays an important role in virulence and antimicrobial resistance in several fungal pathogens. The wheat pathogen Zymoseptoria tritici is important worldwide, but little is known about the genetic architecture of pathogenicity, including the production of melanin. Because melanin production can exhibit complex inheritance, we used quantitative trait locus (QTL) mapping in two crosses to identify the underlying genes. Restriction site-associated DNA sequencing was used to genotype 263 (cross 1) and 261 (cross 2) progeny at ~8500 single-nucleotide polymorphisms and construct two dense linkage maps. We measured gray values, representing degrees of melanization, for single-spore colonies growing on Petri dishes by using a novel image-processing approach that enabled high-throughput phenotyping. Because melanin production can be affected by stress, each offspring was grown in two stressful environments and one control environment. We detected six significant QTL in cross 1 and nine in cross 2, with three QTL shared between the crosses. Different QTL were identified in different environments and at different colony ages. By obtaining complete genome sequences for the four parents and analyzing sequence variation in the QTL confidence intervals, we identified 16 candidate genes likely to affect melanization. One of these candidates was PKS1, a polyketide synthase gene known to play a role in the synthesis of dihydroxynaphthalene melanin. Three candidate quantitative trait nucleotides were identified in PKS1. Many of the other candidate genes were not previously associated with melanization.

Measuring quantitative virulence in the wheat pathogen Zymoseptoria tritici using high-throughput automated image analysis.

Zymoseptoria tritici, causal agent of Septoria tritici blotch on wheat, produces pycnidia in chlorotic and necrotic lesions on infected leaves. A high-throughput phenotyping method was developed based on automated digital image analysis that accurately measures the percentage of leaf area covered by lesions (PLACL) as well as pycnidia size and number. A seedling inoculation assay was conducted using 361 Z. tritici isolates originating from a controlled cross and two different winter wheat cultivars. Pycnidia size and density were found to be quantitative traits that showed a continuous distribution in the progeny. There was a weak correlation between pycnidia density and size (r = -0.27) and between pycnidia density and PLACL (r = 0.37). There were significant differences in PLACL and pycnidia density on resistant and susceptible cultivars. In all, >20% of the offspring exhibited significantly different pycnidia density on the two cultivars, consistent with host specialization. Automated image analysis provided greater accuracy and precision compared with traditional visual estimates of virulence. These results show that digital image analysis provides a powerful tool for measuring differences in quantitative virulence among strains of Z. tritici.