Pathogens and pests are one of the major threats to agricultural productivity worldwide. For decades, targeted resistance breeding was used to create crop cultivars that resist pathogens and environmental stress while retaining yields. The often decade-long process of crossing, selection, and field trials to create a new cultivar is challenged by the rapid rise of pathogens overcoming resistance. Similarly, antimicrobial compounds can rapidly lose efficacy due to resistance evolution. Here, we review three major areas where computational, imaging and experimental approaches are revolutionizing the management of pathogen damage on crops. Recognizing and scoring plant diseases have dramatically improved through high-throughput imaging techniques applicable both under well-controlled greenhouse conditions and directly in the field. However, computer vision of complex disease phenotypes will require significant improvements. In parallel, experimental setups similar to high-throughput drug discovery screens make it possible to screen thousands of pathogen strains for variation in resistance and other relevant phenotypic traits. Confocal microscopy and fluorescence can capture rich phenotypic information across pathogen genotypes. Through genome-wide association mapping approaches, phenotypic data helps to unravel the genetic architecture of stress- and virulence-related traits accelerating resistance breeding. Finally, joint, large-scale screenings of trait variation in crops and pathogens can yield fundamental insights into how pathogens face trade-offs in the adaptation to resistant crop varieties. We discuss how future implementations of such innovative approaches in breeding and pathogen screening can lead to more durable disease control.

中文翻译：

---

通过综合成像和计算方法应对农业中的微生物威胁


病原体和害虫是全球农业生产力的主要威胁之一。几十年来，有针对性的抗性育种被用来培育能够抵抗病原体和环境压力，同时保持产量的作物品种。创造新品种的杂交、选择和田间试验的过程通常长达十年，但由于克服抗药性的病原体迅速增加而面临挑战。同样，抗菌化合物会由于耐药性的演变而迅速失去功效。在这里，我们回顾了计算、成像和实验方法正在彻底改变作物病原体损害管理的三个主要领域。通过适用于良好控制的温室条件和直接在田间的高通量成像技术，植物病害的识别和评分得到了显着改善。然而，复杂疾病表型的计算机视觉需要显着改进。与此同时，类似于高通量药物发现筛选的实验设置使得筛选数千种病原体菌株的耐药性和其他相关表型性状的变化成为可能。共聚焦显微镜和荧光可以捕获跨病原体基因型的丰富表型信息。通过全基因组关联图谱方法，表型数据有助于揭示应激和毒力相关性状的遗传结构，从而加速抗性育种。最后，对作物和病原体的性状变异进行联合、大规模筛选可以对病原体在适应抗性作物品种时如何面临权衡产生基本见解。我们讨论了未来在育种和病原体筛查中实施此类创新方法如何能够实现更持久的疾病控制。