Plant pathology has developed a wide range of concepts and tools for improving plant disease management, including models for understanding and responding to new risks from climate change. Most of these tools can be improved using new advances in artificial intelligence (AI), such as machine learning to integrate massive data sets in predictive models. There is the potential to develop automated analyses of risk that alert decision-makers, from farm managers to national plant protection organizations, to the likely need for action and provide decision support for targeting responses. We review machine-learning applications in plant pathology and synthesize ideas for the next steps to make the most of these tools in digital agriculture. Global projects, such as the proposed global surveillance system for plant disease, will be strengthened by the integration of the wide range of new data, including data from tools like remote sensors, that are used to evaluate the risk of plant disease. There is exciting potential for the use of AI to strengthen global capacity building as well, from image analysis for disease diagnostics and associated management recommendations on farmers’ phones to future training methodologies for plant pathologists that are customized in real-time for management needs in response to the current risks. International cooperation in integrating data and models will help develop the most effective responses to new challenges from climate change. Expected final online publication date for the Annual Review of Phytopathology, Volume 60 is August 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Policymakers and donors often need to identify the locations where technologies are most likely to have important effects, to increase the benefits from agricultural development or extension efforts. Higher quality information may help to target the high-benefit locations, but often actions are needed with limited information. The value of information (VOI) in this context is formalized by evaluating the results of decision making guided by a set of specific information compared to the results of acting without considering that information. We present a framework for management performance mapping that includes evaluating the VOI for decision making about geographic priorities in regional intervention strategies, in case studies of Andean and Kenyan potato seed systems. We illustrate use of recursive partitioning, XGBoost, and Bayesian network models to characterize the relationships among seed health and yield responses and environmental and management predictors used in studies of seed degeneration. These analyses address the expected performance of an intervention based on geographic predictor variables. In the Andean example, positive selection of seed from asymptomatic plants was more effective at high altitudes in Ecuador. In the Kenyan example, there was the potential to target locations with higher technology adoption rates and with higher potato cropland connectivity, i.e., a likely more important role in regional epidemics. Targeting training to high management performance areas would often provide more benefits than would random selection of target areas. We illustrate how assessing the VOI can contribute to targeted development programs and support a culture of continuous improvement for interventions.

No abstract

Virome analysis via high-throughput sequencing (HTS) allows rapid and massive virus identification and diagnoses, expanding our focus from individual samples to the ecological distribution of viruses in agroecological landscapes. Decreases in sequencing costs combined with technological advances, such as automation and robotics, allow for efficient processing and analysis of numerous samples in plant disease clinics, tissue culture laboratories, and breeding programs. There are many opportunities for translating virome analysis to support plant health. For example, virome analysis can be employed in the development of biosecurity strategies and policies, including the implementation of virome risk assessments to support regulation and reduce the movement of infected plant material. A challenge is to identify which new viruses discovered through HTS require regulation and which can be allowed to move in germplasm and trade. On-farm management strategies can incorporate information from high-throughput surveillance, monitoring for new and known viruses across scales, to rapidly identify important agricultural viruses and understand their abundance and spread. Virome indexing programs can be used to generate clean germplasm and seed, crucial for the maintenance of seed system production and health, particularly in vegetatively propagated crops such as roots, tubers, and bananas. Virome analysis in breeding programs can provide insight into virus expression levels by generating relative abundance data, aiding in breeding cultivars resistant, or at least tolerant, to viruses. The integration of network analysis and machine learning techniques can facilitate designing and implementing management strategies, using novel forms of information to provide a scalable, replicable, and practical approach to developing management strategies for viromes. In the long run, these management strategies will be designed by generating sequence databases and building on the foundation of pre-existing knowledge about virus taxonomy, distribution, and host range. In conclusion, virome analysis will support the early adoption and implementation of integrated control strategies, impacting global markets, reducing the risk of introducing novel viruses, and limiting virus spread. The effective translation of virome analysis depends on capacity building to make benefits available globally.