Bayesian analysis in plant pathology

Plant science in the age of simulation intelligence

Historically, plant and crop sciences have been quantitative fields that intensively use measurements and modeling. Traditionally, researchers choose between two dominant modeling approaches: mechanistic plant growth models or data-driven, statistical methodologies. At the intersection of both paradigms, a novel approach referred to as "simulation intelligence", has emerged as a powerful tool for comprehending and controlling complex systems, including plants and crops. This work explores the transformative potential for the plant science community of the nine simulation intelligence motifs, from understanding molecular plant processes to optimizing greenhouse control. Many of these concepts, such as surrogate models and agent-based modeling, have gained prominence in plant and crop sciences. In contrast, some motifs, such as open-ended optimization or program synthesis, still need to be explored further. The motifs of simulation intelligence can potentially revolutionize breeding and precision farming towards more sustainable food production.

Rising Out of the Ashes: Additive Genetic Variation for Crown and Collar Resistance to Hymenoscyphus fraxineus in Fraxinus excelsior

Since the early 1990s, ash dieback due to the invasive ascomycete Hymenoscyphus fraxineus is threatening Fraxinus excelsior in most of its natural range. Previous studies reported significant levels of genetic variability in susceptibility in F. excelsior either in field or inoculation experiments. The present study was based on a field experiment planted in 1995, 15 years before onset of the disease. Crown and collar status were monitored on 777 trees from 23 open-pollinated progenies originating from three French provenances. Health status was modeled using a Bayesian approach where spatiotemporal effects were explicitly taken into account. Moderate narrow-sense heritability was found for crown dieback (h² = 0.42). This study is first to show that resistance at the collar level is also heritable (h² = 0.49 for collar lesions prevalence and h² = 0.42 for their severity) and that there is significant genetic correlation (r = 0.40) between the severities of crown and collar symptoms. There was no evidence for differences between provenances. Family effects were detected, but computing individual breeding values showed that most of the genetic variation lies within families. In agreement with previous reports, early flushing correlates with healthier crown. Implications of these results in disease management and breeding are discussed.

Estimation of leaf wetness duration requirements of foliar fungal pathogens with uncertain data-an application to Mycosphaerella nawae

Wetness of the host surface is a critical environmental factor for the development of foliar fungal diseases, but it is difficult to estimate the wetness durations required by pathogens for infection when only few experimental data are available. In this paper, we propose a method to estimate wetness duration requirements of foliar fungal pathogens when precise experimental data are not available. The proposed method is based on approximate Bayesian computation. It only requires lower and upper bounds of wetness duration requirements for one or fewer temperatures. We describe the method, show how to apply it to an infection model, and then present a case study on Mycosphaerella nawae, the causal agent of circular leaf spot of persimmon. In this example, the parameters of a simple infection model were estimated using experimental data found in the literature for the pathogen, and the model was applied to assess the risk in a Spanish area recently affected by the disease. The results showed that the probability of successful infection was higher than 0.5 for 32% of the on-site wetness durations recorded in the affected area. Results obtained with simulated data showed that our method was able to improve the estimation of wetness duration requirement. Given the flexibility of the proposed method, we expect it to become adopted for assessing the risk of introduction of exotic fungal plant pathogens.

Risk factors for crop health under global change and agricultural shifts: a framework of analyses using rice in tropical and subtropical Asia as a model

Plant disease epidemiology requires expansion of its current methodological and theoretical underpinnings in order to produce full contributions to global food security and global changes. Here, we outline a framework which we applied to farmers' field survey data set on rice diseases in the tropical and subtropical lowlands of Asia. Crop health risks arise from individual diseases, as well as their combinations in syndromes. Four key drivers of agricultural change were examined: labor, water, fertilizer, and land availability that translate into crop establishment method, water shortage, fertilizer input, and fallow period duration, respectively, as well as their combinations in production situations. Various statistical approaches, within a hierarchical structure, proceeding from higher levels of hierarchy (production situations and disease syndromes) to lower ones (individual components of production situations and individual diseases) were used. These analyses showed that (i) production situations, as wholes, represent very large risk factors (positive or negative) for occurrence of disease syndromes; (ii) production situations are strong risk factors for individual diseases; (iii) drivers of agricultural change represent strong risk factors of disease syndromes; and (iv) drivers of change, taken individually, represent small but significant risk factors for individual diseases. The latter analysis indicates that different diseases are positively or negatively associated with shifts in these drivers. We also report scenario analyses, in which drivers of agricultural change are varied in response to possible climate and global changes, generating predictions of shifts in rice health risks. The overall set of analyses emphasizes the need for large-scale ground data to define research priorities for plant protection in rapidly evolving contexts. They illustrate how a structured theoretical framework can be used to analyze emergent features of agronomic and socioecological systems. We suggest that the concept of "disease syndrome" can be borrowed in botanical epidemiology from public health to emphasize a holistic view of disease in shifting production situations in combination with the conventional, individual disease-centered perspective.

An Evidence-Based Approach to Plum Pox Virus Detection by DASI-ELISA and RT-PCR in Dormant Period

An evidence-based approach, such as those developed in clinical and veterinary medicine, was applied to the detection of Plum pox virus (PPV) during the dormant period. A standardized methodology was used for the calculation of parameters of the operational capacity of DASI-ELISA and RT-PCR in wintertime. These methods are routinely handled to test the sanitary status of plants in national or international trading and in those cases concerning export-import of plant materials. Diagnosis often has to be performed during the dormant period, when plant material is commercialized. Some guidelines to interpret diagnostic results of wintertime are provided in an attempt to minimize risks associated with the methods and over-reliance on the binary outcome of a single assay. In order to evaluate if a complementary test increased the confidence of PPV diagnosis when discordant results between DASI-ELISA and RT-PCR are obtained, NASBA-FH also was included. Likelihood ratios of each method were estimated based on the sensitivity and specificity obtained in wintertime. Subsequently, a Bayesian approach was performed to calculate post-test probability of PPV infection in spring. Results of evidence-based approach show that different PPV prevalences require different screening tests. Thus, at very low PPV prevalence levels DASI-ELISA should be used as the election method, whilst at the highest PPV prevalence levels RT-PCR should be performed. NASBA-FH could be used at medium prevalences to clarify discordances between DASI-ELISA and RT-PCR.

Use of bayesian methods to improve prediction of panicle and shoot blight severity of pistachio in california

ABSTRACT Panicle and shoot blight, caused by a Fusicoccum sp., is one of the major aboveground diseases of pistachio in California. The effects of temperature, number of continuous rainy days in April and May, irrigation system, and incidence of latent infection of the Fusicoccum sp. on severity of panicle and shoot blight of pistachio leaves and fruit have been quantified previously, using data collected from 1999 through 2001. A predictive model for leaves and another model for fruit with good explanatory power were generated. In 2003 and 2004, newly collected data were used to evaluate the two models with non-Bayesian and Bayesian methods. The 95% credible (i.e., confidence) intervals of initial (before modification with non-Bayesian and Bayesian methods) and updated parameter estimates were used to investigate their prognostic validity. In 2003, the non-Bayesian analysis resulted in all parameter estimates, with the exception of cumulative daily mean temperature from 1 June until harvest, having different 95% confidence intervals than the parameter estimates of the original models. In addition, the parameter estimates for drip irrigation for the leaf infection and the parameter estimates for drip irrigation and number of continuous rainy days in April and May for fruit infection were not statistically significant. With Bayesian methods, the reestimated model parameters had overlapping 95% credible intervals with the initial estimated parameters, except for the number of continuous rainy days in April and May. When the two sets of modified parameter estimates were used to predict disease severity, statistically significant (alpha = 0.05) differences between observed and predicted disease severities were found with non-Bayesian analysis for leaf infection in three locations and with Bayesian analysis for fruit infection in one orchard. The parameter estimates were modified again at the end of the 2004 season and were all statistically significant with both non-Bayesian and Bayesian methods. Both sets of parameter estimates gave predictions that were not significantly different from observed disease severity on leaves and fruit in all monitored orchards in 2004. In summary, Bayesian methods gave more consistent results when used to update parameter estimates with new information and yielded predictions not statistically different from observed disease severity in more cases than the non-Bayesian analysis.

New applications of statistical tools in plant pathology

ABSTRACT The series of papers introduced by this one address a range of statistical applications in plant pathology, including survival analysis, nonparametric analysis of disease associations, multivariate analyses, neural networks, meta-analysis, and Bayesian statistics. Here we present an overview of additional applications of statistics in plant pathology. An analysis of variance based on the assumption of normally distributed responses with equal variances has been a standard approach in biology for decades. Advances in statistical theory and computation now make it convenient to appropriately deal with discrete responses using generalized linear models, with adjustments for overdispersion as needed. New nonparametric approaches are available for analysis of ordinal data such as disease ratings. Many experiments require the use of models with fixed and random effects for data analysis. New or expanded computing packages, such as SAS PROC MIXED, coupled with extensive advances in statistical theory, allow for appropriate analyses of normally distributed data using linear mixed models, and discrete data with generalized linear mixed models. Decision theory offers a framework in plant pathology for contexts such as the decision about whether to apply or withhold a treatment. Model selection can be performed using Akaike's information criterion. Plant pathologists studying pathogens at the population level have traditionally been the main consumers of statistical approaches in plant pathology, but new technologies such as microarrays supply estimates of gene expression for thousands of genes simultaneously and present challenges for statistical analysis. Applications to the study of the landscape of the field and of the genome share the risk of pseudoreplication, the problem of determining the appropriate scale of the experimental unit and of obtaining sufficient replication at that scale.