Source: AUBURN UNIVERSITY submitted to NRP
THE EVOLUTIONARY DIVERSITY OF AGRICULTURAL PESTS
Sponsoring Institution
National Institute of Food and Agriculture
Project Status
ACTIVE
Funding Source
Reporting Frequency
Annual
Accession No.
1025663
Grant No.
(N/A)
Cumulative Award Amt.
(N/A)
Proposal No.
(N/A)
Multistate No.
(N/A)
Project Start Date
Feb 1, 2021
Project End Date
Jan 31, 2026
Grant Year
(N/A)
Program Code
[(N/A)]- (N/A)
Recipient Organization
AUBURN UNIVERSITY
108 M. WHITE SMITH HALL
AUBURN,AL 36849
Performing Department
Entomology
Non Technical Summary
Many herbivorous insect species are pests of agriculture. They compete with us for food and fiber, and in the process vector plant pathogens. Many more herbivorous insect species will one day become problems for agriculture. As societies and economies have become more connected, so too have biotas. Consequently, each year, new herbivorous insect species are introduced to the US and emerge as pests. To make matters worse, many species that are currently under control will escape it by evolving resistance, and many insect-vectored plant pathogens will evolve greater virulence. Thus, herbivorous insect diversity poses a constantly changing set of high-stakes challenges for agriculture. Our capacity to respond to those challenges is hampered by an incomplete understanding of the scope of herbivorous insect diversity, and the processes by which that diversity changes over time.This project has two overarching goals. The first is to further resolve our view of the scope of herbivorous insect diversity. Most herbivorous insect species have yet to be documented and many of those that have been remain poorly delineated. This makes it harder to respond to new pest problems when they arise. We will address this challenge by working on the systematics of sap-sucking bugs, a group that is especially invasive and destructive. The second goal is to improve our understanding of the evolution of pest populations so that we can be smarter managers of our farms and natural areas. Ultimately, this work aims to support the development of more productive, profitable and sustainable agriculture.p { margin-bottom: 0.1in; direction: ltr; line-height: 115%; text-align: left; orphans: 2; widows: 2; background: transparent }p.western { ; so-language: en-US }p.cjk { ; so-language: zh-CN }p.ctl { ; so-language: hi-IN }a:link { text-decoration: underline }
Animal Health Component
20%
Research Effort Categories
Basic
80%
Applied
20%
Developmental
(N/A)
Classification

Knowledge Area (KA)Subject of Investigation (SOI)Field of Science (FOS)Percent
21131101130100%
Goals / Objectives
Goal 1. Advance the systematics of sap-sucking bugs.Objective 1. Delimit and describe realized and potential pest species. Pest management is more effective when managers have a clear view of pest species diversity.Objective 2. Estimate phylogenetic relationships among species. Phylogenies are predictive and can be used to infer potential pest problems and solutions.Objective 3. Develop diagnostics capacity. Many dangerous pest species are difficult to recognize and distinguish from benign species. Making such distinctions easier will foster more efficient and sustainable agricultural production.Goal 2. Advance understanding of the evolution of pest populations.Objective 4. Identify features that govern the evolution of the interactions between herbivorous insects, their hosts plants, and synthetic insecticides. This will help us predict and manage against the emergence of new pest problems.Objective 5. Develop the theory of virulence evolution. The fact is that most herbivorous insects and insect-vectored parasites do little harm to their host plants. In other words, virulence is exceptional. The theory of virulence evolution poorly describes and predicts the dynamics of plant pathosystems. Thus, the final objective of this project is to help make the theory of virulence evolution more realistic and useful for managing agricultural systems.Each project goal resides in a distinct field of biodiversity research and has its own justification. Goal 1 is to advance the systematics of sap-sucking bug species. This will improve our ability to characterize and recognize species that are (or could become) pests of agricultural in Alabama and elsewhere in the United States, and thus aligns with the mission of AAES. The most fundable (and expensive) facet of this research is the phylogenetics, and the most likely source for external funding is the NSF, specifically the Phylogenetic Systematics Cluster in the Division of Environmental Biology. We have had previous NSF support for work on sap-sucking bug phylogenetics and are in a strong position to seek further funding. Support for Goal 1 can also be sought from commodity groups. We currently have funding from the Alabama Soybean Commission to develop machine-learning based pest diagnostics. Machine-learning systems promise to democratize many pest diagnostics problems, and success in this area should put us in a strong position for additional commodity funding.The systematic research of Goal 1 builds the foundational knowledge needed for Goal 2, that is, to better understand the evolution of pest populations, in particular the evolution of their relationships with host plants. Research in this area has the potential to increase our power to assess the risk of potential pest problems, and to make smarter interventions to prevent or solve them. Goal 2 complements Goal 1 by expanding the potential intellectual impacts of the proposed research. It has the potential of yielding general insights into the evolution of host-pathogen interactions, niche specificity, and virulence. That will bolster proposals to NSF DEB grant programs. It will also open up NIFA funding opportunities, in particular the Foundation Program's Pests and Beneficial Species in Agricultural Production Systems competition.To summarize, over the next five years we plan to use basic research on the systematics of sap-sucking bugs to increase our ability to characterize and detect herbivorous insect species that threaten efficient agricultural production. We will also leverage that systematic research to bootstrap investigations of the evolution of pest populations and pest traits. This research will benefit all Alabamians and has the potential to attract support from regional commodity groups and federal science funding agencies.p { margin-bottom: 0.1in; direction: ltr; line-height: 115%; text-align: left; orphans: 2; widows: 2; background: transparent }p.western { ; so-language: en-US }p.cjk { ; so-language: zh-CN }p.ctl { ; so-language: hi-IN }a:link { text-decoration: underline }
Project Methods
Method for Goal 1: Systematics of sap-sucking bugsObjective 1. Delimit and describe realized and potential pest species. Species will be delimited and described using standard morphological and DNA-sequence approaches as in Hardy et al., 2019 and Hardy & Williams, 2018?.Objective 2. Estimate phylogenetic relationships among sap-sucking bug species.We will use target-enriched genomic sequencing, with a probe set designed to enrich hemipteran samples for Ultra-Conserved Elements (Faircloth, 2017)?. With NSF support we have previously used this approach for Nearctic aphid phylogeny estimation (Hardy et al. in prep). Two priority research lines are to extend our aphid work to the Palaeartic, and to tackle the higher level phylogeny of scale insects.Objective 3. Build diagnostic capacity.The rise of machine learning has opened up a new frontier for diagnostic tool development. In a nutshell, we aim to apply to pest diagnostics the same kinds of algorithms that are used in spam filters and self-driving cars. These algorithms have proven particularly useful for vision-based classification problems, such as the detection of lung cancers from x-ray images, and are beginning to be used to diagnose plant diseases.The biggest challenge in developing a machine learning application is obtaining sufficient training data. In general, the more training data the better, and the best performing computer-vision based classification systems have been training on hundreds of thousands of labeled images. The research focus of this project is to experiment with ways of making artificial neural networks learn more efficiently from fewer training images. This is an active area of software research and several promising approaches have been developed. Examples include image segmentation, data augmentation, transfer learning, and capsule networks. With these approaches, useful machine learning applications have been developed with training data sets consisting of only hundreds to a few thousand training images - a scale feasible for many agricultural pest diagnostics problems.Methods for Goal 2: Advance understanding of the evolution of pest populations.Identify features that govern the evolution of the interactions between herbivorous insects, their host plants, and synthetic insecticides. To address our questions about the evolutionary ecology of herbivorous insects, we will primarily use comparative statistical approaches. For example, we have used comparative phylogenetic analysis to predict the evolution of insecticide resistance (Hardy et al. 2018; Crossley et al. 2020). And we are currently exploring the role of host-use and other types of niche divergence as drivers of speciation in Nearctic aphids with hierarchical path models (a.k.a., structured linear equations) that combine an estimate of aphid phylogeny with correlative-GIS models of aphid niches based on hundreds of thousands of specimen data and rich maps of climate, soil and land-use variables. With such models we can infer causal relationships between niche variables and species diversity, and have been able to perform some of the first tests of long-standing assumptions about the role of host-parasite co-evolutionary antagonism in driving herbivorous insect speciation (Hardy et al. in prep).Another mode of comparative analysis that we have and will continue to use is meta-analysis. For example, we have used meta-analysis to improve our understanding of what governs competition between herbivorous insects, and for how combinations of pesticides and parasites affect the health of pollinators (Bird et al., 2019; Bird et al. in press). And we are currently using meta-analysis to evaluate potential evolutionary-ecology feedbacks between a herbivorous insect's diet breadth, and the diversity and interaction strength of its natural enemy assemblage. This study could yield basic insights into the process of host-use evolution, along with actionable insights into what kinds of natural enemy species and communities deliver the most effective control of herbivorous pest populations.In addition to these comparative statistical approaches, we will take two other types of approaches. We have recently acquired the capacity to address evolutionary biology questions with individual-based, forward time, population genetic simulations. Our first application of this approach was to evaluate the role of antagonistic pleiotropy in driving the evolution of specialization in a polygenic trait, such as host-use performance in herbivorous insects (Hardy and Forister in review). Against the prevailing wisdom, we found that, with polygeny, genetic constraints such as pleiotropy probably do little to push for ecological specialization. Conducting this study opened our eyes to the vast potential of these models to test and extend much of the classic theory of evolution biology.We will also use experimental approaches. Specifically, we have developed a sugarcane aphid system that will allow us to experimentally evolve host-use in the greenhouse, and test hypotheses about the genetic contingencies and consequences of adaptation to a novel host plant.Develop the theory of virulence evolutionAs for the previous objective, for this we will use a variety of comparative statistical approaches including meta-analysis (to test the assumptions of the trade-off hypothesis), and comparative phylogenetics (to test for correlations between virulence and variables such as host longevity, vector community composition, vector community transmission efficiency, pathogen community diversity, etc.). We also use population genetic simulations to gain intuition about how vector biology could affect the evolution of virulence. For example, we will explore the extent to which virulence could evolve as a byproduct of competition for vector berths.Also, in collaboration with Dr Leo De La Fuente in Plant Pathology, we will develop experiments to check the assumptions and insights of the statistical and simulation analyses. We will use the xylem-limited bacteria Xylella fastidiosa as a model system.h5 { margin-top: 0.08in; margin-bottom: 0.04in; direction: ltr; text-align: left; orphans: 2; widows: 2; background: transparent; page-break-after: avoid }h5.western { ; so-language: en-US; font-weight: bold }h5.cjk { ; so-language: zh-CN; font-weight: bold }h5.ctl { ; so-language: hi-IN; font-weight: bold }p { margin-bottom: 0.1in; direction: ltr; line-height: 115%; text-align: left; orphans: 2; widows: 2; background: transparent }p.western { ; so-language: en-US }p.cjk { ; so-language: zh-CN }p.ctl { ; so-language: hi-IN }a:link { text-decoration: underline }

Progress 02/01/21 to 09/30/21

Outputs
Target Audience:In 2021, we primarily reached other academics, in particular, ecologists and evolutionary biologists interested in the evolution of pesticide resistance, and the factors contributing to the decline of pollinator populations. That being said, our research on pollinator health did attract the attention of the bee keeping community. Also, some unpublished advances in disease diagnostics were shared with local soybean producers. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided?The project supported the substantial training and professional development of three graduate students, who learned about machine learning, population genetic modeling, statistical analysis and science communication. Our work on the evolution of virulence and pesticide resistance was also icorportated into a new graduate-level course on applied evolutionary biology, and helped build the theoretical biology skills of six additional graduate students. How have the results been disseminated to communities of interest?Last year, our meta-analysis of how multiple stressors combine to affect honeybee health was published in a peer-reviewed jounral, and shared among bee keepers. We also published in a peer-reviewed journal our latest statistical efforts to understand what factors promote the evolution of insectice resistance. Several additional manuscripts are in review, describing our comparative phylogenetic work on aphids, our meta-analysis of the relationship between viurlence and pathogen replication in plant pathosystems, and our population genetic simulation-based exploration of how host-use evolution in herbivorous insects depends on the genetic architecture of host-use preference and performance. What do you plan to do during the next reporting period to accomplish the goals?Goal 1. Objective 1. We will uses evolutionary theory to direct empirical research on plant pathogens. In particular, in the comming year, we aim to better characterize the competive interactions between strains of plant-pathogen bacterial strains that are co-infecting hosts. We also aim to evaluate the extent to which performance within insect vectors trades off against performance within host plants. Objective 2. Here our main efforts are for dissemination. We aim to publish our findings about aphid phylogeny, and what it tells us about the ecological drivers of herbivorous insect speciation. Objective 3. We plan to continue refining our deep-learning based classifier of soybean foliar diseases. This coming year we also plant to publish a paper describing the development process. And we will be considering other possible applications of deep-learning in agricultural diagnostics. Goal 2. Objective 4. The main thrust for this comming year will be to finish and disseminate a theoretical investigation of how we can vary the application of insecticides in space and time so as to delay the evolution of insecticide resistance for as long as possible. Objective 5. This will be our main focus this coming year. We aim to use populaiton genetics modeling to advance our understanding of how interactions between strains within a host might affect the evolution of virulence. We also aim to develop our capacity to investigate the ecology of virulence in the plant pathogen Xylella fastidiosa.

Impacts
What was accomplished under these goals? Goal 1. Objective 1. Nothing to report. Objective 2. In 2021, we completed a phylogenetic analysis of DNA-sequence data sampled from more than a thousand loci from more than 400 species of Nearctic aphids. We are currently working on manuscripts to present an updated view of the phylogenetics systematics of aphids, and to describe new insights into the ecological drivers of hervirous insect speciation. Objective 3. In the past year, we made significant progress with the development of an deep-learning based automated classifier of digital images of soybean foliar diseases. It can reliably distinguish between about half a dozen of the most common diseases in the southern states. Goal 2. Objective 4. Last year, we conducted research that improved our ability to predict the evolution of insecticide resistance evolution. Leveraging data from large public datasets, and using state-of the art machine learning approaches, we found large effects from (1) the overall-structural similarity between insecticidal compounds and the secondary compounds found in a herbivores diet, and (2) a herbivore's diet breadth. We also used population genetic simulation models to gain insights in how the evolution of a herbivorous insect's diet depends on the genetic architectures of host-use preference and performance. And we used meta-analysis to better understand how multiple stressors, including incidental exposure to insecticides, affect the health of honeybees. Objective 5. Using indivual-based, forward time, quantitative genetic models, we have demonstrated that high virulence can evolve in a population of plant pathogens via antagonstic pleiotropy affecting performance within host plants and insect vectors. Using meta-analysis, we found the a core assumption of the standard theory of virulence evolution -- namely, that virulence is the byproduct of aggresive replication -- does not hold for plant pathosystems.

Publications

  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Bird, G., Wilson, A. E., Williams, G. R., & Hardy, N. B. (2021). Parasites and pesticides act antagonistically on honey bee health. Journal of Applied Ecology, 58(5), 997-1005.
  • Type: Journal Articles Status: Published Year Published: 2021 Citation: Crossley, M. S., Snyder, W. E., & Hardy, N. B. (2021). Insectplant relationships predict the speed of insecticide adaptation. Evolutionary applications, 14(2), 290-296.
  • Type: Journal Articles Status: Under Review Year Published: 2021 Citation: Hardy, N. B., & Forister, M. (2021). Niche specificity, polygeny, and pleiotropy in herbivorous insects. bioRxiv.