BT RESISTANCE MANAGEMENT: SOFTWARE FOR RESEARCH, EDUCATION, AND OUTREACH

• Sponsoring Institution: National Institute of Food and Agriculture
• Project Status: COMPLETE
• Funding Source: SMALL BUSINESS GRANT
• Reporting Frequency: Annual
• Accession No.: 1010254
• Grant No.: 2016-33610-25667
• Cumulative Award Amt.: $599,999.00
• Proposal No.: 2016-03977
• Project Start Date: Sep 1, 2016
• Project End Date: Aug 31, 2019
• Grant Year: 2016
• Program Code: [8.13]- Plant Production and Protection-Engineering

Recipient Organization
APPLIED BIOMATHEMATICS
100 NORTH COUNTRY ROAD
SETAUKET,NY 11733

Performing Department
(N/A)

Non Technical Summary
1. Crops that have been genetically modified to protect against insect pests, such as those expressing Bt toxins, have gained tremendous economic and societal importance. These biotechnology products improve the stability of agricultural production through highly effective crop protection while reducing the use of traditional pesticides, fuel, and water by farmers in the US and internationally. Significantly, Bt technology has improved economic outcomes for growers. However, the durability of any given Bt toxin as a protective agent is shortened by the evolution of resistance in the target pest population. Strategies for delaying the evolution of resistance, called insect resistance mangement (IRM), need to be paired intelligently with the crop type, the climate, the target pest, and local agricultural practices. Research into IRM increasingly relies on sophisticated mathematical models that allow many numerical experiments to be done in a short amount of time. The complexity of these models has brought about a number of drawbacks, including 1) inconsistent assumptions that make model comparison difficult, 2) a lack of transparency due to the sheer difficulty of reproducing these models, and 3) the limitation of IRM modeling to a relatively small number of highly skilled researchers.With this USDA-NIFA SBIR Phase II award, Applied Biomathematics intends to research and develop the computational algorithms necessary to support a program that allows users to model the evolution of resistance to Bt or similar transgenic technologies in complex, multi-crop landscapes for a wide variety of target pests. The main goal of the project is to build a simulation engine capable of tracking responses to selection for resistance at up to 12 different genes, each of which may be associated with resistance to one or more Bt toxins. This task is difficult because such genetic complexity can consume a large amount of computing power and memory. Making the performance of the program good enough to use on typical computing platforms requires careful and creative optimization. Our approach will mix algorithms that simulate individual pests in great detail when necessary with faster approaches that lump individuals into populations with similar characteristics. This approach will be faster than models that are purely individual based but should preserve biological detail better than population-level approximations typically used in models with more than one gene. As part of a program that allows users to build IRM models without special training in mathematics, population genetics, or programming and that defines IRM models through a standard set of user inputs, the improved genetics algorithm will lead to a simulation tool that is powerful, transparent, and widely accessible.

Animal Health Component

50%

Research Effort Categories

Basic

0%

Applied

50%

Developmental

50%

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
216	2410	1070	50%
216	2410	2080	50%

Knowledge Area
216 - Integrated Pest Management Systems;

Subject Of Investigation
2410 - Cross-commodity research--multiple crops;

Field Of Science
1070 - Ecology; 2080 - Mathematics and computer sciences;

Keywords

insect resistance management

agricultural biotechnology

regulatory modeling

software

Goals / Objectives
Major GoalsOur major goal is to produce a software tool that improves both the capability and transparency of models used to project the risk of pest resistance to Bt crops. Insect resistance management (IRM) models are used by regulators, USDA and academic scientists, and industry researchers for three main purposes: to identify strategies for resistance management, to assess the risk of resistance for specific crop products under specific scenarios, and to help educate growers about the importance of compliance with IRM guidelines. These mathematical models are complex and include a large number of both parameters and assumptions. A software tool that clearly documents assumptions, standardizes inputs and options, makes models easily reproducible, and is trusted by stakeholders will improve the pace, availability, uniformity, and transparency of IRM modeling. The increasing complexity of inidividual Bt hybrids and the landscape of Bt crops in which they are deployed also requires the development of a tool that enables users to address responses to strong selection on a fairly large (6-12) number of genes.ObjectivesIncrease the number of genes the model can simulateOur existing prototype can handle up to four genes. Recent and emerging models have addressed a larger number but have not yet been used in a published study or registration document for the purpose of Bt resistance management. The main challenge in increasing genetic complexity is its exponential demands on computing resources, including both CPU cycles and memory. In an effort to reduce complexity, some existing models make simplifying assumptions, such as linkage equilibrium, that are known to fundamentally change model predictions. Avoiding such assumptions places a premium on optimization. Toward this goals, we will use a pre-processing step to reduce the number of computations needed to carry out the genetics algorithms later in the simulation and will resort to using sparse data structures to save both CPU cycles and memory requirements.Explore cloud deploymentCloud deployment would provide access to a scalable computing resource, increasing the availability of complex modeling to researchers worldwide.Introduce new modeling optionsWe will expand the options for dispersal, management, and cultural practices. These will include local and long distance dispersal parameters, temporal trends in crop deployment, and responsive management strategies.Reproduce published studiesThe reproduction of published studies gives us insight into the assumptions and methods used in accepted IRM models. It also validates and demonstrates the trustworthiness of the software tool. We expect that reproduction will have a third benefit, namely the illustration of how much more easily models can be shared, reproduced, and expanded when they are constructed with a widely available and standardized platform.DisseminationWe will prepare two manuscripts for peer publication, the first describing our technical approach to modeling more loci and the second summarizing our reproduction of 10 published studies. Findings from our project will be presented at two national meetings, as well. We will also produce white papers detailing worked examples of model reproductions. Finally, we will develop a new website for the project/product and populate it with examples.

Project Methods
Increase in the number of loci in the modelThe increase in the number of loci the model can simulate will be achieved through an algorithm that precalculates the probability of recombination events between pairs of gametotypes and then uses this precalculated result through a population-level simulation. This approach differs from various existing approaches in three specific ways.Some evolutionary models allow a large (>5) number of loci to be simulated by assuming that there is no linkage disequilibrium. This assumption allows the model to address genotypes rather than gametotypes, thereby reducing the number of calculations necessary and the amount of memory required. However, linkage disequilibrium can be an important component of evolutionary responses to selection and has the potential to either forestall or accelerate the evolution of resistance in agricultural pests. Our approach captures linkage disequilibrium and, further, allows users to specify the linkage between loci.To our knowledge, existing models that address a large (>5) number loci without using quantitative genetics do so through the use of individual based simulation. This approach allows open-ended biological complexity but is computationally inefficient, particularly for the large population sizes achieved by agricultural pests. Our approach uses individual based simulation when it is efficient during a precalculation step, but completes the body of the simulation using population-level algorithms.Some previous models have overcome the complexity of polygenic responses to selection by using quantitative genetics. However, quantitative genetic models explicitly assume that the number of loci under selection is very large (approximately infinite) and that selection on each locus is weak. Our approach does not include quantitative genetics expressly because we assume the number of loci, though large, is certainly finite and selection on at least some of the loci is strong.ValidationThe main genetic algorithm will be validated using two-locus cases in which the dynamics can be easily modeled with widely-used deterministic equations. The program as a whole will be validated through reproduction of published studies. Where a difference is found between our program and the results of a published study, we will determine the cause of the difference by 1) verifying that our code is free of errors, 2) checking for differences in assumptions between our model and the published study, and 3) determining whether a change in our program is warranted through a revision of the algorithm or the addition of a user option.

Progress 09/01/16 to 08/31/19

Outputs
Target Audience:Audiences reached during this project include researchers from domestic industry, academic, and agency organizations, as well as researchers from international organizations. These audiences were reached through presentations at the 2016 International Congress of Entomolology and the 2017, 2018, and 2019 annual conferences of the Entomological Society of America, where we presented several oral papers, organized a symposium, and delivered a virtual poster. We had direct conversations about project goals and outcomes with scientists now (after company mergers) at Corteva, Bayer, and Syngenta and involved in the industry groups IRAC and ABSTC. We also spoke with scientists at EPA and USDA-ARS. We presented the software and discussed the potential for conservation-focused pest management with an international group of graduate students at the Student Conference on Conservation Science at the American Museum of Natural History. We also presented the project to board members of the PIVOT Foundation and Partners In Health, philanthropic organizations with a focus on global health. Finally, we presented the software to a business development group from a reinsurance company with the proposal to use it as a tool for crop insurance risk assessment. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? Our USDA Phase II SBIR award permitted the PD, Nicholas Friedenberg, professional development through contact with scientific peers and attendance at national conferences. The funding also allowed us to hire two programmers, who gained valuable early-career experience on the project. Finally, the award allowed us to take on a summer intern in 2018. A computer science / biology double-major, the intern went on to join a graduate program in entomology. How have the results been disseminated to communities of interest?Project goals and results have primarily been disseminated through national conferences, individual discussion with scientists in academia, industry, and government agencies, and through a project portal on our company website. We also presented our software concept to investors with links to the crop insurance industry. The product of our research has also reached its target audience through sales of the advanced prototype, including an ag-biotech startup in the US and a public/private research center in South America. Knowledge gained through project activities from both Phases I and II were also incorporated into worked examples for education. These have been used in a workshop and an undergraduate biology course. The workshop demonstrated potential links between resistance management and conservation goals. The biology course, in the City Univeristy of New York system, serves a diverse student population not normally exposed to questions of agricultural management. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? IMPACT Crops that have been gentically modified to protect against instect pests, such has those producing Bt toxin or RNAi fragments, have gained tremendous economic and societal importance both domestically and around the world. These biotechnology products improve the stability of agricultural systems through effective crop protections while, ideally, reducing the use of traditional pesticides, fuel, and water. As with pesticides, and even crop rotation, transgenic crop protectants can become ineffective over time due to evolutionary changes in pest populations. Resistance to various Bt traits has already evolved in the US, India, and South America, pushing growers adn seed companies to find not only new methods for protecting crops but also new strategies for managing the evolution of resistance, itself. With our USDA-NIFA Phase II award, Applied Biomathematics is building the components of a powerful software tool to simulate resistance evolution, understand its drivers, and explore management strategies. The software, RAMAS IRM, builds on the success of existing mathematical models of insect resistance management (IRM) with the goals of improving consistency, transparency, and acessibilty. The software removes the need for math and programming. But we also aim to exceed the capability of most existing models. As landscapes, and even individual plants, contain more and more transgenic traits, there is a need to model more and more genes. Doing this correctly and efficiently is difficult, as the mathematical problems behind the genetics become astronomically large. RAMAS IRM will be able to simulate up to 12 traits in realistic agriultural landscapes containing billions of insect pests. As we grow the program's capabilities, it will move into the Cloud to utilize scalable memory and processing power to tackle the supercomputer-sized problems posed by biotechnology in agriculture. ACCOMPLISHMENTS, PER OBJECTIVE Objective 1: More genes. We devised an approach to efficiently pre-calculate and store the parameters necessary to simulate a key source of variability driving the evolutionary process, called recombination. Compared with a brute-force approach with 10 genes, our solution is 20 times faster to calculate and uses less than 0.2% as much space in memory. The significance of this improvement is that the complexity of computing genetic recombination is no longer the factor limiting the number of genes that can be simulated. Objective 2: Cloud deployment cability and feasibility. We built a workflow for developing infrastructure for Cloud application development and deployment. The process was tested by converting a simple legacy software product into a web application. The exercise was carried through all stages necessary for the eventual Cloud deployment of RAMAS IRM, including testing scalability and multiple users and license management. We made the resulting case study available as a new commercial product, RAMAS EcoLab Online, which accompanies two electronic textbooks we offer for undergraduate education. The new product was offered as an optional alternative to the legacy software and was chosen by students and professors in almost all cases. Users report better experiences due to cross-platform, cross-device compatibility. The cost of delivering this software via the Cloud has been negligible and costs associated with user support have actually decreased. The greater computational intesnsity of IRM models will introduce greater costs. Costs can be controlled by putting quotas on users that limit the memory and CPUs they can use. Flexbility will be possible through the provision of different quotas at different prices. Hence, what was initially a concern about the Cloud application model now appears to be an advantage for both us and the user; by forcing us to price the product according to expected use, Cloud delivery gives rise to customer choice and creates a ladder by which users will increase their spend as they progress in model complexity. Objective 3: New modeling features. We have dramatically inreased the sophistication of the model of agricultural dynamics. In addition to crop rotations, users can create complex trends in what is planted over time as well as calendar-day-specific applications of pesticide for additional control. Both crop rotations and spraying can be triggered by scouts that occur on specific days of the year, greatly strengthening the linkage between the model and real-world techniques of gathering pest information and makng management decisions. Meanwhile, long distance dispersal is introduced via the one-time randomization of a fraction of adults in the landscape either upon emergence or immediately after mating. Objective 4: Reproduction of published studies. We attempted to use our model to reproduce a set of published studies. Our research indicated that about the half of the studies we sought to reproduce reported slower resistance evolution than we observed in our reproductions. The common feature of these models was the overlap of generations. Intuitively, overlapping generations can lead to mating between generations, which will slow the rate of adaptation. Our model previously treated the population as a single cohort. Even when multiple generations per year were specificed, all individuals resided in a single generation at at time. Development of a demographic model that allows generational overlap is nearly complete and will be presented in the spring of 2020. Objective 5: Dissemination. Dissemination efforts over the course of our project have included five oral papers and a virtual poster at four annual conferences. We organized a symposium at the 2018 annual conference of the Entomological Society of America and led a workshop at the American Museum of Natural History. We featured our software in a peer-reviewed publication on the importance of pest population dynamics to the rate of pesticide resistance evolution, using our software to validate the results of an EPA model and to generate results for pyramided traits of varying efficacy. Our second planned publication is delayed due to our difficulties in reproducing published studies, which we expect to overcome with the completion of our revised demographic model. We built a website to disseminate general information on the topic of IRM modeling and the software features. The site, ramas.com/irm, is updated with news on our activities and provides links to our conference presentations.

Publications

• Type: Conference Papers and Presentations Status: Published Year Published: 2019 Citation: Friedenberg, Nicholas A. 2019. RAMAS IRM: Simulation of insect resistance evolution and management. Virtual Poster, Annual Conference of the Entomological Society of America, St. Louis, MO.

Progress 09/01/17 to 08/31/18

Outputs
Target Audience:We communicated research results to an international gathering of scientists from industry, government agencies, and universities attending the national annual conference of the Entomological Society of America, as well as with specific researchers and collaborators at universities, ag biotech companies, the EPA, and USDA. We also discussed potential linkages between pesticide resistance management and species conservation with graduate students and professors from around the world at the Student Conference on Conservation Science at the American Museum of Natural History. Finally, we have helped to organize a symposium on the role of density dependence in pesticide resistance evolution at the upcoming annual conference of the Entomological Society of America in Vancouver, BC. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? The award has permitted the PD, Nicholas Friedenberg, professional development through contact with scientific peersand the attendance of national conferences. The award allowed us to hire two programmers, one still and undergraduate and one a recent college graduate with degrees in engineering and computer science. Both have gained practical experience on the project. How have the results been disseminated to communities of interest?Project goals and restults have been disseminated throug national conferences, individual discussions with scientists in academia, industry, and government agencies, and through a project portal on our company website. Knowledge gained through project activities from Phases I and II were also incorporated into a workshop on insect resistance management for conservation biology graduate students. What do you plan to do during the next reporting period to accomplish the goals?We expect to focus on integrating the various submodels of the program over the first quarter of the next year, shifting to testing, documentation, and validation after that to complete the project goals. Importantly, validation will include the reproduction of published models, the results of which will be disseminated in the peer-reviewed literature.

Impacts
What was accomplished under these goals? IMPACT Crops that have been genetically modified to protect against insect pests, such as those expressing Bt or RNAi, have gained tremendous economic and societal importance both domestically and around the world. These biotechnology products improve the stability of agricultural systems through effective crop protection while reducing the use of traditional pesticides, fuel, and water. As with pesticides, and even crop rotation, transgenic crop protectants can become ineffective over time due to evolutionary changes in pests. Resistance to various Bt traits has already evolved in the US, India, and South America, pushing growers and seed companies to find not only new methods for protecting crops but also new strategies for managing the evolution of resistance, itself. With our USDA-NIFA SBIR Phase II award, Applied Biomathematics is building a powerful software tool to simulate resistance evolution, understand its drivers, and explore management strategies. The software, RAMAS IRM, builds on the success of existing mathematical models of insect resistance management with the goals of making IRM modeling consistent, transparent, and more accessible. The software removes the need for math or programming. But we also aim to exceed the capability of most existing models. As landscapes, and even individual plants, contain more and more traits, there is a need to model more genes. Doing this correctly and efficiently is difficult, as the mathematical problems behind the genetics become astronomically large. RAMAS IRM will be able to simulate up to 12 traits in realistic agricultural landscapes containing billions of insect pests. A future version utilizing scalable cloud-based computing will be able to tackle the super-computer-sized problems the ag biotech industry faces in its expanding global market. ACCOMPLISHMENTS, per objective: 1) We have further optimized our routine for simulating many genes. Opimization is the main challenge to increasing genetic comlexity, as both memory and CPU cycles can be limiting factors. 2) We used a simple existing piece of software in our library, a unit for teaching undergraduate students about population dynamics in spread over space, as a testbed for developing our workflow and resources for cloud-based software delivery. The result, now commercially distributed, is RAMAS EcoLab Online. Browser-based and programmed to react to the user's device size, it is available across all platforms on computers, tablets, and phones. In the process of building this new software, we developed the tools for license and session management, data delivery, and resource scaling that will be essential to a cloud-based version of RAMAS IRM. 3) We have dramatically increased the sophistication of the software's model of agriculural dynamics. In addition to crop rotations, there are long-term trends in what is planted as well as calendar-day-specific applications of pesticide for additional control. Both crop rotations and spraying can be triggered by scouts that occur on specific days of the year, greatly strengthening the linkage between the model and real-world techniques of gathering pest information and making management decisions. 4) The reproduction of published studies has been further delayed as we wait to finish necessary components of the program. However, a paper we published this year used a simplified version of RAMAS IRM to validate model results produced by an EPA scientist. 5) Dissemination effots have included a conference presentation, organizing a symposium for this coming fall, discussions with companies in the crop protection and ag biotech industries, and the publication of a paper using a simplified version of RAMAS IRM.

Publications

• Type: Journal Articles Status: Published Year Published: 2018 Citation: Martinez, J.C., M.A. Caprio, and N.A. Friedenberg. 2018. Density dependence and growth rate: evolutionary effects on resistance development to Bt (Bacillus thuringiensis). Journal of Economic Entomology 111: 382-390.
• Type: Conference Papers and Presentations Status: Published Year Published: 2017 Citation: Martinez, J.C., M.A. Caprio, and N.A. Friedenberg. 2017. An ecological perspective on dose for insect resistance management. Oral paper. Annual Conference of the Entomological Society of America, Denver, Colorado, USA
• Type: Websites Status: Published Year Published: 2018 Citation: http://www.ramas.com/IRM
• Type: Conference Papers and Presentations Status: Accepted Year Published: 2018 Citation: Friedenberg, N.A. 2018. Ignorance is bliss: How do we model density dependence on a daily time scale? Annual Conference of the Entomological Society of America, Vancouver, BC, Canada

Progress 09/01/16 to 08/31/17

Outputs
Target Audience:We communicated project goals, activities, and/or findings to a global collection of scientists attending the International Congress of Entomology as well as to specific researchers and collaborators at universities, ag biotech companies, the EPA, and USDA. In addition to these technical discussions, we presented an overview of the proposed product to investors in the crop insurance space and had several discussions around the potential use of the product with NGOs with active programs in African smallholder agriculture. Changes/Problems: Nothing Reported What opportunities for training and professional development has the project provided? The award has permitted the PD, Nicholas Friedenberg, professional development through contact with scientific peers and the attendance of a national conference. We have hired two programmers to work on the project. One is just out of college and the other is still and undergraduate. They have shown tremendous growth in skill, knowledge, and confidence through the past year. The PD also led a workshop about insect resistance management and modeling for graduate students attending the American Museum of Natural History's Student Conference on Conservation Science, providing a direct training experience to the attendees. How have the results been disseminated to communities of interest?Results have been disseminated through a national conference, individual discussions with scientists in academia, industry, and government, and through a project portal on our company website. Knowledge gained through project activities from Phase I and II were also incorporated into our workshop on IRM for conservation biology graduate students. What do you plan to do during the next reporting period to accomplish the goals?We consider our overall progress to be slightly ahead of schedule. We expect to adhere to our work plan, which emphasizes cloud deployment research, additional user options, case studies, and dissemination in the second reporting period.

Impacts
What was accomplished under these goals? By objective: 1) We have completed development of the submodel that directly handles the genetics of sexual reproduction for any number of genes. The main accomplishment in this area has been optimization. We reduced both the time required to compute gamete production and the memory required to store the results. 2) We have begin an exploration of cloud software development and deployment using an existing software product as a testbed. The product, which was chosen for its relative simplicity, was rewritten in Java and deployed on Amazon Web Servcies using Tomcat server instances. The front-end user interface is delivered through a web page and uses responsive design elements for mobile compatibility. In addition to establishing a set of tools, a user-authentication system, and a workflow for cloud deployment, this work gave us insight into the value of rewriting the proposed program in a new language that can better exploit the open-source tools available for web applications. 3) Interfaces, classes, and data structures are imlemented for a number of additional modeling options including responsive management actions and temporal trends in frequencies of crop varieties in the landscape. 4) We have been organizing the tasks required to reproduce selected published studies. We have rescheduled the actual reproductions so that we can first integrate the necessary features into the program. 5) Dissemination efforts have included a conference presentation, a workshop, a presentation to an investment group, and personal discussions with scientists in acedemia, industry, and government agencies. We also shared the project goals with individuals at NGOs leading agricultural programs for smallholders in Africa. We are preparing material for a manuscript on our technical approach.

Publications

• Type: Websites Status: Published Year Published: 2017 Citation: www.ramas.com/IRM