Progress 07/01/22 to 06/30/23
Outputs
Target Audience:Dr. Machado: presented the development of this work at several national swine meetings, including U.S. swine industry preparedness for emerging diseases (GIS Week Lightning Talk (NCSU, Raleigh, NC)Swine Innovation Forum (Goldsboro, NC)-Feb-2023; Gentle Introduction to RABapp (Portsmouth, NH)-2023; Major Enhancement of U.S. Swine Industry Biosecurity (Cross-Border Threat Screening and Supply Chain Defense (CBTS) Center of Excellence (COE), Distinguished Speaker Series, Online)-2023. Video lectures about this tool, on-farm biosecurity, major swine diseases, and epidemiology can be found herehttps://machado-lab.github.io/research/biosecurity/ 2023-We have updated the R package associated with this project, making it more accessible to the scientific public https://nfj1380.github.io/mrIML/index.html Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?As a course coordinator at DVM and Graduate level, Dr. Machado introduced machine learning to four-year veterinary students (Spring-2023) and the Ph.D. level (Fall 2022) via lectures and hands-on (R programming classes). How have the results been disseminated to communities of interest?Dr. Machado and Ph.D. student Abagael Sykes presented at national conferences. Two websites continue to be updated: 1) with the guidance on how to use the R package and 2) With tutorials and lectures for producers and swine veterinarians https://machado-lab.github.io/research/biosecurity/. What do you plan to do during the next reporting period to accomplish the goals?Promote the use of the tool by extending access to RABapp to other states.
Impacts
What was accomplished under these goals?
Aim 1 is completed and published, seehttps://doi.org/10.1111/tbed.14369 Aim 2the reporting systems to communicate with our public will be integrated into the Machado lab website; the previously developed Rapid Access Biosecurity (RAB) app (RABapp) https://machado-lab.github.io/rabapp/ to host this tool within the same ecosystem. We have developed the biosecurity machine learning tool available for RABapp users. Extension activities via web-call have been delivered to more than 40 swine production companies, with production in 22 US states.
Publications
- Type:
Websites
Status:
Other
Year Published:
2023
Citation:
Machado, Gustavo. Multi Response Interpretable Machine Learning. Github. mrIML: Multivariatemulti-responseinterpretable machine learning, 2023. https://nfj1380.github.io/mrIML/.
|
Progress 07/01/19 to 06/30/23
Outputs
Target Audience:Swine Industries,Swine Producers, andSwine Veterinarians. Seminars and Annual Meetings of the American Association of Swine Veterinarians (AASV),North Carolina Swine VeterinaryAssociation (NCVMA),National Pork Board biosecurity meeting, veterinarians, pharma companies, andchief executive officer's ofpig companies,Swine Virtual Conference and Forum,Pennsylvania Farm Bureau Virtual Conference, NCSU Center for Geospatial Analytics Research Roundtable Lightning Talk Changes/Problems:Previous changes/problems we required the project extension and internal budget modification due to delayed stages of the project impacted by COVID-19, process to complete hiring, lull in receiving initial funds, and then a re-budget to hire (Iowa) grad students instead of (Iowa) post doc, and a new hire (North Carolina) post doc for data collection . What opportunities for training and professional development has the project provided?Professional development and training course activities for Veterinarians and DVM students. How have the results been disseminated to communities of interest?Presentations at conferences, seminars, meetings, two websites with guidance on how to use the R package and tutorials and lectures forproducers and swineveterinarians. What do you plan to do during the next reporting period to accomplish the goals?
Nothing Reported
Impacts
What was accomplished under these goals?
We developed and validated biosecurity machine learning machine-learning algorithms tool to calculate the predicted risk of PRRSv, carried out the analysis of (150) breeding farms from (5) US states, data collection has been achieved, and a farm-level ranked list of the relative relevance of each biosecurity measure was generated. The package was further refined and improved to provide a detailed report for each individual biosecurity aspect that has a positive or negative impact on the risk of a new PRRSv case, the way each biosecurity contribution to predicting PRRSv is measured. We added a method for allowing the introduction of uncertainty to our predictions. We advanced the benchmarking functionality not only to the company level, but down to the individualized farm. This identified key biosecurity features that either contribute to a farm, to remaining infected with PRRSV, or to continue negative (absence of outbreaks); or farms of system vs. all other farms in the region; within-system benchmarking; and within the top 20 biosecurity features by the order of relevance over the risk of PRRSv of for all farms. Extension activities via web-call have been delivered to more than 40 swine production companies, with production in 22 US states. We created a vignette and an R package (mrIML) (https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html), reporting systems communicate with the public with integration into Dr Machado's Lab website, and the developed Rapid Access Biosecurity application (RABapp) (https://machado-lab.github.io/rabapp/) to host this tool within the same ecosystem, connecting all R software infrastructure. This delivers an easy-to-use interface and fast results. We held meetings and workshops at national and international pig conferences, for pig producers and other machine learning experts, provided a how-to-interpret guide on the biosecurity tool, how-to-use videos for RABapp users, and distributed learning material online via a newsletter and website: (https://machado-lab.github.io/Software/). We also developed videos about machine learning to make it accessible to the less-technological-savvy stakeholders, and proposed a series of continuing education about PRRSV and epidemiology available to the online public via website: (https://machado-lab.github.io/research/biosecurity/ )The well documented material is open-source free-to-use computer codes embedded within an R package, developed by Dr. G. Machado and A. Sykes, Ph.D. student. For in-depth package details visit the maintained website: (https://nfj1380.github.io/mrIML/)and publication: (https://doi.org/10.1111/tbed.14369).
Publications
- Type:
Journal Articles
Status:
Published
Year Published:
2019
Citation:
SILVA GSS, MACHADO G, BAKER KL, HOLTKAMP DJ, LINHARES DCL. Machine-learning algorithms to identify key biosecurity practices and factors associated with breeding herds reporting PRRS outbreak. Prev Vet Med. 2019; 171:104749.
- Type:
Journal Articles
Status:
Published
Year Published:
2019
Citation:
PEREZ A, LINHARES DCL, ARRUDA AG, VANDERWAAL K, MACHADO G, VILALTA C, SANHUEZA J, TORRISON J, TORREMORELL M, CORZO C. Individual or common good? Voluntary data sharing to inform disease surveillance systems in food animals. Frontiers in Vet Sci. 2019; 6(194): 1-7.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2020
Citation:
ALMEIDA MN, GAVIS JA, MACHADO G, ZIMMERMAN JJ, LINHARES DCL. PRRSV farrowing room surveillance - what do negative PCRs really mean? Proceedings of the Allen D. Leman Swine Conference. Sep 14, 2020. Poster 69. Saint Paul, MN.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2020
Citation:
MAUCH BW, SILVA G, MACHADO G, HOLTKAMP DJ, LINHARES DCL. Screening the vulnerability of porcine reproductive and respiratory syndrome virus (PRRSV) introduction in breeding herds using a short survey. Annual Meeting of the American Association of Swine Veterinarians. Mar 7-10, 2020. Atlanta. P79-80.
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2019
Citation:
SILVA GS, MACHADO G, BAKER K, HOLTKAMP D, LINHARES DCL. Investigating biosecurity aspects related to PRRSV outbreaks. Proc 50th American Association of Swine Veterinarians Annual Meeting. Lake Buena Vista, FL. March 9-12, 2019. P40-42.
|
Progress 07/01/21 to 06/30/22
Outputs
Target Audience:Dr. Machado: presented at the Pennsylvania Farm Bureau to members of the local swine industry titled "swine biosecurity and FAD preparedness" Online. 2021 Food Animal Innovation Summit, Abagael Sykes, the Ph.D. student presented a posted and short talk titled: "Development of an interpretable machine learning-based online tool to guide swine on-farm biosecurity implementation", Raleigh NC. The new mrIML: Multivariate (multi-response) interpretable machine learning has been released with new documentation of how to use the local model interpretation developed in this project for other food-animal biosecurity https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html A couple of talks to pharma and pig companies have been delivered about the R package this project is developing. Changes/Problems:The impact of COVID-19 delayed the developments under the extension aim, in addition, we requested a project modification to hire a post-doctoral researcher to complete the development of the tool, while the team in Iowa (sub-award) also requested a re-budget for two graduate students instead of postdoctoral research to collect the third round of biosecurity data from participating companies. What opportunities for training and professional development has the project provided?Dr. Machado as course coordinator lectured senior DVM students on how to use the proposed tool (course number VMP-971) the class included one day on machine learning applied to food-animals. Veterinarians of the 4 new participating companies received information on how to read and use the report they received, this was online series of calls. The new educational videos are now also available here https://machado-lab.github.io/research/biosecurity/. How have the results been disseminated to communities of interest?Dr. Machado and Ph.D. student Abagael Sykes presented at national conferences. We have also written a vignette hosted here https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html, in addition to the newly recorded videos available in the link below https://machado-lab.github.io/research/biosecurity/. What do you plan to do during the next reporting period to accomplish the goals?Fully develop and deliver the biosecurity machine-learning tool via the RABapp portal. Continue to collect biosecurity information from the current participants and expand to other commercial pig-producing companies. Promote the use of the tool during the upcoming year.
Impacts
What was accomplished under these goals?
Our specific project objectives were: Research: 1. Select and validate machine-learning algorithms to calculate the predicted risk of PRRSv introduction; 1.1. Generate a farm-level ranked list of the relative relevance of each biosecurity measure. 2. Extension: 2: Generate farm-level predictions that can be shared with producers to promote disease management, control, and facilitate comparison between farms. 3. Develop a reporting system for producers to access their own risks. 4. Promote the use of the predicted risk of PRRSv introduction in disease management and prevention through producer-oriented extension. To address objectives 1 & 1.1. Both aims have been completed. We have gone beyond the proposed aims to provide well-documented material via open-source free-to-use computer codes embedded within an R package, developed by Dr. Machado and Abagael Sykes, Ph.D. student. For in-depth details visit the package website we developed and maintain https://nfj1380.github.io/mrIML/ Extension: To address objectives 2 & ultimately 3. Has been completed and reports sent. Objective 3, the reporting systems to communicate with our public will be integrated into the Machado lab website the previously developed The Rapid Access Biosecurity (RAB) app (RABapp) https://machado-lab.github.io/rabapp/ to host this tool within the same ecosystem. We have developed the base shiny code for the biosecurity machine learning tool. Finally, for objective 4, because of the continuous challenges posed by covid19 we decided to develop virtual videos in replacement for workshops, we have developed the proposed series of continuing education about PRRSV and epidemiology available to the public at this link https://machado-lab.github.io/research/biosecurity/. We also included videos about machine learning to make it accessible to not-tech savvy stakeholders.
Publications
- Type:
Websites
Status:
Published
Year Published:
2022
Citation:
Machado, G., 2022. Assessing biosecurity vulnerabilities to predict the risk of new Porcine Reproductive and Respiratory Syndrome outbreaks [website]. Machado Lab. URL https://machado-lab.github.io/research/biosecurity/ (accessed 6.12.22)
|
Progress 07/01/20 to 06/30/21
Outputs
Target Audience:Dr. Machado: presented as keynote speaker at 2020 Allen D. Leman Swine Virtual Conference to more than 100 attendees on how the current project will provide the producer with a useful tool, the title of the presentation was "On the ability to predict PRRSV outbreaks at the farm-level based on biosecurity practices". 2020 AASV Annual Meeting Program, Abagael Sykes, Ph.D. student presented a posted and short talk title: "Ensemble machine learning modeling and biosecurity practices" NCVMA Swine Veterinary Seminar, November 2020 at Raleigh, NC. Machado holds meetings with NC veterinarians and pig company veterinarians. In one meeting before covid-19, the projects were presented at Clinton, NC. More than 150 farms among 15 pig companies have provided data for phase I of this projects, all received their detailed benchmarking report, similar to what we have hosted in the R package web page https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html More than 4 invited talks to pharma and pig companies have been delivered about the R package this project is developing. Changes/Problems:The impact of COVID-19 some delayed hiring the graduate students and did not allow us to visit with veterinarians and collected the full round of biosecurity data we have anticipated to collected. We had issues hiring scientists due to covid-19 restrictions in traveling. We anticipate that we will now 2021 be able to hire the necessary team members and move the project aims forward. What opportunities for training and professional development has the project provided?The NC veterinarians received training on the interpretation of the proposed tool. Dr. Machado as course coordinator lectured senior DVM students on how to use the proposed tool (course number VMP-971) The class included one day on machine learning applied to food animals. Veterinarians of the 15 companies that are participating in this project received information on how to read and use the report they received. How have the results been disseminated to communities of interest?A brief proof of concept was presented at NC Swine Vet Meeting in November 2020 at Raleigh, NC. Machado also presented an invited talk at NCVMA Swine Veterinary Seminars all online given covid-19 to pharma companies nationally and internationally. Dr. Machado and Ph.D. student Abagael Sykes presented at national conferences, Allen D. Leman Swine and AASV. We have also written a vignette hosted here https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html What do you plan to do during the next reporting period to accomplish the goals?Improve the machine learning R package website and also its efficacy in computation. Continue to collect biosecurity information from the current participants and expand. Deliver a workshop in Iowa. Make the shiny app available to the general public along with how to use the tool, under a self-contain environment that can protect user identity and confidentiality. Promote the use of the tool during the upcoming year and publish results in peer-reviewed journals
Impacts
What was accomplished under these goals?
Our specific project objectives were: Research: 1. Select and validate machine-learning algorithms to calculate the predicted risk of PRRSv introduction; 1.1. Generate a farm-level ranked list of the relative relevance of each biosecurity measure. 2. Extension: Generate farm-level predictions that can be shared with producers to promote disease management, control, and facilitate comparison between farms. 3. Develop a reporting system for producers to access their own risks. 4. Promote the use of the predicted risk of PRRSv introduction in disease management and prevention through producer-oriented extension. To address objectives 1 & 1.1. We have collected round #2 of data and were able to further improve the current machine learning approaches we have first anticipated to develop. We actually advanced the field of machine learning by providing benchmarking functionality not only to the company level, as first we proposed, we now actually can go all the way down to the individualized farm. Briefly, we can identify key biosecurity features that either contributing to a farm to reaming infected with PRRSV or to continue negative (absence of outbreaks). We created a vignette and an R package name mrIML (https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html) in the next year we will start to hold workshops not only for pig producers but given the interest in the machine learning field on our approach, we will also present the other to machine learning experts. Extension: To address objectives 2 & ultimately 3, we have developed version one of the producer-oriented repost that has been sent to participating pig production systems, as of June 2021, 150 reports have been produced. The items each producer or pig company receives are described on our R package website https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html Objective 3, the reporting systems to communicate with our public will be integrated into Machado lab-website (https://machado-lab.github.io/Software/) that will connect with all R software infrastructure, which will provide easy to use interface and fast results. An R shiny app is in development and will become available soon along with how to use videos. Finally, for objective 4, we will hold meetings at national pig conferences in the coming year, which will include a guide on how to interpret the biosecurity tool. Dr. Machado presented it during the 2020 Allen D. Leman Swine Virtual. By the end of this year, we expect to hold a workshop in Iowa during the ISU James D. Mckean Swine Disease Conference. In addition, we realized with covid-19 may impact extension activities, we will record videos about the tool and PRRSV biosecurity. The material will be distributed online via a newsletter and website (https://machado-lab.github.io/Software/). The expectation remains that this project continues to improve our national on-farm biosecurity capacity, which will protect U.S. agriculture and food supply.
Publications
- Type:
Websites
Status:
Published
Year Published:
2021
Citation:
Vignette of how to use the mrIML R package
Biosecurity working example URL https://nfj1380.github.io/mrIML/articles/Vignette_biosecurity.html (accessed 6.5.21)
- Type:
Other
Status:
Other
Year Published:
2021
Citation:
An R package named: mrIML: Multivariate (multi-response) interpretable machine learning. 2021. https://github.com/nfj1380/mrIML.
- Type:
Journal Articles
Status:
Under Review
Year Published:
2021
Citation:
Sykes, A.L., Silva, G.S., Holtkamp, D.J., Mauch, B.W., Osemeke, O., Linhares, D.C.L., Machado, G., 2021. Interpretable machine learning applied to on-farm biosecurity and porcine reproductive and respiratory syndrome virus. https://arxiv.org/abs/2106.06506
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Sykes, A.L., Silva, G.S., Holtkamp, D.J., Mauch, B.W., Osemeke, O., Linhares, D.C.L., Machado, G.,On the ability to predict the farm-level risk of PRRSV outbreaks by ensemble machine learning modeling and biosecurity practices, AASV Annual Meeting Proceedings
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2021
Citation:
Gustavo Machado, oral presentation "On the ability to predict PRRSV outbreaks at the farm level based on biosecurity practices" available at
https://www.youtube.com/watch?v=0k2tGhJujQk
|
Progress 07/01/19 to 06/30/20
Outputs
Target Audience:Machado: presented to veterinarians at the NC Swine Vet Meeting in November 2019, at Raleigh, NC. Machado attended AASV March 2020, Atlanta, presented to the National Pork Board biosecurity meeting. Machado: presented an invited talk at NCVMA Swine Veterinary Seminar, November 2019 at Raleigh, NC. Machado holds meetings with NC veterinarians and pig company COE's to engage with more pig producers at Clinton, NC. Changes/Problems:We realized early on that it would take one PRRSv cycle to collect most biosecurity data, thus we spend the first year collecting the second round of data and developing the machine learning methods further. The impact of COVID-19 some delay in hiring which may lead to a request to a no-cost extension. Also, the delay in receiving the funds has impacted the timeline. What opportunities for training and professional development has the project provided?The NC veterinarians received training on the interpretation of the proposed tool. A select curse developed by Machado offered to NCSU-DVM studentsincluded one day on machine learning applied to food-animals. How have the results been disseminated to communities of interest?A brief proof of concept was presented to NC Swine Vet Meeting in November 2019 at Raleigh, NC.Machado alsopresented an invited talk at NCVMA Swine Veterinary Seminar, November 2019 at Raleigh, NC where the approach was presented.Machado also had meetings with NC veterinarians and pig company COE'sat Clinton, NC. What do you plan to do during the next reporting period to accomplish the goals?Add machine-learning algorithms to the currently selected ones and analyze all the biosecurity plans we have collected in year one and provide an automatic and detailed report to participant farms. 2) Finalize and make codes available for the scientific community 3) Transfer the models to a web-based platform 4) Attend the North American PRRSV and CRWAD conferences in Chicago in November 2020 and IPVS 2020 in Rio November 2020. Both talks and posters will be given.5) promote the use of the tool during the upcoming year, 6) Publish results in peer-reviewed journals
Impacts
What was accomplished under these goals?
Our specific project objectives were: Research: 1. Select and validate machine-learning algorithms to calculate the predicted risk of PRRSv introduction; 1.1. Generate a farm-level ranked list of the relative relevance of each biosecurity measure. 2. Extension: Generate farm-level predictions that can be shared with producers to promote disease management, control, and facilitate comparison between farms. 3. Develop a reporting system for producers to access their own risks. 4. Promote the use of the predicted risk of PRRSv introduction in disease management and prevention through producer-oriented extension. To address objectives 1 & 1.1, we carried out the analysis of 150 breeding farms from 5 US states from which the first round of data collection has been achieved. More specifically, we discover that the variation in biosecurity among states are important and have implemented/considered this in our modeling. Objective 1 has been improved, not our ability to predict PRRSv increased by 9 points. Objective 1.1 we have expanded it based on the recent success of Machado (See paper here https://doi.org/10.1111/1365-2656.13076) in which it was refined the way each biosecurity contribution to predicting PRRSv is measured. Now weprovide an improved anddetailed report for each individual biosecurity aspect that has apositive or negative impact on the risk of a new PRRSv case. In addition, in year 2 we hope to add another approach that will allow the introduction of uncertainty to our predictions. Extension: To address objectives 2 & ultimately 3, we have developed the version one of the producer oriented repost that has been sent to participating pig production systems. This report included section named: a) Between-systems benchmarking - farms of system XX vs. all other farms in the region, b) within-system benchmarking, c) the top 20 biosecurity features by the order of relevance over the risk of PRRSv of for all farms. Objective 3, the reporting systems to communicate with our public will be integrated into Machado lab-website that will connect with all R software infrastructure, which will provide easy to use interface and fast results. We have identified the steps to make this available in the next year. Finally, for objective 4, we will hold meetings atnational pig conferences in the coming year, which will include a guide on how to interpret the biosecurity tool. We also plan to share it internationally, this will be achieved as Machado will present during IPVS 2020, in Rio. In year 2, we wish to continue applying our machine learning pipeline to identify which biosecurity gaps will need to be prioritized in order to reduce the incidence of PRRSv in the US. Another expectation is that this project continue to improve our national on-farm biosecurity capacity, which will protectU.S. agriculture and food supply.
Publications
- Type:
Conference Papers and Presentations
Status:
Submitted
Year Published:
2020
Citation:
IPVS, 2020, Rio de Janeiro, Brazil.
Tittle: On the ability to predict farm-level risk of PRRSv by ensemble machine learning modeling and biosecurity practices
Gustavo Machado*, Derald J. Holtkamp; Broc Mauch, Daniel Linhares, Gustavo S. Silva
Department of Population Heath and Pathobiology, North Carolina State University
Veterinary Diagnostic and Production Animal Medicine Department, College of Veterinary Medicine, Iowa State University
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2019
Citation:
AASV, 2019, Florida
Da Silva, G., Machado, G., Linhares, D. Investigating the biosecurity aspects related to PRRSv outbreaks.
Veterinary Diagnostic and Production Animal Medicine Department, Iowa State University, Ames, Iowa; 2
College of Veterinary Medicine, North Carolina State University, Raleigh, North Carolina
|
Progress 05/01/19 to 04/30/20
Outputs
Target Audience:Machado: presented to veterinarians at the NC Swine Vet Meeting in November 2019, at Raleigh, NC. Machado attended AASV March 2020, Atlanta, presented to the National Pork Board biosecurity meeting. Machado: presented an invited talk at NCVMA Swine Veterinary Seminar, November 2019 at Raleigh, NC. Machado holds meetings with NC veterinarians and pig company COE's to engage with more pig producers at Clinton, NC. Changes/Problems:We realized early on that it would take one PRRSv cycle to collect most biosecurity data, thus we spend the first year collecting the second round of data and developing the machine learning methods further. The impact of COVID-19 some delay in hiring which may lead to a request to a no-cost extension. Also, the delay in receiving the funds has impacted the timeline. What opportunities for training and professional development has the project provided?The NC veterinarians received training on the interpretation of the proposed tool. A select curse developed by Machado offered to NCSU-DVM studentsincluded one day on machine learning applied to food-animals. How have the results been disseminated to communities of interest?A brief proof of concept was presented to NC Swine Vet Meeting in November 2019 at Raleigh, NC.Machado alsopresented an invited talk at NCVMA Swine Veterinary Seminar, November 2019 at Raleigh, NC where the approach was presented.Machado also had meetings with NC veterinarians and pig company COE'sat Clinton, NC. What do you plan to do during the next reporting period to accomplish the goals?Add machine-learning algorithms to the currently selected ones and analyze all the biosecurity plans we have collected in year one and provide an automatic and detailed report to participant farms. 2) Finalize and make codes available for the scientific community 3) Transfer the models to a web-based platform 4) Attend the North American PRRSV and CRWAD conferences in Chicago in November 2020 and IPVS 2020 in Rio November 2020. Both talks and posters will be given.5) promote the use of the tool during the upcoming year, 6) Publish results in peer-reviewed journals
Impacts
What was accomplished under these goals?
Our specific project objectives were: Research: 1. Select and validate machine-learning algorithms to calculate the predicted risk of PRRSv introduction; 1.1. Generate a farm-level ranked list of the relative relevance of each biosecurity measure. 2. Extension: Generate farm-level predictions that can be shared with producers to promote disease management, control, and facilitate comparison between farms. 3. Develop a reporting system for producers to access their own risks. 4. Promote the use of the predicted risk of PRRSv introduction in disease management and prevention through producer-oriented extension. To address objectives 1 & 1.1, we carried out the analysis of 150 breeding farms from 5 US states from which the first round of data collection has been achieved. More specifically, we discover that the variation in biosecurity among states are important and have implemented/considered this in our modeling. Objective 1 has been improved, not our ability to predict PRRSv increased by 9 points. Objective 1.1 we have expanded it based on the recent success of Machado (See paper here https://doi.org/10.1111/1365-2656.13076) in which it was refined the way each biosecurity contribution to predicting PRRSv is measured. Now weprovide an improved anddetailed report for each individual biosecurity aspect that has apositive or negative impact on the risk of a new PRRSv case. In addition, in year 2 we hope to add another approach that will allow the introduction of uncertainty to our predictions. Extension: To address objectives 2 & ultimately 3, we have developed the version one of the producer oriented repost that has been sent to participating pig production systems. This report included section named: a) Between-systems benchmarking - farms of system XX vs. all other farms in the region, b) within-system benchmarking, c) the top 20 biosecurity features by the order of relevance over the risk of PRRSv of for all farms. Objective 3, the reporting systems to communicate with our public will be integrated into Machado lab-website that will connect with all R software infrastructure, which will provide easy to use interface and fast results. We have identified the steps to make this available in the next year. Finally, for objective 4, we will hold meetings atnational pig conferences in the coming year, which will include a guide on how to interpret the biosecurity tool. We also plan to share it internationally, this will be achieved as Machado will present during IPVS 2020, in Rio. In year 2, we wish to continue applying our machine learning pipeline to identify which biosecurity gaps will need to be prioritized in order to reduce the incidence of PRRSv in the US. Another expectation is that this project continue to improve our national on-farm biosecurity capacity, which will protectU.S. agriculture and food supply.
Publications
- Type:
Conference Papers and Presentations
Status:
Submitted
Year Published:
2020
Citation:
IPVS, 2020, Rio de Janeiro, Brazil.
Tittle: On the ability to predict farm-level risk of PRRSv by ensemble machine learning modeling and biosecurity practices
Gustavo Machado*, Derald J. Holtkamp; Broc Mauch, Daniel Linhares, Gustavo S. Silva
Department of Population Heath and Pathobiology, North Carolina State University
Veterinary Diagnostic and Production Animal Medicine Department, College of Veterinary Medicine, Iowa State University
- Type:
Conference Papers and Presentations
Status:
Accepted
Year Published:
2019
Citation:
AASV, 2019, Florida
Da Silva, G., Machado, G., Linhares, D. Investigating the biosecurity aspects related to PRRSv outbreaks.
Veterinary Diagnostic and Production Animal Medicine Department, Iowa State University, Ames, Iowa; 2
College of Veterinary Medicine, North Carolina State University, Raleigh, North Carolina
|
|