Progress 07/01/23 to 08/31/24

Outputs
Target Audience: Our target audiences were veterinarians and producers who are making disease management decisions. Our goal was to provide value to our future customers through the creation of better surveillance tools and improved vaccine design. The Genvax Sequence Database will allow veterinarians and producers to view and manage their diagnostic test results and pathogen sequence data. The platform will allow them to assess disease prevalence and ultimately have their choice of producer-specific vaccine products, including customized prescription saRNA-nanoparticle vaccines made by Genvax. Our collaborators and investors are deeply rooted in almost every phase of production agriculture. Our goal is to continue to develop new solutions to combat current and emerging infectious diseases in livestock. Changes/Problems:We have had two different events delay the progress of our project. The first was that initial surveillance sampling was scheduled to begin in July 2023, but was delayed until late September/early October 2023. Once Genvax became aware of the SBIR award, we contacted several potential collaborators to see if they could help with the sample collections. As part of that process, Genvax created a three-page information flier summarizing our Phase I SBIR project, the needs of the study, and the benefits the collaborators could expect. Once we had organized the sample collection process and sample submission process with the commercial sites and veterinarians, we were several months delayed. The second delay was due to negative testing results for the first 3 collection timepoints. Without positive samples we were unable to begin making a custom vaccine that was 100% matched to a sample taken from a commercial herd. We had also proposed to run an in vivo pig study with that custom vaccine. Our first positive samples were taken in January 2024 and we were several months behind our proposed timeline had to request an extension of the grant end date to accommodate the full in vivo study. The only swine IAV positive surveillance samples came from Farrow to Wean Sites. Therefore, we did not detect any viral spread between Farrow to Wean and 4-month-old pigs within a production site. When we did detect a positive sample during a collection, a majority of the samples were positive (Site 1 1/23/24 49/50+, Site 1 3/22/24 32/50+, Site 2 3/21/24 31/50+). The positive detections represented widespread infection among the animals that were sampled. In total 10 samples from each positive collection timepoint with sufficiently low CT values underwent full-length sequencing of the HA gene. All 30 HA sequences were identified as US Clade H1 alpha (Global Clade 1A.1.1). Nucleotide and amino acid sequences from all three positive collection timepoints were aligned and compared for similarity. The sequences were very similar to each other with the lowest pairwise nucleotide match percentage being 97.58% and the lowest pairwise amino acid match was 96.99%. Some of the proposed data analyses were limited to the fact that we only detected one group of highly similar strains during our surveillance testing. One additional achievement of this Phase I SBIR project that was not originally planned was that we collaborated with P3+Uplift to begin implementation of the Genvax Sequence Database 2.0 which is geared toward external users and commercialization. The new database utilized some of the existing infrastructure we had initially built during the first part of the Phase I SBIR project, but added key features that will enable vaccine ordering and tracking of the vaccine manufacturing process. This includes an outward facing customer ordering interface as well as an internal order tracking platform. The Genvax Sequence Database 2.0 will add substantial value to future customers and to Genvax. What opportunities for training and professional development has the project provided? Nothing Reported How have the results been disseminated to communities of interest?Research results from this project have been shared with Carthage Veterinary Services who collaborated on collecting swine influenza surveillance samples, Vitality Robotics (Joe Webb - project consultant), P3/Uplift who helped with additional development of the Genvax Sequence Database, and United Animal Health who is a lead investor in Genvax Technologies. What do you plan to do during the next reporting period to accomplish the goals? Nothing Reported

Impacts
What was accomplished under these goals? Objective 1: Demonstrate the feasibility of a swine influenza A surveillance and viral evolution prediction platform. How do novel IAV strains emerge and circulate within production systems? What is the prevalence/diversity of circulating swine IAV within production systems? We only had 3 positive collection timepoints and all the sequenced strains were highly related. How do flu viruses persist on farms, and how do they move within/between systems? The detected strain remained on the farm at least several months and were highly similar between sites in Illinois and Missouri. We did not detect any positive samples in 4-month-old growing pigs so there was no detected movement within a production site. Is there significant overlap of related strains found in Farrow to Wean Site 1 facilities and 4-month-old growing pigs? We did not detect any positive samples in 4-month-old growing pigs so there was no detected movement within a production site. What is the accuracy of various ML models and ensembles of models in predicting swine IAV evolution? The goal of our first ML model was to predict which swine IAV strains would persist from the previous year to the next. We tested many different ML algorithms and models and found that AdaBoost and XGBoost were the most accurate. Can we identify important metadata and ML feature sets to help improve the accuracy of swine IAV evolution predictions? We were able to create new feature sets from the sequencing data that helped to improve our ML model accuracy. Future efforts should be made to improve metadata collection to improve future ML models. To what extent does genetic diversity impact antigenic diversity? Using historical swine IAV data, we compared genetic diversity and antigenic diversity. Antigenic information is critical to understanding vaccine protection and future vaccine design. What is the antigenic diversity of circulating swine IAV and how can that help to inform us about possible protection offered from commercial vaccines and future efficacious vaccine design in possible Phase II projects? Phase II projects will include additional tools to track antigenic diversity and summarize that information for the veterinarian and producer to help with disease management strategies. Can improved producer strain surveillance data be paired with vaccine strain data to prove vaccine efficacy? Improved surveillance data will improve ML models and strain prevalence and spread information. Producer specific surveillance will allow for 100% matched pathogen specific prescription vaccines to be made. What improvements can be made to accelerate the turn-around time of sample collection, sequence deposition, and data analytics and ML outputs? We have implemented important infrastructure to accelerate the turn-around time of custom prescription vaccines that are 100% matched to the pathogen at the production site. Sequencing data that is generated is directly added to the Genvax Sequence database via an API. The database is directly connected to our manufacturing software so that fully tracked manufacturing can begin as soon as a veterinarian places an order. What are the best methods for sequence deposition into the Genvax Sequence Database? The best methods for sequence and diagnostic testing data deposition are to be directly linked to the diagnostic center generating the data. We created a system that automatically uploads data from the Iowa State University Veterinary Diagnostic Laboratory directly to the Genvax Sequence Database via an API. Can we begin to formulate a plan to improve the process via APIs? Electronic data transfer via an API will be the preferred method of data deposition into the Genvax Sequence Database. We will continue to work with diagnostic centers to implement electronic data transfer methods. Additionally, we plan to build automated methods for gathering publicly available sequencing data. Milestone: We collected 1096 nasal swab samples from two different commercial swine production sites. The only swine IAV positive samples came from Farrow to Wean Sites. Therefore, we did not detect any viral spread between Farrow to Wean and 4-month-old pigs within a production site. When we did detect a positive sample during a collection, many of the samples were positive (HPP 1/23/24 49/50+, HPP 3/22/24 32/50+, TH 3/21/24 31/50+). The positive detections represented widespread infection among the animals that were sampled. In total 10 samples from each positive collection timepoint with sufficiently low CT values underwent full-length sequencing of the HA gene. We successfully detected swine IAV at the two commercial sites when swine IAV was present. The lack of diversity found within the positive samples showed that both production sites were dealing with a group of highly related viruses that were identified as US Clade H1 alpha (Global Clade 1A.1.1). Our initial sequence database was used to train our ML models. With our first iteration of the Genvax Sequence Database we also implemented tools to look at both sequence and amino acid identity (100% match) between two sequences. We also implemented a tool to look at amino acid similarity between two sequences which gives information about how amino acid changes impact strain relatedness. We were also able to start development of data analytic tools to track the prevalence, relatedness, and evolution of their specific collected gene sequences. One of the most popular online tools for viewing and interacting with pathogen data is Nextstrain (https://nextstrain.org/). Their software provides a real-time snapshot of evolving pathogen populations and provides interactive data visualizations to virologists, epidemiologists, public health officials and citizen scientists. We used their open-source code to create a local instance of Nextstrain for swine IAV data. The goal of our first ML model was to predict which swine IAV strains would persist from the previous year to the next. We tested many different ML algorithms and models and found that AdaBoost and XGBoost were the most accurate. We were able to successfully achieve >85% accuracy and precision with several of our ML models. We were able to create new feature sets from the swine IAV HA sequencing data that helped to improve our ML models. Objective 2: Demonstrate the feasibility of a rapidly produced self-amplifying RNA nanoparticle vaccine designed from the surveillance and prediction of circulating swine IAV strains for a specific production system. Can we design, develop, and deploy a vaccine once new producer-specific HA gene sequences have been added to the database in less than 4 weeks? We successfully designed and developed a new producer specific HA saRNA in less than 4 weeks. That saRNA was later used to create a 100% matched vaccine for an in vivo pig study in April 2024. Will the vaccine produce efficacious (HI titers >=1:40) immune responses? We completed this success metric with samples from the HA treated group on study Day 53 (5/5 piglets with HI titers >= 1:40). Does vaccination result in consistent and reproducible HI titers (All HI titers either >= 1:40 or < 1:40 in each experimental group and each timepoint)? All HA treated group pigs had non-zero HI titers on study Days 42 and 53. All Day 53 samples had HI titers >= 1:40). Milestone: The success metric for efficacy of the deployed vaccine will be HI titers >= 1:40 against an autologous swine IAV strain. Overall, vaccine efficacy will be considered successful if a majority of the piglets (Ex. 3/5) show HI titers >= 1:40 for at least one of the collection dates post-vaccination. We successfully designed and developed a new producer specific HA saRNA in less than 4 weeks. That saRNA was later used to create a 100% matched vaccine for an in vivo pig study in April 2024. We completed this success metric with samples from the HA treated group on study Day 53 (5/5 piglets with HI titers >= 1:40).

Publications