Non Technical Summary
Antimicrobial use (AMU) in food animals is a contributor to antimicrobial resistance (AMR) globally. The effects of AMU reduction in animal production and welfare, food safety, and environmental AMR mitigation are poorly understood. The objective of this proposal is to determine the impact of AMU on microbial diversity, prevalence of AMR genes, and genes involved in horizontal gene transfer in environmental samples from antimicrobial-free and conventional (i.e., farms that use antimicrobials as prophylactics) farms. We will also compare bird morbidity and mortality between these two categories and identify factors that contribute to health and disease prevention. This information will be used to inform producers of important factors contributing to AMR and how to control it. It will also inform the sources of AMR transmission from outside to inside poultry houses and how to improve biosecurity to avoid these events. Our specific objectives are to (1) Identify population diversity and prevalence of AMR genes in environmental samples from broiler, breeder, and pullet farms and hatcheries using metagenomic approaches; (2) Identify transmission pathways of AMR genes between sample types and location inside and outside the poultry house and between houses; (3) Identify the impact of AMU and farm practices on microbiome connections and transmission of AMR genes. This research will inform best practices to reduce AMU without compromising animal health and productivity. Mitigating AMR in poultry farms will improve food safety in the USA by avoiding farm-to-fork transmission of AMR pathogens and improving food production efficiency by reducing recalls.

Classification

Knowledge Area (KA)	Subject of Investigation (SOI)	Field of Science (FOS)	Percent
307	3299	1170	100%

Knowledge Area
307 - Animal Management Systems;

Subject Of Investigation
3299 - Poultry, general/other;

Field Of Science
1170 - Epidemiology;

Goals / Objectives
Antimicrobial resistance (AMR) is a threat to food safety and to the health of animals and humans. Decreasing the antimicrobial use (AMU) in animals can help mitigate AMR in foodborne pathogens. However, the effect of reducing AMU in poultry on the spread of AMR across the food chain and to the environment of farms is not fully elucidated. Ouroverall goalis to identify the prevalence and potential transmission pathways of AMR across the food chain and its associations with farm practices and AMU. Findings will allow for biosecurity measures to avoid the AMR spread within and between farms, reducing the risk of contamination of carcasses with foodborne pathogens carrying AMR. Thisseedproject will provide information on the AMR prevalence and transmission in poultry farms.Future projectswill build on this seed proposal and will, through extension work, evaluate the effect of improving biosecurity and flock health status in relation to the AMU and AMR at farms. In addition, future work will build on results from this proposal and measure the risk reduction of human contamination with foodborne pathogens carrying AMR due to these measures.Specific objectives:Objective 1:To identify microbial population diversity and prevalence of AMR genes (ARG) in environmental samples from hatchery, broiler, breeder, and pullet farms and from carcasses in processing plants.Objective 2:To identify transmission patterns of ARG between sample types and locations inside and outside the poultry house, between houses, and to the processing plant.Objective 3:To identify the impact of AMU and farm practices on microbiome connections and transmission of ARG.

Project Methods
Task 1.1.Sample collection:We enrolled 10% of the estimated number of antimicrobial-free and conventional poultry farms from 2 poultry-producing companies located in Alabama and Mississippi (see letters of support). Dr. Macklin is the main contact with enrolled farms. In total, approximately 4-10 farms from each farm type - pullet, breeder, and broiler - and 2 hatcheries will be sampled in the form of up to 5 chicken houses per farm. For each chicken house, soil, litter, animal feces surrounding the poultry houses, and paper used in chick transport cages will be collected. For litter collection, each barn will be divided into 4 quadrants, with 20g of litter collected from each quadrant. A total of four 20g soil samples from the surrounding of the chicken house will be collected, alongside any animal fecal samples present in the premises. Moreover, on the day of chick placement, 20 sections of 20cm2of chick cage paper will be collected from each transportation truck. For each chicken house,samples from each sample type (litter, soil, feces, and paper) will be pooled and vortexed, so that there is one representative composite sample per sample type per house. For the processing plant samples, ten poultry carcasses will be collected at the post-pick and post-chill stages. Carcasses will be individually bagged and rinsed with sterile water by shaking for 1-2 minutes. Additionally, a composite sample (10 grab samples per truck) will be collected from the transport truck. All samples will be stored at -80°C until processing.Task 1.2.Illumina Shotgun (short read) and Nanopore (long read) sequencing:DNA from each representative sample will be extracted using PowerSoil® or PowerWater®kits (Qiagen, Hilden. Germany). For shotgun sequencing, libraries from each representative sample will be prepared using Illumina DNA Prep commercial kit. Samples will be pair-end sequenced in Illumina NovaSeq (PE150) aiming at a depth of >40 million reads/sample, which is close to deep sequencing and adequate to support the proposed analysis.For long-read sequencing, libraries will be prepared using the Oxford Nanopore Technologies Ligation library preparation kit and representative samples will be sequenced using the MinION sequencer at 3X depth.Task 1.3.Microbial community and AMR gene profile reconstruction (Fig. 2):All reads will be trimmed, clipped, and low-quality and short reads will be discarded so that only good-quality reads are used. The microbial community will be reconstructed from each representative sample reads using the software MetaPhlAn 4.For AMR gene profiling, the reads of each sample type per farm will be pooled and used for a combined assembly. The resulting contigs will be mined forARGusing Hidden Markov Models (HMM) trained against AMR databases (CARDand MEGARes). These metagenomicARGwill be combined with referenceARGto form a comprehensive and specific curated AMR gene collection. For each sample, reads will be mapped against the curated AMR collection to construct the AMR profile using ShortBRED.The microbiomes and resistomes from the farm environments will be reconstructed and grouped according to sample-type and farm-type to allow for statistical comparisons using ordination, PERMANOVA, and correlation analyses between profiles across sample and farm types. Libraries will be normalized for size and depth prior to the analysis, as needed. The statistical significance threshold will be set to P≤0.05.To identify transmission patterns of ARG between sample types and locations inside and outside the poultry house, between houses, and to the processing plant. PI Santana-Pereira will perform the applicable bioinformatics analysis.Task 2.1.Create the mobile genetic profiles and compare between farms and sample types.MGEs of interest (i.e. plasmids, integrative conjugate elements - ICEs, transposons, and integrons) can be mined using recombinase genes as markers.The assembled contigs fromtask 1.3will be mined for MGE's recombinases using HMM profiles. The mined genes will be added to reference recombinases from protein databases (InterPro, Pfam, and EggNOG). This collection will be declustered and categorized by MGE family and subfamily to form a curated marker database. Reads from each sample will be mapped against this curated database to create an MGE profile using ShortBRED. The profiles will be grouped by sample-type and farm-type and used for statistical comparisonsTask 2.2.Investigate MGE and ARG transmission.The MGE, AMR, and community profiles will be used to estimate gene and taxa transfer between environments using SourceTracker.Although initially developed for bacterial community tracking, the algorithm will be modified to track AMR and recombinase gene transmission, as previously.Then plasmids will be reconstructedto establish a link between MGEs and ARG cargo, shedding light into the mechanisms of AMR transfer between environments. AMR and MGE profiles will be correlated to bacterial community profiles to identify potential microbial vectors.Task 2.3.Investigate pathogen transmission with culture-dependent analysis.To complement the metagenomic analysis, Avian PathogenicEscherichia coli(APEC) will be isolated from all samples and submitted for whole genome sequencing. Due to its genomic diversity,versatility (commensal and pathogenic bacteria for poultry and other animals, environment, and humans), and the emergence of AMR,APEC are amenable to transmission tracking and a good model organism to propose transmission pathways of AMR across the food chain. The APEC genomes will bede novoassembled and compared via whole genome alignment and SNP calling to identify lineage-specific signatures and to track APEC transmission across environments. The whole genome alignment will also be used to construct an APEC phylogenetic tree to investigate the correlation between lineages and their environments.The profiles from different farm environments will be reconstructed and grouped according to sample type and farm-type to allow for statistic comparisons using ordination, PERMANOVA, and statistical correlation analyses of microbiome and AMR profiles within each group. Libraries will be normalized for size and depth prior to the analysis, as needed.SourceTracker will estimate the contribution of the source profiles to the other environments. These analyses will be done across sample types and farms. Correlation between transferred AMR/MGEs genes and colonizing taxa to infer possible vectors will be estimated via network analysis. All statistical significance will be set to P≤0.05. APEC phylogeny and transmission will be used to corroborate the bioinformatic transmission analysis.Task 3.1.Data collection through interview.For this task, we will administer a questionnaire in interview format. We will collect information from the producer and the company manager including the amount and classes of antimicrobials used in the last 3 months, presence and diagnosis of flock disease in the last 3 months, percentage of mortality, use of feed supplement, vaccination protocols, sanitization methods used, and waste disposal methods.Task 3.2.Co-analysis of farm practices and AMR, microbiome and MGE profiles. Questionnaire data will be parsed and correlated with the profile changes observed in objectives 1 and 2.How data will be analyzed or interpreted.A stepwise regression will be used to find the best model fit, usingP≤0.01 as the variable elimination threshold. Model fit will be evaluated and compared by the adjusted R squared and by diagnostic plots visualization. The significance level will be set asP≤0.05. Differences in microbiome and AMR will be correlated with AMU to give insights into this practice's impacts in the environment and AMR selection.