Quantication of antibiotic resistance genes and mobile genetic elements in manure from dairy farms in California

Antibiotic resistance genes (ARGs) are emerging environmental contaminants of concern to both human and animal health. Dairy manure is considered reservoir of ARGs. This study is focused on investigating prevalence of ARGs in California dairy farm manure under current common manure management. A total of 33 manure samples were collected from multiple manure treatment conditions: 1) ushed manure (FM), 2) fresh pile (FP), 3) compost pile (CP), 4) primary lagoon (PL), and 5) secondary lagoon (SL). After DNA extraction, all fecal samples were screened by PCR for the presence of 8 ARGs: four sulfonamide ARGs (sulI, sulII, sulIII, sulA), two tetracycline ARGs (tetW, tetO), two macrolide-lincosamide-streptogramin B (MLS B ) ARGs (ermB, ermF). Samples were also screened for two mobile genetic elements (MGEs) (intI1, tnpA), which are responsible for dissemination of ARGs. Quantitative PCR was then used to screen all samples for ve ARGs (sulII, tetW, ermF, tnpA and intI1). Prevalence of genes varied among sample types, but all genes were detectable in different manure types. Results showed that liquid-solid separation, piling, and lagoon conditions had limited effects on reducing ARGs and MGEs, and the effect was only found signicant on tetW (p = 0.01). Besides, network analysis indicated that sulII was associated with tnpA (p < 0.05), and Psychrobacter and Pseudomonas as opportunistic human pathogens, were potential ARG/MGE hosts (p < 0.05). This research indicated current manure management practices in California dairy farms has limited effects on reducing ARGs and MGEs. Improvement of manure management in dairy farms is thus important to mitigate dissemination of ARGs into the environment. transposons. and Anderson-Darling test of normality (α = 0.05). Tukey’s multiple comparison test was used for comparing gene levels under different conditions (α = 0.05). Multiplicity adjusted p-value was reported for each comparison. Principal Component Analysis (PCA) Plot and Hierarchical Clustering Plot were conducted by MetaboAnalyst 3.0 to nd similarity among samples. Correlation networks were created by MetScape 3.1.3 and Cytoscape 3.4.0. CorrelationCalculator 1.0.1 was used based on Debiased Sparse Partial Correlation (DSPC) method to calculate partial correlation values and p-values for each pair in the network. Range for edges was set to partial correlation values (corr. pcor) of < -0.20 or > + 0.20. to a variety of antimicrobials due to multidrug eux pumps, chromosomal mutations and the acquisition of resistance genes via horizontal gene transfer (Poole 2011). Although current animal waste treatment systems in dairy farms are not specically designed to remove ARGs, it is important to understand the potential impacts of existing manure management practices on removal of ARGs. ARGs are considered as an environmental contaminant, which may adversely impact human health. In this research, we studied the prevalence of ARGs and MGEs in ushed manure, primary lagoon manure, secondary lagoon manure, fresh pile manure, and compost pile manure. Manure samples were obtained from multiple dairy farms located in Central Valley, California. Prevalence of genes varied among sample types, but all of the studied genes were detectable in different manure types. Among ve genes quantied, only tetW was found at signicantly lower concentration in compost pile comparing with ushed manure (adj. p = 0.02) and primary lagoon samples (adj. p = 0.02). Network analysis showed that sulII was signicantly correlated with HGT by transposase gene, and certain pathogens (Psychrobacter and Pseudomonas) were potential ARG and MGE hosts (p < 0.05). Results of this study showed that ARGs are widely present in liquid (lagoon samples) and solid dairy farm manure (fresh and compost piles). Manure management such as liquid-solid separation, piling, and lagoon storage may not have signicant impacts on ARG and MGE reductions. Current manure management practices need to be improved to mitigate the transmission of ARGs into the environment.


Introduction
Antibiotic resistance is an emerging threat to public health which compromises the success and costs of therapeutic treatments (CDC 2013;Frieri et al. 2017;Pang et al. 2019;Zaman et al. 2017). In the United States, it is estimated that more than 2.8 million people are infected by antibiotic-resistant bacteria each year, and more than 35,000 died as a direct result of infection (CDC 2019). The total economic cost to the U.S. economy is estimated up to $55 billion a year due to lost wages, extended hospital stays and premature deaths (CDC 2013; Roberts et al. 2009). Use of antibiotics in animal husbandry is one of the leading factors causing the widespread antibiotic resistance (CDC 2019). Every year, over 13 million kilograms, or approximately 80% of all antimicrobial drugs are applied in animal farming to treat/prevent infectious diseases and promote animal growth in the United States (FDA 2012;FDA 2013). Out of this, approximately 70% is used for nontherapeutic purposes (UCS 2001).
Manure fertilizer is commonly used in cropland, and possible impacts of manure borne antibiotic resistance genes (ARGs) on environment are yet to be fully understood (Baquero et al. 2008;Han et al. 2018;Kumar et al. 2005;Wind et al. 2018;Zhao et al. 2017). Dairy cattle are considered potential mediators, reservoirs, and disseminators of resistant bacteria and/or ARGs (Allen et al. 2010;Guardabassi et al. 2004). When dairy manure is applied as fertilizer or directly deposited on land by grazing livestock, antibiotic resistant bacteria, ARGs and antibiotic residues may transfer into soil and ambient waterbodies such as river and lakes during rainfall and runoff events (Bennett 2008;Gogarten and Townsend 2005). Antibiotic resistance in environmental bacteria is selected by antibiotic residues or other stress environmental factors (Baquero et al. 2008;Pruden et al. 2013). Once transported, ARGs are persistent in environment for a prolonged period. They proliferate in the host bacteria, and transfer to other microbes including human pathogens through horizontal gene transfers (HGTs), which are mediated by transposons, integrons, and plasmids (Bennett 2008;Gogarten and Townsend 2005).
More than 369 million tons of manure are produced in the USA annually (USDA 2012), and majority of this quantity is used as fertilizer in cropland. Considering the amount of dairy manure produced, and its subsequent use as fertilizers, improved understanding of ARGs in manure could help in identifying the potential impacts on environment. Currently, California is the top milk producing state in the United States, and produces around 60 million tons of manure annually (USDA 2016). Flushed system is one of the most commonly adopted methods for manure handling and management in dairy farms in California due to many bene ts, including low labor, ease of handling and reduced operating cost (CARB 2017;Kaffka et al. 2016;Meyer et al. 2011). In a ushed system (Fig. 1), a dairy barn is ushed with recycled water from a lagoon, and then ushed manure passes through a solid separator, where it is separated into solid and liquid waste streams. Solid manure is piled, and in some cases, it is composted in dairy farms before applied into cropland as fertilizers. Liquid portion of manure is stored in lagoon systems for 3-6 months, and eventually it is used as fertilizers. Despite the intensive management of ushed manure, the understanding of regulating ARG in the process is still unknown.
Previous reports emphasize the importance of understanding the fate of ARGs in livestock manure treatments (Flores-Orozco et al. 2020;Gou et al. 2018;Howes 2017;Ma et al. 2018). Abundance of ARGs in livestock waste varies among farm types and locations (He et al. 2020). McKinney et al. (2010) examined the behavior of ARGs in eight livestock lagoon systems. Authors found tet and sul ARGs in chicken layer lagoons were lowest compared with lagoons of swine and dairy facilities. Hurst et al. (2019) studied the abundance of 13 ARGs in untreated manure blend pits and long-term storage systems in Northeastern U.S dairy farms, and a majority of farms use a scrape system to collect manure. The authors found ARGs abundance varied among farms, and ARG concentrations generally did not correlate to average antimicrobial usage due to complex environmental factors. Most manure conditions were simulated at the bench scale, and only few analyzed commonly used practices in dairy farms (Flores-Orozco et al. 2020;Huang et al. 2019;Pei et al. 2007;Selvam et al. 2012;Sun et al. 2016;Wang et al. 2012). Wang et al. (2012) simulated the environmental conditions of swine manure treatment by lab scale thermophilic composting and ambient temperature lagoon storage with modest aeration over a 48-day period. Authors found ve erm genes and ve tet genes dramatically declined after composting, while no signi cant reduction of erm or tet genes was observed during the lagoon treatment. Selvam et al. (2012) reported that resistance genes for tetracycline, sulfonamide, and uoroquinolone were undetectable after 28-42 days of swine manure composting in lab scale. An investigation in three pig farm wastewater treatment systems in China showed relative abundance of most ARGs were signi cantly higher in wastewater lagoon than in fresh manures even after treatment (Cheng et al. 2013). Though these studies do provide preliminary understanding, knowledge about the ARGs in California ushed system in dairy farm is yet to be understood.
A major technology to detect antibiotic resistance is by bacteria culture. However, data from current literature showed a wide range use of culture media, incubation time and antibiotic concentrations. Guidelines of standard culture for antibiotic resistance are lacking (Allen et al. 2010). Besides, many non-culturable species and unexpressed ARGs could not be identi ed (D'Costa et al. 2006;Ghosh and LaPara 2007), and these ARGs can be activated under suitable environmental conditions. It was reported only less than 1% of bacteria are culturable by standard methods (Allen et al. 2010). Knowledge of species and antibiotic resistance pro le of unculturable bacteria is lacking. Therefore, the use of culture-independent methods, such as polymerase chain reaction (PCR) and quantitative polymerase chain reaction (qPCR), has a potential to produce relatively more comprehensive and reproducible knowledge of ARG pro les.
The goal of this study was to investigate the fate of ARGs and MGEs under dairy farm manure management. The speci c objectives of this study were: 1) estimate the prevalence of ARGs and MGEs in different manure management conditions, including ushed manure, fresh pile, compost pile, primary lagoon, and secondary lagoon; 2) quantify ARGs and MGEs in dairy manure by qPCR and compare abundance of targeted genes in different manure management conditions; and 3) evaluate the relationships among ARGs, MGEs, and microbial communities by network analysis. To the author's knowledge, this study is the rst to quantify ARGs by qPCR in dairy manure from California dairy farms, and authors attempt will help improving existing understanding of ARGs in ushed dairy manure, lagoon system, and piled solid dairy manure.

Materials And Methods
Solid and liquid manure sampling in dairy farms Liquid and solid dairy manure samples were collected from multiple dairy farms in Central Valley, California. Sample information is described elsewhere and 16S rRNA gene sequencing results were published (Pandey et al. 2018). Brie y, thirty-three solid/liquid manure samples were collected from Tulare, Glenn, and Merced counties in California Central Valley, which has approximately 91 percent of dairy cows and over 80 percent of dairy facilities in California (CARB 2017). Solid samples were collected from fresh/old piles (0 to 6 months old) (n = 14), and liquid samples were collected from ushed manure pits and primary/secondary lagoons (0 to 6 months old storage) (n = 19). The solid samples collected from fresh piles (less than 2 weeks old pile) were termed as Fresh Pile (FP). The solid samples collected from old piles were termed as Compost Pile (CP). The studied CP here does not necessarily mean the piles were maintained under thermophilic temperature and mixing conditions of a standard composting process. Similarly, lagoon system in dairy farms were not necessarily under strict anaerobic environments. The liquid manure samples collected from ushed manure pit were termed as Flushed Manure (FM), while the liquid manure samples collected from primary lagoons and secondary lagoons were termed as Primary Lagoon (PL) and Secondary Lagoon (SL), respectively. In each dairy facility, 1 liter of liquid manure sample from each pond, and 600 g of solid manure from each pile were collected in sterile bottles. Samples were transported on wet ice after collection and stored at -20 ºC before DNA extraction.
DNA extraction for dairy farm manure samples Genomic DNA was extracted either using the MO BIO PowerSoil® DNA Isolation Kit or MO BIO PowerWater® DNA Isolation Kit (MO BIO Laboratories Inc.), depending on the sample consistency. All solid samples and liquid samples with turbid and sludge-like consistency were processed by the MO BIO PowerSoil® kit. For sludge-like liquid samples, 10 mL of each sample were centrifuged in 50 mL polypropylene tubes at 13,000 rpm for 10 minutes and 0.25 g of the resulting pellet was used for bead beating. Liquid samples with clear-to-low turbidity were processed by the MO BIO PowerWater® kit, and 10-200 mL of each was ltered through a Millipore lter (0.45-µm pore size). The quality and concentration of the DNA were assessed by NanoDrop 1000 spectrophotometer (Thermo Scienti c). All extracted DNA samples were stored at -20 °C before PCR ampli cation.

PCR assays for detection of resistance genes
It was reported sul, tet and erm are three of the most frequently detected ARGs classes in livestock waste, which match the major classes of antibiotics used in animal growth promotion and disease control (He et al., 2020). Primers designed in previous work targeting sul, tet and erm genes were used to amplify ARGs (Garder et al. 2014;Hu et al. 2015;Pei et al. 2007) as in Table 1. Subsequently, PCR assays were performed to determine gene detectability in the study samples. These assays were carried out using the KAPA2G Robust HotStart Ready Mix PCR Kit (KAPA) in a 25 µL volume reaction. The PCR reaction consisted of 12.5 µL 2 x KAPA2G Robust Hotstart Ready Mix, 1.25 µL 10 mM each primer, and 2 µL of the template. The temperature program consisted of initial denaturation at 95 °C, followed by 40 cycles of 15 s at 95 °C; 30 s at the 60 °C (55 °C for tetO, tetW, ermB and ermF); 30 s at 72 °C, and a nal extension step for 1 min at 72 °C. PCR products were veri ed by gel electrophoresis, puri ed, and cloned using the TOPO TA Cloning Kit (Invitrogen) according to the manufacturer's instructions. Clones were sequenced to verify the insert of the targeted gene (sequencing and veri cation results are shown in Fig. S1-S6). Plasmids carrying the target genes were extracted and used as positive controls for qPCR to generate standard curves. triplicates. The average copy and standard deviation were calculated among triplicates for each reaction. Melting curve analysis was used to detect nonspeci c ampli cation. Standard curves were included in each qPCR plate by performing serial 10-fold dilutions of the standards. The e ciency of the PCR was calculated by E ciency = 10 −(1/slope) -1. All standard curves had a r 2 > 0.99 and an ampli cation e ciency of 90-110%. The detection limit for each gene was determined by the highest dilution that produced a consistent C T value (within 5% deviation). If the standard deviation was more than 5% then two samples with the smallest difference were used for calculation.
The absolute copy number of genes was quanti ed by referring to the corresponding standard curve obtained by plotting threshold cycles versus log-copy number of genes. Levels of targeted genes were normalized as the percentage of copy number of a gene/copy number of 16S rRNA gene for each sample to emphasize the relative abundance in environmental samples (Alexander et al. 2011;Marti et al. 2013;Selvam et al. 2012).

16S rRNA gene sequencing
The high-throughput sequencing for 16S rRNA gene is described elsewhere (Pandey et al. 2018). Sequencing was performed by DNA Technologies Core Facility of the Genome Center at the University of California-Davis using the Illumina MiSeq platform. The V3 and V4 hypervariable region of the 16S rRNA gene was ampli ed using the forward primer: (5′-TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGCCTACGGGNGGCWGCAG-3′) and the reverse primer: (5′-GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGGACTACHVGGGTATCTAATCC-3′) (Klindworth et al. 2013). For quality control, barcodes and primers were allowed to have 1 and 4 mismatches, respectively. Primer sequence reads were then trimmed, and sequences were merged into a single amplicon sequence using FLASH2. Assignment of sequence to phylotypes was performed in the RDP database using the RDP Bayesian classi er (bootstrap con dence score > 50%). Further, covariates were generated using relative abundance of bacterial taxa in each sample. Stepwise discriminant analysis models built in JMP Pro 13.0 were performed until only variables with a p-value < 0.05 were retained (Pandey et al., 2018).

Data analysis
Statistical analysis on gene abundance data was performed as described in previous studies (Burch et al. 2016;Sandberg and LaPara 2016;Sun et al. 2016

PCR for gene presence
Firstly, PCR assays were applied to explore whether the gene was detectable or not in each sample. PCR screening results in all 33 samples are shown in Table 2. Four manure management groups, FP, FM, PL, and SL, had similar positive percentages of gene types. CP group had a signi cantly lower percentage (p = 0.02), with an average of 47% types of targeted genes. One sample (PL2) was found with no detection. Four samples (FP1, FP5, FP7, and SL4) were found with all ten genes. The most abundant gene was sulII, and it was present in a total of 93.9% among all samples, with a percentage of 92.9% in solid samples and 94.7% in liquid samples. The lowest one, sulA, was positive in a total of 12.1% among all samples, and was detected in 21.4% solid samples and 5.3% liquid samples. Liquid samples normally had a higher percentage of detectable genes, except for sulIII and sulA. sulII, tetW, ermF, tnpA and intI1 were selected for further study to quantify the gene concentrations because of their representation of different antibiotic resistance mechanisms and high prevalence among the samples. Fisher's exact test for contingency table analysis showed the overall gene detection rate in CP group was signi cantly lower than FP, FM, PL, and SL (p < 0.01). Quanti cation of resistance related genes Five genes (sulII, tetW, ermF, tnpA and intI1) were quanti ed by qPCR in 33 dairy manure samples taken from different manure management conditions. DNA templates for qPCR were the same batch of extractions as PCR. The numbers of copies of the ve resistance related genes quanti ed at each sample were then normalized to the number of copies of bacterial 16S rRNA gene. Data is shown in Table 3. As shown in Fig. 2 (Table 3), the average gene concentrations for sulII, tetW, and intI1 were similar (~ 1 10 −4 gene copies/16S rDNA copies). The tetW was the highest (1.43 10 −4 gene copies/16S rDNA copies). The concentrations of ermF and tnpA were 5.98 10 −6 and 4.67 10 −5 (gene copies/16S rDNA copies), respectively (lower by one and two order of magnitudes). In ordinary one-way ANOVA, diagnostic of residuals showed data passed Brown-Forsythe test and Anderson-Darling test (α = 0.05). One-way ANOVA showed manure management had no signi cant effect on four of the ve genes. Effect of manure management practices was only found signi cant on tetW (p = 0.01). Tukey test for multiple comparisons showed tetW in Compost Pile were signi cantly lower than Flushed Manure (adjusted p = 0.02) and Primary Lagoon (adjusted p = 0.02).
PCA and cluster plots for relative abundance of ve genes were drawn by MetaboAnalyst 3.0 (Xia et al. 2015) as shown in Fig. 3. Relative abundance of ve genes were log transformed and then normalized by median, followed by mean centering as the data scaling method. Figure 3 (a) shows PC 1 captured 40.1% of the variation between samples, and PC 2 captured 23.4%. These two PCs captured 63.5% of the variation between the samples. The CP, FP, PL, SL, and FM groups were overlapped, which means they were not signi cantly different from each other. In agglomerative hierarchical cluster analysis shown in Fig. 3 (b), each sample began as a separate cluster and the algorithm proceeded to combine them until all samples belonged to one cluster. Results showed that PL and FM, CP and FP were similar, as they tended to cluster together. However, different manure conditions did not fall into separate clusters, indicating their ARG pro les were not signi cantly different from each other. As the CP, FP, PL, SL, and FM groups were overlapped in Fig. 3 (a) and did not fall into separate clusters in Fig. 3 (b), it can be inferred that liquid-solid separation, lagoon system and piling process may have limited to no impacts on ARGs reductions.
Cooccurrence of ARGs, MGEs, and microbial communities

Discussion
Prevalence and quanti cation of resistance related genes The PCR results showed manure under different conditions possessed variety of ARGs and MGEs. Both traditional PCR and RT-qPCR were able to amplify DNA. RT-qPCR provided both qualitative and quantitative data by measuring the kinetics of the reaction in the exponential phase. Traditional method by agarose gels provided only qualitative results by measuring ampli cation products at endpoint of the PCR reaction (Parashar et al. 2006). In our study, targeted genes were screened rstly by PCR and selected gene were then quanti ed by RT-qPCR. It was noticed that some of genes were not detectable in PCR, and the same genes were detectable in qPCR. For example, sulII, tnpA and intI1 in PL2 were detectable in qPCR but were not detectable in PCR. This may be due to the limitation of UV visualization because some bands in agarose gels were not visible clearly under UV light. Relative abundance of intI1 in PL2 and SL5 samples was both above average in qPCR but intI1 gene in these samples was not detected in PCR.
The results showed that sulII, ermF, tnpA, and intI1 concentrations were not signi cantly different among ve manure conditions (FP, CP, FM, PL, SL), and only one gene-tetW, was found at a signi cantly lower concentration in CP compared with the FM and PL. Previous studies showed various responses of ARGs to biological conditions such as anaerobic lagoons and composting. This may be due to different experimental conditions and complex microbial ecologies involved (Pruden et al. 2013). McKinney et al. (2010) observed reductions of tet ARGs but increases of sul ARGs in anaerobic lagoons. Zhang et al. (2017) found that absolute abundances of 13 out of 14 ARGs and two integrase genes increased after 52 days of anaerobic digestion of swine manure. Sun et al. (2016) stated that 4 out of 10 detected ARGs declined during dairy manure anaerobic digestion under 20 °C. Storteboom et al. (2007) reported reduction of tetO but increase of tetW during horse manure composting process. Previous studies reported a higher decrease of cultivated antibiotic resistant bacteria in composting process compared to lagoon system (Wang et al. 2012).
It was noticed that average sulII, tetW, and intI1 concentrations identi ed in this study were lower than previous ndings. As an example, McKinney et al. (2010) reported sulII and tetW of ~ 10 − 1 and 10 − 2 copies/16S rRNA respectively in a dairy lagoon samples in Colorado. Dungan et al. (2018) reported intI1 gene copies of 10 − 2 /16S rRNA gene in the dairy wastewater in Idaho. Differences in ARG levels may be due to site-speci c physical/chemical conditions, manure handing methods, and historical intensity of antibiotic use (He et al. 2020).
However, tet and sul were reported to be the most abundant ARGs in livestock waste (He et al. 2020), which is aligned to the ndings of this study.
× × × × Cooccurrence of ARGs, MGEs, and microbial communities Network analysis indicates intI1 and Psychrobacter, ermF and Pseudomonas, were signi cantly correlated (p < 0.05). This suggests Psychrobacter was potential hosts of intI1, and abundance of ermF could be attributable to the presence of Pseudomonas.
In general, ARGs are persistent in the environment, since they not only proliferate in the host bacteria, but also transfer to other microbes and pathogens through HGT mechanisms by transposons, integrons, and plasmids (Bennett 2008;Gogarten and Townsend 2005). It was reported that integrons and transposons are responsible for the acquisition and dissemination of ARGs by HGT (Han et al. 2016;Sandberg and LaPara 2016). The sulII gene was reported on a broad-host plasmid RSF1010 (Rådström and Swedberg 1988;Yau et al. 2010). The plasmid was also found to be integrated into transposons (Cain et al. 2010). Huang et al. (2019) reported abundance of ARGs and transposase genes, which were decreased during anaerobic digestion of swine manure. Most ARGs including sul genes were signi cantly correlated with transposase genes. This study showed that tnpA and sulII abundance were signi cantly correlated, which indicates that the sulII was possibly related to the HGT by transposons. Correlations between other genes were not signi cant, and this may due to resistance genes not located in integrons/transposons and non-speci c selection agents in the manure (Andersson and Hughes 2010;Di Cesare et al. 2016;McKinney et al. 2010).
Psychrobacter species have been found in various environmental conditions including extremely low temperatures and highly saline ecosystems. These species are considered as rare opportunistic human pathogens. One of the species in Pseudomonas genus, Pseudomonas aeruginosa, is an opportunistic pathogen that causes infections in humans with a high mortality rate. Presence of ermF in Pseudomonas could compromise clinical treatment by MLS B antibiotics. Pseudomonas is resistant to a variety of antimicrobials due to multidrug e ux pumps, chromosomal mutations and the acquisition of resistance genes via horizontal gene transfer (Poole 2011).
Although current animal waste treatment systems in dairy farms are not speci cally designed to remove ARGs, it is important to understand the potential impacts of existing manure management practices on removal of ARGs. ARGs are considered as an environmental contaminant, which may adversely impact human health. In this research, we studied the prevalence of ARGs and MGEs in ushed manure, primary lagoon manure, secondary lagoon manure, fresh pile manure, and compost pile manure. Manure samples were obtained from multiple dairy farms located in Central Valley, California. Prevalence of genes varied among sample types, but all of the studied genes were detectable in different manure types. Among ve genes quanti ed, only tetW was found at signi cantly lower concentration in compost pile comparing with ushed manure (adj. p = 0.02) and primary lagoon samples (adj. p = 0.02). Network analysis showed that sulII was signi cantly correlated with HGT by transposase gene, and certain pathogens (Psychrobacter and Pseudomonas) were potential ARG and MGE hosts (p < 0.05). Results of this study showed that ARGs are widely present in liquid (lagoon samples) and solid dairy farm manure (fresh and compost piles). Manure management such as liquid-solid separation, piling, and lagoon storage may not have signi cant impacts on ARG and MGE reductions. Current manure management practices need to be improved to mitigate the transmission of ARGs into the environment.

Declarations
Ethics approval and consent to participate This article does not contain any studies with either human participants or animals. Ethical approval and consent to participate is not required. Copies of resistance related genes normalized to the number of bacterial 16S rRNA gene genes in different dairy manure. X-axis labels indicate the type of dairy treatments, rectangular boxes indicate the interquartile range of the data; median value is indicated by the horizontal line inside the box. Whiskers show min to max of data. Extreme outliers (< Q1 -3 IQ or > Q3 + 3 IQ) were removed and shown as "--" in Table 3