Metagenomic insights into isolable bacterial communities and antimicrobial resistance in airborne dust from pig farms

This study aims to investigate bacterial communities and antimicrobial resistance (AMR) in airborne dust from pig farms. Airborne dust, pig feces and feed were collected from nine pig farms in Thailand. Airborne dust samples were collected from upwind and downwind (25 meters from pig house), and inside (in the middle of the pig house) of the selected pig house. Pig feces and feed samples were individually collected from the pen floor and feed trough from the same pig house where airborne dust was collected. A direct total bacteria count on each sampling plate was conducted and averaged. The ESKAPE pathogens together with Escherichia coli, Salmonella, and Streptococcus were examined. A total of 163 bacterial isolates were collected and tested for MICs. Pooled bacteria from the inside airborne dust samples were analyzed using Metagenomic Sequencing. The highest bacterial concentration (1.9–11.2 × 103 CFU/m3) was found inside pig houses. Staphylococcus (n = 37) and Enterococcus (n = 36) were most frequent bacterial species. Salmonella (n = 3) were exclusively isolated from feed and feces. Target bacteria showed a variety of resistance phenotypes, and the same bacterial species with the same resistance phenotype were found in airborne dust, feed and fecal from each farm. Metagenomic Sequencing analysis revealed 1,652 bacterial species across all pig farms, of which the predominant bacterial phylum was Bacillota. One hundred fifty-nine AMR genes of 12 different antibiotic classes were identified, with aminoglycoside resistance genes (24%) being the most prevalent. A total of 251 different plasmids were discovered, and the same plasmid was detected in multiple farms. In conclusion, the phenotypic and metagenomic results demonstrated that airborne dust from pig farms contained a diverse array of bacterial species and genes encoding resistance to a range of clinically important antimicrobial agents, indicating the significant role in the spread of AMR bacterial pathogens with potential hazards to human health. Policy measurements to address AMR in airborne dust from livestock farms are mandatory.


Introduction
Antimicrobial resistance (AMR) has been listed as one of the six global emerging environmental challenges by the United Nations Environment Programme (1).Transmission of AMR bacteria and their resistance determinants in environmental settings has been extensively researched in two primary habitats, aquaculture, and soils (2) e.g., lakes (3), soil with sewage and chicken manure (4), sediments, wastewater treatment plants (5), hospital wastewater (6).Prior studies have shown that ambient air contained a variety of bacteria and AMR determinants that may move long distances and across borders (7,8).These place airborne AMR bacteria and their resistance determinants as an additional important route for the spread of AMR on a continental or global scale (8,9), necessitating policies and regulation to reduce the spread of AMR through airborne dust.
WHO published a list of highly virulent and AMR bacterial pathogens as global targets for development of novel antibiotics, i.e., the ESKAPE pathogens (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp.) that are also high priority organisms for AMR monitoring due to their ability to acquire high levels of resistance.The existence and spread of airborne AMR in ESKAPE pathogens in healthcare settings has been the subject of numerous research (10).The problem is far more extensive due to their expansion in animal farms and the environment (11).Transmission of AMR bacteria, including ESKAPE organisms, from animal farms to nearby communities through the air was demonstrated (12, 13).
The role of livestock farming, in particular pig farms, as a significant source and reservoir for AMR bacteria and determinants has been scientifically demonstrated (14).It is primarily caused by a high bacterial load and an increase in antimicrobial use at the farms, which creates a significant selective antimicrobial pressure and raises the risk of bacterial resistance development and spread.In addition to meat, soil and farm wastewater, airborne dust from pig farms has reportedly contained AMR bacteria and determinants, inevitably reaching the general population (7,15).AMR bacteriafilled dust has the potential to become airborne and spread across the farms.It may also be released into the outside air by forced or natural ventilation, endangering neighboring people, farm animals, and the environment (16).To date, relatively few research has focused on AMR in ambient air, in comparison to food animals and their products.Despite the extensive antimicrobial use for a long time, little is still known about the bacterial community and AMR in airborne dust from livestock farms.
AMR studies typically rely on culture-based approaches, of which data on AMR phenotype and prevalence could be assessed.These approaches are usually time-consuming, laborious, and information limited.An advanced high-throughput technology and bioinformatic tool, Metagenomic sequencing methods, has proved the ways to overcome the limitations of the culturebased approach and to enhance the likelihood of finding AMR determinants, including novel resistance genes.It is anticipated that bacterial population and AMR traits in air samples from livestock farms would vary and be complex and therefore, culture-based methods and metagenomic sequencing analysis should be used in complementary for in-depth understanding.
The correlation between total levels of AMR and the use of antimicrobial agents was previously indicated (17).It was estimated that pigs consumed the highest antimicrobial quantity of 193 mg/PCU in 2017, accounting for 45% of the global increase in antimicrobial consumption from 2017 to 2030 (18).This is consistent with Thailand's 2020 One Health Report on Antimicrobial Consumption and AMR, which found that pigs consumed the highest amount of antibacterial agents in medicated feed (19).In addition, pig farm dust was previously shown to have higher levels of AMR bacteria and resistance genes in comparison to other livestock (20).Therefore, this study aimed to investigate bacterial communities and AMR in airborne pig farm dust.

Materials and methods . Sampling plan and location
This cross-sectional study was conducted in 9 pig farms located in Nakhon Sawan, Chainat, Supan Buri, Lopburi, and Saraburi of Thailand during the rainy season from late May to mid-October 2022 (Table 1).Samples were taken on a single visit to each farm.Due to the farm biosecurity, the selection of farms depended on the owners' willingness to participate in the study and the availability of veterinarians or animal health practitioners.All participating farms adopt a close house system with the number of pigs varying from 50 to 13,000.Three farms (Farm 4, Farm 6, and Farm 9) had large-scale commercial farming operations with between 11,000 and 13,000 pigs.Six farms were smaller scale farms, of which four farms had 2,000-5,000 pigs and another two farms had 50-200 pigs.Sampling was carried out at one pig house in each farm according to the owner of the farm's advice.For antimicrobial usage, amoxicillin was the most used antimicrobials by adding to feed either alone or in combination with tiamulin.Two farms (Farm 2 and Farm 3) declared using penicillin -streptomycin injection for treatment of ill pigs.Amoxicillin injection alone was used on one farm (Farm 2) for sick pigs.Information on the use of antimicrobial feed additives was unavailable.

. Airborne dust sample collection
Airborne dust samples were collected using a BioStage R singlestage viable cascade impactor (SKC Inc., Eighty-Four, PA, USA) with 400 pores of 0.6 µm in size and a Quick Take 30 pump with a flow rate of 28.3 L/min.Tryptic soy agar (TSA, Difco TM , MD, USA) in 100 mm × 15 mm petri dishes were placed inside the impactor for collecting dust.The impactor was positioned approximately 1.5 m above the ground, which is a person's average respiratory height (21).Airborne dust samples were taken from the selected pig house at three separate locations, inside (at the center of the pig house), upwind (25 m from the pig house), and downwind (25 m from the pig house).Sampling time was 2 min per plate.Three TSA plates were utilized to collect airborne dust samples at each location for a total of nine plates per farm.The Kestrel 3000 Weather Meter (Nielsen-Kellermen, PA, USA) was used to record the environmental variables potentially affecting the composition and diversity of bacterial communities, including .

Collection of pig feces and feed samples
Pig feces and feed samples at least 25 g of each were taken from the same pig house where airborne dust was collected.Fecal droppings were collected from the pen floor, and feed was collected from feeding trough.The samples were immediately placed in an ice box and transported to the laboratory within 6 h of collection.

. Quantification of total bacteria
Total bacteria were directly counted on TSA plates after a 24h incubation at 37 • C (22).The average colony-forming units per cubic meter of air (CFU/m 3 ) were calculated from the counting results of three plates at each sampling position.Bioaerosol concentrations were estimated using the following formula (23).
Bioaerosol concentration (CFU/m 3 ) = Number of colonies / (Adjusted flow rate of the sampling machine × Sampling duration in minutes).
A single purified colony of each bacterial species was collected from each sampling position in each pig house and stored as a 20% glycerol stock at −80 • C for further analysis.

. Antimicrobial susceptibility testing
A total of 163 bacterial isolates (n = 163) were collected (Table 4) and tested for their antimicrobial susceptibilities by broth microdilution method using the Sensititre TM Complete Automated AST System (Thermo Fisher Scientific, MA, USA).Different Sensititre TM MIC plates were used for different bacteria as follows: EUVSEC2, ASSECAF, and ASSECB for E. coli, Salmonella, Klebsiella species, and Enterobacter species; GNX2F for Acinetobacter and Pseudomonas species; EUST2 for Staphylococcus species and STP67 for Streptococcus species.All antibiotic plates were purchased from Trek Diagnostic Systems, West Sussex, UK.Two-fold agar dilution method was used for the susceptibility testing of Enterococcus species (36).The CLSI interpretive criteria was used for defining bacterial isolates as resistant or susceptible (36).E. coli ATCC 25922, P. aeruginosa ATCC 27853, S. aureus ATCC 29213 and E. faecalis ATCC 29212 served as quality control.

. Metagenomic sequencing analysis
A loopful of the pooled airborne dust samples was taken directly from the pooled three TSA plates collected at the inside position of each pig house (n = 7) before the enrichment for isolation of target bacterial species to extract genomic DNA using ZymoBIOMICS TM DNA Miniprep Kit (Zymo Research, Irvine, CA, USA) following the manufacturer's instructions.The genomic DNA from Farm 1 and 2 samples were of poor quality and unusable for Metagenomic Sequencing.The concentration and quality of the extracted DNA were measured using a NanoDrop ND 1000 spectrophotometer (Thermo Fisher Scientific).The quality and degradation were additionally evaluated by running 1 µl of the genomic DNA on 0.8% agarose gel electrophoresis stained with RedSafe nucleic acid staining solution (Thermo Fisher Scientific).The genomic DNA was submitted for Metagenomic Sequencing at Siriraj Long-read Lab (Si-LoL), Faculty of Medicine Siriraj Hospital, Mahidol University, Bangkok, Thailand, using an Oxford Nanopore sequencing platform.The Metagenomic Sequencing results were analyzed as previously described (37).Briefly, the output fastq files were uploaded to BugSeq workflow v20.07.1 (https://bugseq.com)for metagenomic classification.The reads were quality controlled with fastp v20.07.1, a FASTQ data preprocessing tool, using a minimum average read quality of Phred 7, a minimum read length of 100 bp, and the default low complexity filter.The reads were then mapped with minimap2 v2.17 to identify consensus with all microbes in the NCBI nt database.The obtained results were visualized in MultiQC reporting tool.AMR genes were identified by aligning the reads against the Resfinder (http:// genomicepidemiology.org) with minimap2.Analysis from BugSeq outputs and visualizations were performed in Pavian R package v1.2.0 in RStudio (R version 4.1.0)(https://www.r-project.org/).Kraken 2 was used to analyze bacterial species and visualize the Pavian R package results.

. Bacterial species isolated from airborne dust in pig farms
Using conventional-standard methods, the abundance and distribution of the ESKAPE species, E. coli, Salmonella and Streptococcus species were determined (Table 2).Staphylococcus (n = 37) and Enterococcus species (n = 36) were the most frequently detected bacterial species in all samples.Of all five Klebsiella isolates, two were obtained from airborne samples, while the others were isolated from feed samples.Salmonella species (n = 3) were found only in feed and feces but not in the airborne dust samples.Only one Pseudomonas isolate, which came from the feed sample, was obtained.

. Phenotypic antimicrobial susceptibilities
Overall, the bacterial isolates obtained from pig farm environment in this study exhibited various AMR phenotype and rates (Table 3).The number of target bacterial species varied greatly amongst the farms, therefore the comparison between farms is not appropriate.
High resistance rates to erythromycin (94%) and tetracycline (92%) were observed for the Enterococcus isolates from airborne dust, feed, and feces (n = 36).The Staphylococcus isolates (n = 37) showed high resistance rates to tiamulin (100%), penicillin (95%), and clindamycin (86%).Notably, Staphylococcus isolates collected from airborne dust displayed resistance to a wider range of antibiotics compared to those from feed and feces samples.  .TABLE Antimicrobial resistance rates (%) of bacterial species isolated from airborne dust, feed, and feces from the pig farms.
The only one isolate of Pseudomonas species, obtained from a feed sample of Farm 5, was resistant to a wide range of antimicrobials, including aztreonam, cefotaxime, and ciprofloxacin.Enterobacter isolates (n = 9) demonstrated resistance to most antimicrobials, including ampicillin, azithromycin, cefoxitin, chloramphenicol, and gentamicin, with high resistance rates ranging from 89% to 100%.Resistance to cefepime, cefotaxime, and ceftazidime was also observed at a lower frequency.
The bacteria isolate with various AMR phenotypes widely distributed among farms (Table 4).For each farm, the same bacterial species with the same resistance phenotype were found in airborne dust, feed, and fecal samples, for example, Enterococcus resistant to ampicillin, azithromycin, cefotaxime, ceftazidime, chloramphenicol, gentamicin, tigecycline and trimethoprim were found in airborne dust, feed and fecal samples obtained from Farm 3.

. Metagenomic characteristics of airborne dust inside pig farms . . Bacterial community compositions
Based on the Kraken 2 analysis, a total number of 1,652 bacterial species were found in all seven pig farms (Supplementary Table S1), of which Staphylococcus chromogenes was the most common bacterial species, followed by Mammaliicoccus sciuri and Staphylococcus haemolyticus.The Sankey visualization revealed that the predominant bacterial phyla observed inside pig house were Bacillota and Pseudomonadota (Figure 1).The Burkholderales order was discovered in Farms 4, 6, and 8.The relative proportions of phyla and genus of bacterial community compositions present in airborne dust from all pig farms are shown in Figure 2. The Bacillota phylum occupied the highest proportion of all farms (19.7-73.7%), of which Farm 9 had the highest percentage (73.7%).Staphylococcus was the most detected genus (50.8%).This bacterial pathogen was identified in all farms, except for Farm 4 where Acinetobacter was predominant.The number of bacterial species identified in each farm varied from 1,293 in Farm 3 to 325 in Farm 5 (Table 1).
Seventy-four resistance genes were identified in ESKAPE species, of which Staphylococcus species (32 resistance genes) and Acinetobacter species (25 resistance genes) carried most AMR genes.The mcr 6.1 gene was detected in Klebsiella species, while Pseudomonas and Enterobacter species did not host any resistance genes as indicated by Bugseq metagenomic sequencing analysis data (Supplementary Table S3).

. . Plasmid diversity
Two hundred-fifty-one plasmids were identified in all pig farms (Supplementary Table S4).Different numbers of plasmids were detected in different farms.The highest number of plasmids were detected in Farm 6 (n = 72), followed by Farm 3 (n = 56) and Farm 4 (n = 53).S. aureus (n = 48), A. baumannii (n = 19), and E. coli (n = 22) served as the most common hosts for the plasmids across all pig farms (Supplementary Table S4).No plasmids were identified in Pseudomonas and Enterobacter.

Discussion
One of the major findings in this study was the presence of various bacterial pathogens and plasmids in airborne dust within and around pig farms.The concentrations of culturable airborne bacteria in pig houses either average (4.8 10 3 CFU/m 3 ) or individual farms (1.9-11.2× 10 3 CFU/m 3 ) in this study were comparable to a study conducted in farrowing, weaning, and fattening pig houses in China (38)    Research Institute Robert Sauvé (IRSST) recommended that the concentration of total bacteria present during agricultural activity should not exceed 10,000 CFU/m 3 in the air throughout an 8-h work period (41).The bacterial concentrations in all participating pig farms, except Farm 4 (11.2 × 10 3 CFU/m 3 ), were lower than the suggested level.Farm 4 had the most total pigs (13,000 heads), as well as the most pigs (700 heads) in the sampling pen.Typically, larger pig populations in a confined space can result in higher microbial shedding, increased fecal matter and greater respiratory secretions, leading to an increased bacterial load in the air.These factors could explain the observation of the higher bacterial counts in airborne dust in Farm 4 that are greater than the advised level and those in other farms.
Staphylococcus has a remarkable ability to adapt their metabolic processes and composition to overcome obstacles, resulting in survival and persistence in an environment where many factors are unfavorable for growth and proliferation (42).Using either a culture-based approach or metagenomic sequencing analysis,

Glycopoptide bleO
Staphylococcus species were most prevalent in the airborne dust inside pig farms (Figure 1), in agreement with previous studies (43,44).In contrast, Salmonella species was isolated only from fecal and feed samples, in agreement with previous studies conducted in pig buildings in the US (45) and Eastern Canada (46).The absence of Salmonella in airborne dust samples observed was possibly due to the Salmonella level being so low that they were undetectable or the culture media being inadequate for culturing airborne samples (46).The latter emphasizes the need for the development of more efficient methods to recover cultivable airborne Salmonella species.It was previously demonstrated that Gram-positive bacteria predominated in airborne microbial contents whereas Gramnegative bacteria were in relatively low numbers (47).This is consistent with the current findings, which found that the most common species in all samples were Staphylococcus and Enterococcus species.This is likely because Gram-positive can produce spore that enhance their resistance to environmental stresses, leading to better growing in cultivation conditions (48).At the same time, Gram-negative bacteria tend to have short survival times in an airborne environment (49).
Pathogenic bacteria isolated from airborne dust in this study exhibited high AMR rates and most were resistant to multiple drugs (multidrug resistance, MDR), in agreement with previous studies (43,50).The presence of MDR bacteria in the airborne dust raises a particular concern of the risk of occupational infections, and vice versa, there is a possibility that farm workers serve as asymptomatic reservoirs of MDR bacteria (50, 51).Airborne MDR bacteria have broader public health implications, as they can spread to the community and environment, including people who live close to pig farms.
Despite the limited number, the Acinetobacter and Pseudomonas isolates displayed resistance to various antimicrobials, e.g., aztreonam, cefotaxime, and ciprofloxacin.These two bacteria are opportunistic pathogens that may inflict diseases in people with impaired lung condition.Clindamycin is commonly used for the treatment of Staphylococcal and Streptococcal pneumonia, including community-acquired methicillin-resistant S. aureus (MRSA) infections (52,53).Therefore, the presence of clindamycin-resistant Staphylococcus species in the airborne dust is of particular concern for public health.Taken together, the findings raise the alarm for the occupational and public health threat associated with airborne dust from pig farms and call for the development of effective measures to minimize the spread of AMR in farm environments.
Metagenomics provides a massive amount of data, yet a lot of it may be irrelevant or ambiguous.In this study, the combination of culture-dependent and -independent techniques was used.The total DNA for metagenomic sequencing analysis was prepared from the pooled bacteria, providing microbiome samples that mostly comprised bacteria DNA rather than other microorganisms.This combination was previously shown to improve understanding of human-associated microbial communities in relation to human health and disease (54).TSA, a non-selective media, contains nutrients required to support the growth of a wide range of bacteria and has been used for culturing bacteria from dust in several previous studies (22,43,55).The limitation was that TSA examined only culturable bacteria, resulting in the lack of non-culturable bacteria, fastidious bacteria with high nutrient requirements and bacteria with slow growth rate.The incubation conditions were aerobic; therefore, anaerobic bacteria are overlooked.
Metagenomic sequencing analysis in this study focused on the pooled bacteria collected inside each pig house to investigate the indoor microbial composition and its possible association with antimicrobial usage.Analyzing the microbiome composition of the outdoor environment of pig houses could provide insights into variations in microbial communities but the contributions of variables from other sources must be considered.
Bacillota was a major phylum in this study providing the evidence for the predominance of Gram-positive bacteria in airborne dust.This is in agreement with previous studies demonstrating the presence of the Bacillota in various atmospheric environments including animal farms (56) and animal feces served as a significant source of bacteria in this phylum (57).Therefore, the prolonged storage of manure in pits before its removal could be attributed to the higher bacteria level in the air within pig farms.
Indoor bioaerosols in food animal farms played a crucial role in the dissemination of AMR genes (12), in agreement with the observation of several genes encoding resistance to a wide range of antibiotic classes in this study (Supplementary Table S2).Aminoglycosides resistance encoding genes were at the highest abundance, in agreement with previous studies conducted in urban environment and pig farm (8,58).A previous study showed that aminoglycoside resistance genes, such as aph(3 ′ )-I, aadE, and aad were dominant in total suspended particulate samples collected from pig and chicken farms (58).The predominance of aadD, aadE, and aad(6) was previously observed in bioaerosols in chicken farms (59).High concentrations of aac6 ′ -II, aadA1, aphA3, and strB in the air of various animal farms, e.g., pig, cattle, layer, and broiler farms were recently reported (13).These studies collectively underscore the importance of airborne dust from pig farms as a significant transmission pathway for AMR genes.Interestingly, no resistance genes were detected by Bugseq Metagenomic Sequencing Analysis in Pseudomonas and Enterobacter species, which are notorious for being MDR and containing multiple AMR genes (Supplementary Table S3).This may be attributed to low DNA content of these two species in airborne samples and the lack of certain AMR genes in the database.Further studies are warranted to explore the mechanisms and implications of AMR transmission through the atmosphere in livestock settings.
Previous studies showed that the global atmosphere is being polluted by AMR genes, in particular β-lactam resistance genes (7) and the occurrence of the genes in airborne particulate matter, dust, and human airways was positively correlated with Staphylococcus spp.(60).These agree with the metagenomic sequencing results in this study.A particular concern was the existence of bla OXA genes encoding OXA-type carbapenemases known to play a crucial role in carbapenem resistance in Acinetobacter and Pseudomonas species (61).The airborne transmission of these genes possibly provides the context for the clinical significance and potential treatment challenges in the clinical setting.
Colistin is one of the highest priorities critically important antibiotics but has been extensively used in pig farming.Colistinresistant E. coli strains carrying plasmid-based mcr-1 were reported in food animals within China in 2016 (62).Since then, various mcr variants have been uncovered in livestock populations worldwide.The dissemination of colistin resistance gene including mcr-1, mcr-2 and mcr-3 genes were previously reported in Thai pig farms, of which mcr-3 was most common (63,64).The Metagenomic Sequence analysis in this study revealed several plasmid-borne colistin resistance genes (e.g., mcr-1.1,mcr-2.2,mcr-3.19,mcr-3.5, mcr-4.3, and mcr-6).To our knowledge, mcr-4.3 and mcr-6 have never been reported in pig farms.No participating farms in this study explicitly mentioned the use of colistin.It is worth noting that colistin was frequently incorporated into medicated feed for suckling and nursery piglets in Thailand as a preventive measure against gastrointestinal tract infections.In Thai pig production, an estimated 40 tons of colistin were combined with medicated feed, of which 87.2% was designated for piglets (65).Extensive use of colistin may lead to the high colistin resistance in the pig production and can spread to the other farms via environmental contamination.The results underline the importance of further studies to comprehend the dynamics and implications of colistin resistance in the interconnected ecosystems of human, animal, and environmental health and to guide the development of effective strategies to preserve the efficacy of this critically important antibiotic.
Based on the metagenomic sequencing analysis, the same plasmids were detected in multiple farms, suggesting the spread of these plasmids among farms.It is well perceived that horizontal transfer of R plasmids plays a crucial role in the wide distribution of AMR.However, in vitro horizontal transmission of R plasmids was not pursued in this study.
The plasmids pSALNBL118, pKKS49, and an unnamed plasmid (Accession no.CP027181) were detected in up to 5 farms.pSALNBL118 is a phage like plasmid originated from S. aureus strain B3-4A isolated from beef liver (66) and might be important for horizontal gene transfer (67) and transmission of virulence factors.The plasmid pKKS49 carrying apmA encoding apramycin resistance, which were originated from an MRSA ST398 isolate obtained from a dust sample taken in a holding with breeding pigs in Portugal (68).Apramycin is frequently used for treatment and control of gastrointestinal tract infection in piglets in Thai pig farms (65), which can create selective pressure leading to resistant of this antibiotic (69).However, it was not disclosed if the antibiotic was used in the participating farm in this study.In addition, the unnamed plasmid carrying ant (2")-Ia was commonly detected in gentamicin-resistant A. baumannii strain and should be further characterized due to the clinical importance of the pathogen.
The finding of mcr-harboring plasmids was limited to Farm 6, despite the farm not disclosing its use of the antibiotic.This suggests the possibility of contamination from external sources such as farm workers, contaminated feed, and water as well as the environmental contamination of AMR bacteria and genes, in addition to co-selection by other antibiotics.pCHL5009T-102k-mcr3 harboring mcr-3.5 and bla CTX−M55 was identified in E. coli.The plasmid was originally found to carry both mcr-1 and mcr-3 and isolated from E. coli in a patient in New Zealand who have experienced travel to Thailand (70).The observation of pCHL5009T-102k-mcr3 corresponded to previous studies that revealed the limited presence of mcr-1 and the predominance of mcr-3 in the E. coli isolates from healthy and sick pigs (64).Even though the reason underlying for the mcr-1 loss remains unclear, the co-localization of mcr and bla CTX−M55 on the same plasmid is an alarm for the distribution of bacterial pathogens resistant to last line antibiotics in the airborne dust in pig farm (71).
Additional limitations to this study are noted.Despite significant advancements in genomics, many resistance genes and bacterial species are still unidentified and unrepresented in databases.Furthermore, some bacteria with low biomass and resistance genes with low copies may not yield sufficient DNA copies for identifying by Metagenomic sequencing.
In conclusion, the results in this study demonstrated that the airborne dust around and within pig houses contained a wide array of microorganisms as well as AMR bacteria, posing potential risks to human, animal, and environmental health.The use of culturebased techniques in combination with metagenomic sequencing analysis enables the detection of a wider range of bacteria and their AMR genotype, providing valuable insights and providing crucial information for risk assessment, intervention planning, and informed decision-making to combat the spread of AMR.Specific guidelines to limit bioaerosol concentrations in livestock farm as well as to protect worker health are required.Future studies are suggested to examine the impact of AMR-contaminated airborne dust exposures in areas where animals are reared on the general public's health.

Bacterial
; I, Inside; D, Downwind; Fe, Feed; Fa, Feces; N, Not found.b One isolate was collected for each bacterial species from every sampling position within each pig house.Frontiers in Veterinary Science frontiersin.orgHein et al. ./fvets.

FIGURE
FIGURESankey visualization from the Kraken analysis of airborne bacterial communities inside pig farms (n = ).It is shown in di erent taxonomy levels.The higher the portion for each phylum, the higher read counts.The number indicates the read counts.D, Domain; K, Kingdom; P, Phylum; C, Class; O, Order; F, Family; G, Genus; S, Species.

FIGURE
FIGUREMetagenomic classifications of bacterial community compositions at Phylum level (A) and Genus level (B) of airborne dust inside seven pig farms byProportion (percentage of sequencing reads that align or map to a specific phyla and genus to the total number of reads) of top phyla and genus.

FIGURE
FIGURE Antimicrobial resistance genes found in airborne dust inside pig farms (n = ) by metagenomic sequencing.(A) Proportions of the number of detected genes for each antibiotic class.Proportion represents the percentage of number of detected genes for each antibiotic class to the total number of detected genes of resistance genes of di erent antibiotic classes from airborne dust collected inside the pig farms.(B) Resistance gene profile in airborne dust inside the pig farms.The stacked bar chart represents the number of detected resistance genes in each antibiotic class in each pig farm.Each colored segment within the bars corresponds to a specific resistance gene.
TABLE General details of the participating pig farms, average environmental parameters, bacterial concentrations, and number of bacterial species in airborne dust at di erent sampling positions in pig farms (n = ).
Klebsiella species, Enterobacter species, and Streptococcus species were confirmed by PCR using the following primer sets; but lower than previous studies in a nursery pig house in South Korea (1.34 × 10 5 CFU/m 3 ) (39) and in pig confinement facility (1.8 ×10 4 CFU/m 3 ) TABLE Phenotypic characteristics of antimicrobial resistance in target bacterial species isolated from airborne dust, feed, and feces in pig farms.
TABLE Antimicrobial resistance genes identified in airborne dust inside pig farms (n =).