Salmonella-induced microbiome profile in response to sanitation by quaternary ammonium chloride

ABSTRACT Salmonella enterica is a prominent cause of foodborne disease in the United States. However, the mechanism and route of pathogen transmission that leads to Salmonella infection in commercial processing plants are poorly understood. This study aimed to investigate the effect of mixed-species biofilms on S. enterica survival and persistence under sanitizer stress [Quaternary ammonium compounds (QACs)] by analyzing 78 floor drain samples from a meat processing facility and three S. enterica strains (serovars Cerro, Montevideo, and Typhimurium) isolated from that facility and an unrelated source. The four test groups were as follows: control, QAC treatment, Salmonella addition, and QAC treatment with Salmonella addition. DNAs were extracted, and 16S rRNA gene based on the variable region V4 amplicon sequencing was performed to analyze the relative abundance, core microbiome, and Alpha and Beta diversity using the qiime2 pipeline. At the genus level, the Brochothrix (45.56%), Pseudomonas (38.94%), Carnobacterium (6.18%), Lactococcus (4.68%), Serratia (3.14%), and Staphylococcus (0.82%) were shown to be the most prevalent in all drain samples. The results demonstrate that the relative abundance of different bacterial genera was affected by both QAC treatment and Salmonella addition, with some genera showing increases or decreases in abundance. Notably, the correlation network was constructed to understand the relationships between the different bacteria. Nitrospira had the greatest number of connections in the floor drain environment network, with two negative and eight positive correlations. The results suggest that Nitrospira in the mixed-species biofilm community may play a role in converting ammonium in the QAC sanitizer into nitrites. Thus, Nitrospira could be a potentially important genus in providing sanitizer resistance to pathogen-encompassed mixed-species biofilms. IMPORTANCE Salmonella contamination in meat processing facilities can lead to foodborne illness outbreaks. Our study characterized the microbiome dynamics in beef facility drains and their response to Salmonella addition and common sanitizer (QAC). Nitrospira could be an important genus in providing sanitizer resistance to pathogen-encompassed mixed-species biofilms. The results provide insight into the impact of mixed-species biofilms on Salmonella survival and persistence under sanitizer stress in meat processing facilities. The results highlight the need to consider mixed-species biofilm effects when developing targeted interventions to enhance food safety.

or strains are better adapted to the processing environment through biofilm formation to survive the sanitization and daily cleaning procedures and subsequently cause product cross-contamination (2).
Recent developments in genome sequencing have revolutionized our capacity to detect and identify bacteria.A common use of this technique is amplicon sequencing, which involves the extraction and sequencing of tiny DNA fragments (called amplicons).This method is often used to examine multi-species biofilms, which are challenging due to their complexity and diversity of microorganisms.Several studies have used amplicon sequencing to describe multi-species biofilms and Salmonella in contexts of food preparation (3)(4)(5).Using amplicon sequencing, Hultman et al. investigated the bacterial populations in meat processing facilities and discovered that diverse species of Leuconostoc predominated in sausages (6).Also, Benjamin et al. conducted a study to determine the types and quantities of microorganisms present at various stages of the meat-processing chain and to identify potential sites of microbial transmission in a slaughterhouse to reduce the risk of cross-contamination effectively (7).
Previous studies have shown that the synergistic and/or antagonistic interactions within the environmental multispecies community might either enhance or inhibit the colonization and tolerance of specific pathogens (8)(9)(10); thus, the composition of environmental microorganisms might pose a critical impact on pathogen tolerance and survival.Specifically, a study demonstrated enhanced tolerance to common sanitizers like QAC for some Salmonella strains in biofilms (11).Other research has revealed reduced QAC susceptibility among Salmonella from meat processing environments versus strains from other sources (2,12).Taken together, these findings suggest that Salmonella that persists in food facilities through biofilm development may acquire resistance to sanitizers over time.Therefore, the presence and the different combinations of the environmental species as well as their interaction with the unique S. enterica strains might alter their stress tolerance.
A food processing facility is a man-made environment, and floor drains are a representation of the activity in the processing facility.In a meat processing plant, the rinsing water from animal carcasses, processing equipment, and other environmental surfaces that may contain foodborne pathogens can accumulate in the drains; thus, the variety of microorganisms found in floor drains is a good representation of the bacterial species present in the processing environment (13).Characterization of the drain biofilm microbiome can thus provide critical insights into reservoirs of spoilage or pathogenic microorganisms within the facility.To further understand the impact of interspecies interactions and relative abundance alteration due to sanitization on S. enterica survival, we performed detailed bioinformatics and statistical analysis using 16S rRNA gene amplicon-based sequencing to analyze four different conditions: water-trea ted drain mixed biofilm without Salmonella addition, water-treated drain mixed biofilm with Salmonella addition, QAC-treated drain mixed biofilm without Salmonella addition, and QAC-treated drain mixed biofilm with Salmonella addition.

RESULTS
A total of 78 drain-associated biofilm samples were analyzed to assess the bacterial communities based on Salmonella addition and QAC treatment condition.A total of 7,151,472 partial 16S rRNA gene sequences were obtained after demultiplexing from the biofilm samples (Table S1) and the number of reads per sample after filtering is shown in Fig. S1.Sequences were de-replicated into unique amplicon sequence variants (ASV; traditionally referred to as OTUs) and a list of representative sequen ces was created with 10,190 features (Table S2).For a better comparison of the background microbiome before and after Salmonella addition samples, reads assigned to the Enterobacteriaceae_unassigned category were removed from the samples.This allowed us to specifically evaluate impacts on the endogenous drain biofilm community, excluding sequences from the exogenously added Salmonella strains.By strategically omitting these Salmonella sequences, we were able to scrutinize changes in the relative abundances of the native genera, without the distortion caused by the substantial number of reads from the introduced strains.After removing these reads, the final features were 9,797 (Table S2).The final findings of the taxonomic assignment were exported as a BIOM file and statistically examined.Each analysis was based on four metadata criteria: location of the floor drain, the inclusion of different S. enterica strains, the source of the S. enterica strains, and the status of QAC treatments.Each of the three S. enterica strains (serovars: Cerro, Montevideo, and Typhimurium) was introduced into drain samples, then sanitizer treatment was performed with portions tested before and after the treatment.

Core microbiome
The core microbiome consists of the bacteria that dominate a certain ecological niche or habitat.The analysis was conducted using relative abundances as the basis.The data were assessed according to the Salmonella and sanitizer conditions.Analysis of the core microbiome was performed using prevalence and relative abundance cutoff values of 20% and 0.01%, respectively.For water-treated drain mixed biofilm without Salmonella addition samples, Pseudomonas, Brochothrix, Serratia, Lactococcus, and Carnobacterium were identified as the five predominant bacterial taxa (Fig. 1a).In water-treated drain mixed biofilm with Salmonella addition samples, the decreasing order of prevalence was as follows: Brochothrix, Pseudomonas, Carnobacterium, Staphylococcus, Lactococcus, and Serratia (Fig. 1b).For QAC-treated drain mixed biofilm without Salmonella addition condition, the sample had the same microbiome as the original biofilms (Fig. 1c).However, only four taxa, Brochothrix, Carnobacterium, Pseudomonas, and Lactococcus, made up the core microbiome in QAC-treated drain mixed biofilm with Salmonella addition (Fig. 1d).

Relative abundance
The effect of QAC treatment and Salmonella addition on relative abundance changes at the genus level in drain samples (Fig. 2).The 16s rRNA amplicon sequencing data indicated that the biofilm samples from drains contained six major taxa within the genus, including Brochothrix (45.56%),Pseudomonas (38.94%),Carnobacterium (6.18%), Lactococcus (4.68%), Serratia (3.14%), and Staphylococcus (0.82%).Only the 10 most abundant genera in this data set are displayed in the graphs; the remaining genera were merged into the data set as "others." In water-treated drain mixed biofilm without Salmonella addition samples, the relative abundance of Brochothrix and Carnobacterium, two families that were present in water-treated drain mixed biofilm with Salmonella addition, increased.However, the relative abundance of Pseudomonas, Lactococcus, and Serratia decreased in the water-treated drain mixed biofilm with Salmonella addition samples.The addition of the Salmonella strain had no significant effect on the relative abundance, although Staphylococcus, which represented less than 1% of the relative abundance, appeared.The observation only occurred in water-treated drain mixed biofilm with Salmonella addition condition.In the water-treated drain mixed biofilm without Salmonella addition sample, Pseudomonas, Serratia, and Lactococcus were more abundant than Brochothrix and Carnobacterium.However, in the QAC-treated drain mixed biofilm without Sal monella addition sample, the relative abundance of Brochothrix and Carnobacterium increased.Yersinia represented 0.3% of the relative abundance in the water-treated drain mixed biofilm without Salmonella addition group but increased up to 0.8% QAC-treated drain mixed biofilm without Salmonella addition group.The relative abundance was unaffected by the inclusion of the Salmonella strain and the QAC treatment.The relative abundance of Lactococcus and Carnobacterium increased appreciably, while the quantity of Pseudomonas declined.Brochothrix's relative abundance increased.Serratia and Staphylococcus stand out as the most significant.They no longer constituted greater than 1% of the samples' relative abundance.Arthrobacter showed a relative abundance of 0.3% and increased 3.73-fold after sanitizer use.While Bacillus exhibited a 5.34-fold rise in relative abundance after sanitizer treatment, reaching 0.3% in the QAC-treated drain mixed biofilm with Salmonella addition.However, the changes were insufficient for any conclusions to be made.

Alpha diversity
Alpha diversity refers to the species diversity of a given sample or community.It measures the richness and evenness of species in a sample.Richness relates to the number of different species present in the sample, whereas evenness refers to the abundance of each species compared to the other species.The Shannon index was utilized to represent Alpha's variety because it accounts for both richness and uniformity.Alpha diversity was assessed based on QAC treatment status, where Salmonella origins and floor drain sites were controlled for.Analysis using the Shannon index revealed that the addition of Salmonella had no statistically significant (P > 0.05) impact on biodiversity (Fig. 3a).In addition, the effects on the source of Salmonella were restricted.Except for the samples including the local Cerro strain and the samples containing the Montevideo strain from an external source, which showed an increase in diversity after sanitizer treatment, the other Salmonella strain addition groups displayed a minor decrease in diversity (Fig. 3b).Comparing the four different strains of Salmonella administered before and after the sanitizer treatment to the majority of samples, there was no influence (P > 0.05) of floor location on the species diversity when comparing the before and after samples.(Fig. 3c).

Beta diversity
Beta diversity refers to the explicit comparison of microbial communities depending on their bacterial composition.It can depict the dissimilarity or distance between each pair of group members.Beta diversity is computed for each pair of samples to generate a distance or dissimilarity matrix that reflects the dissimilarity between specific samples.The data were evaluated with Principal Coordinates Analysis.The beta diversity analysis demonstrated that the community composition of samples without Salmonella addition did not alter significantly (P > 0.05) before and after sanitization.However, the QACtreated drain mixed biofilm with the Salmonella addition group exhibited substantial differences (P < 0.05) compared to the water-treated drain mixed biofilm without the Salmonella addition group (Fig. 4).

Correlation network
Correlation network analysis is a technique used to study the links among the microbial community members of an ecosystem.The association among the abundance of taxa at the genus level in the drain samples was analyzed to detect co-occurrence patterns and identify possible keystone species.The network was evaluated using the Sparse Correlation Coefficient with a threshold of >0. with Aerococcus, Tetragenococcus, and Ralstonia, as revealed by pattern search analysis of each species (Fig. 5a).The correlation among the rest of the low abundance genera is mostly positive correlations (Fig. 5b).Nitrospira had the highest number of associations (two negative and eight positive correlations) and was a possible keystone species in the biofilm community of a drain.Over 80% of the associations were positive.The most noteworthy finding was that the majority of positive correlations in the samples with the presence of Salmonella were positive.

DISCUSSION
In general, meat processing facilities utilize sanitizers that can kill 99% of microorganisms to clean their equipment.However, certain foodborne pathogens are often encased in biofilms, which are bacterial reservoirs, providing additional defense and making it harder to eradicate the bacteria and prevent contamination (14).Multispecies biofilms may also boost resilience to a range of stressors, including tolerance to sanitizers.A single resistant species also offered protection to sensitive biofilm members, making meat processing facility cleanliness even more challenging field (15).The multi-species biofilm interactions among various species of microorganisms in the bacterial commun ity substantially impact the growth capacity, survival, and, more generally, the unique behavior of each strain (16).The partnerships inside the biofilm also greatly affect their sanitizer stress tolerance.Because interactions depend on what species are within, evaluating the relationship between species composition and sanitizer tolerance may assist researchers in developing effective strategies to control food contamination in meat processing plants.Since floor drains collect all washing water and liquid wastes in the plants, the microbial communities in the floor drains reflect the microecological niches that cover the numerous microorganisms in the processing plant environment (13).To explore the influence of mixed-species biofilms by ambient microorganisms on Salmonella survival and persistence under disinfection stress, we analyzed 78 floor drain samples from a beef processing plant with three unique Salmonella prevalent serovars, two strains isolated from the local environment (Cerro and Montevideo), and one strain (Typhimurium) from a different processing plant as a comparison.In the experiment, QACs were used as the disinfectant.QACs contain cations that are attracted to negatively charged bacterial cells, so once adhered to the bacterial cell, it can degrade the cell wall and terminate the cell (17).
Our analyses revealed that QAC treatment and Salmonella introduction individu ally altered the relative abundances of predominant genera within the drain biofilm communities, though no significant changes were observed in alpha diversity.Specifi cally, the relative abundance of Brochothrix increased after QAC application, implying this genus withstands sanitization.Notably, only the samples treated with both Salmo nella addition and subsequent QAC treatment exhibited distinct community structures compared to the control groups without Salmonella (with and without QAC treatment) suggesting that the combination of Salmonella introduction followed by QAC treatment induces tangible shifts in drain biofilm composition.The lack of significant changes in alpha diversity highlights the resilience and functional redundancy inherent to complex drain biofilm communities.Our result also shows that the variety of the external Cerro and external Montevideo strains addition groups increased slightly but the diversity of the other Salmonella addition groups decreased supporting previous studies conducted on Salmonella at the beef processing facility (11).This could also be attributed to innate differences in the susceptibility of the various strains to QACs, but the sample size was too small to draw definitive conclusions.
It is worth noting that most of the samples in this study contained Brochothrix, a common spoilage bacterium known to cause premature spoilage of meat and meat products due to the production of offensive odors in refrigerated products with residual oxygen concentrations greater than 0.2% (18).However, it is non-pathogenic to humans.The presence of Brochothrix in all the samples suggests that it may be a particularly prevalent or hardy bacterium in processing and floor drain environments.In addition to Brochothrix, the samples in this study contained several other genera commonly associated with food spoilage or pathogenicity (19).Pseudomonas is a common environmental bacterium that can reduce the shelf life of foods and opportun istic pathogens (20), and its survival implies biofilms limit efficacy.
Our correlation network analysis revealed that Nitrospira exhibited the highest number of connections, mostly positive, across the various treatment groups.This suggests Nitrospira serves integral functional roles within these complex drain biofilm communities.In particular, this genus may contribute to nitrogen cycling through pathways like nitrification or denitrification, providing a nutritional source for other microbes from the decomposition of nitrogenous compounds including QAC residues (21).Nitrospira also had negative effects on two of the main genera, Pseudomonas and Serratia.The high interconnectivity of Nitrospira suggests it may serve an integral role in modulating community structure.Further studies are needed to fully understand the mechanisms underlying these correlations and how they may be influenced by various factors such as the type of bacteria, the type of stressor, and the specific conditions in the floor drain environment.
There are several limitations to our study that may have affected the results.One limitation is that we only examined the bacterial community at one time point, and it is possible that the composition and relative abundances of the bacteria present may change over time due to various environmental factors.Another limitation is that we only examined samples from one meat processing facility.While other research has shown that there were differences in the bacterial compositions of the biofilms from different facilities, it is possible that these differences may not be representative of biofilms in all meat processing environments.In terms of the perturbation treatments, we only used QAC and Salmonella as perturbation factors.Other factors that could affect the bacterial community, such as temperature, pH, and nutrient availability, were not examined in this study.In addition, we only used one type of sanitizer, QAC, and five different strains of Salmonella, so the results may vary with the use of other types of sanitizers with different functional mechanisms or other strains/serovars of Salmonella.In addition, the sample size of each group was relatively small, which may have affected the statistical power of the analysis.These factors may have influenced the results, particu larly in the alpha and beta diversity analysis.In comparison to other studies, the results of this study are consistent with the finding that Brochothrix is a common dominant genus in floor drain samples (11).This study found that QAC treatment had no significant effect on biofilm diversity which is counterintuitive.It is possible that the observed outcomes are a result of the regulatory mechanisms within the biofilm system having higher environmental stability, which allows for resistance to the stress imposed by sanitizing agents and maintenance of equilibrium within the ecosystem (22).The different methods used to analyze the data, as well as the specific strains of Salmonella used in the study, may cause different results.However, this initial study provides valuable preliminary data to guide the design of expanded investigations.To create generalizable insights about the dynamics of drain microbiomes and their responses to sanitizers that might assist in improved food safety standards across the industry, we intend to carry out a larger-scale, multivariate sampling study in the future.Further studies incorporating additional processing plants, time courses, environmental variables (pH, temperature), sanitizers, and bacterial strains are warranted.Future metabolic and microscopic analyses could also provide functional insights into biofilm interactions.
In conclusion, this study highlights that both QAC treatment and the addition of Salmonella had minimal effects on biofilm diversity when examining the microbiome of floor drain samples collected from a beef processing facility.The data indicate that the source of Salmonella had no effect on biofilm diversity but it did alter the relative abundance of species within various dominant genera.Nitrospira is an important genus within the biofilm communities and may serve as a keystone species.These insights can help inform the development of effective strategies for managing the microbiome of meat processing facilities, to minimize the risk of food spoilage and pathogen contami nation.

Sample preparation, treatment, DNA extraction, and 16S rRNA gene amplicon-based sequencing
A beef processing plant (designated Plant C) with recurring Salmonella incidence was recruited for this study.Floor drain samples were collected from the cooler (n = 42) and hotbox (n = 44) areas in Plant C. Drains selected in each area were at least 25 m apart with no shared drainage lines.All drains were located at low points of the floor for collection and removal of liquid waste, and drain selection was also based on accessibility and minimal interference with plant processing activities.To collect samples, the drain's covering grate was removed and the underside of the grate and an interior surface area of ∼500 cm 2 were vigorously swabbed with the cellulose sponge (Speci-sponge; Nasco, Atkinson, WI) to collect the attached microorganisms.Sponges were sealed in individual whirlpak bags, then transported to the laboratory on wet ice in a cooler and processed as previously described (11).
Two Salmonella strains (Cerro, Montevideo) were isolated from Plant C floor drains with an immunomagnetic concentration of Salmonella using anti-Salmonella Dynabeads (Thermo Fisher, Carlsbad, CA).These strains isolated from the local environment were confirmed by polymerase chain reaction and then serotyped using slide agglutination (O typing) and tube agglutination (flagellar H-typing) methods with commercial antisera (Difco, BD Diagnostic Systems, Sparks, MD) as previously described (11).In addition, a Salmonella Typhimurium strain was obtained from FSIS to be included in the study as a comparison since it was isolated from a different processing plant.Drain samples confirmed to be free of S. enterica were enriched in LB-NS medium for 5 days at 7°C to simulate the chilled processing plant environmental temperature and also best maintain the original microorganism composition of the samples.The enriched samples were aliquoted with 15% sterile glycerol added and stored at −20°C for further biofilm and sanitization experiments.
The biofilm formation and sanitization assays were conducted following the established protocol as described previously (11).Briefly, the above aliquoted drain samples were thawed then 1:50 diluted in LB-NS medium and grown for 5 days at 7°C.Overnight cultures of the S enterica strains were then individually added into the 5-day drain samples at a 1:100 ratio (approximately 5 × 10 6 cells/mL of Salmonella cells added into the mixture).The mixed cultures were allowed to form multispecies biofilms on stainless steel (SS) chips which were immersed and incubated in the cultures for another 5 days at 7°C.Multispecies biofilm formation by drain samples without the addition of S. enterica strains was also included in the study as control samples.
The mixed biofilms with or without Salmonella strain addition were treated with sterile water (pre-sanitization control samples) or with a QAC-based commercial sanitizer Vanquish (Dawn Chemical Corp., Milwaukee, WI) at 300 ppm for 1 min (post-sanitization samples).After treatment, the samples were processed and harvested as previously described (22).DNA extraction and purification were performed for amplicon sequenc ing based on the variable region V4 of the 16S rRNA gene.The 16S RNA sequencing was performed by Novogene with fastq files returned for analysis.

Bioinformatic analysis
FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/) was used to evaluate the quality of raw fastq data sets.Following quality control of all sequences, reads were processed using the qiime2 pipeline (https://qiime2.org/)(15).The raw fastq sequences of all samples were imported into qiime2 for taxon analysis.DADA2 was used to exclude low-quality sequences based on the quality plots (23).Given the fact that a few samples had a quality lower than Q20 around 210, 210 was chosen as the truncation length.Clean forward and backward reads were matched and merged, and the number of sequences for each OTU and the combined sample sequences was estimated.The taxonomic assignment of sample sequences was accomplished using pre-trained classifiers for the 99% OTUs in the SILVA 138 release database (https://data.qiime2.org/2021.11/common/silva13899nbclassifier.qza).The BIOM file was exported and metadata (Table S3) and taxonomic information for future statistical analysis were added.

Statistical analysis
For this data set, the sample rarefaction depth was 127,637.Samples with a feature count less than the depth of the sample rarefaction were omitted from the study.The remaining data set's relative abundance, alpha, and beta diversities were then estimated.The Tukey HSD test (24) was employed to establish the alpha diversity data's significance and the PERMANOVA to establish the significance of the Beta diversity study (25).All plots were generated using MicrobiomeAnalyst (https://www.microbiomeanalyst.ca)(26).

FIG 1
FIG 1 Core microbiome of four different floor drain groups based on Salmonella addition and sanitizer condition.Core drain microbiome profiles at the genus level in groups: (a) water-treated drain mixed biofilm without Salmonella addition, (b) water-treated drain mixed biofilm with Salmonella addition, (c) QAC-treated drain mixed biofilm without Salmonella addition, and (d) QAC-treated drain mixed biofilm without Salmonella addition.Different colors represent different levels of prevalence of the genera.Core microbiome was defined with a minimum relative abundance of 0.01% in each group.

FIG 2
FIG 2Impact of Salmonella strain addition and QAC treatment in relative abundance changes of the mixed biofilm.Relative bacterial genus abundances in response to four groupings."PRE Control" represents water-treated drain mixed biofilm without Salmonella addition samples, "PRE Salmonella" stands for water-treated drain mixed biofilm with Salmonella addition group, "POST Control" stands for QAC-treated drain mixed biofilm without Salmonella addition, and "POST Salmonella" denotes for QAC-treated drain mixed biofilm with Salmonella addition group."PRE Control" samples were the initial biofilms, which also served as the control group for this investigation.Each bar reflects the average relative abundance of different bacterial genera relative abundance within a group, and each color in the bar represents a taxonomic variation of a bacterial genus.The top 10 genera by relative abundance were displayed, while the remaining genera were merged as "others."

FIG 3
FIG 3 Alpha diversity (Shannon index) among different floor drain groups.Results of the Shannon index prior to and following sanitization based on the addition of Salmonella on (a), results (b) based on the source of Salmonella strains added to the biofilms, and results (c) indicating the location of the drain biofilm samples collected.The horizontal bars inside the boxes display the medians.The 75th and 25th percentiles are shown in the top and bottom boxes, respectively.The top and bottom whiskers, respectively, cover data within 1.5 the interquartile range of the box's top and bottom margins.

FIG 4
FIG 4 Beta diversity among different floor drain groups.Principal coordinates analysis (PCoA) based on Bray-Curtis dissimilarity and permutational multivariate analysis of variance (PERMANOVA) of the drain biofilms based on different metadata.Each dot represents a different group."Control PRE" represents water-treated drain mixed biofilm without Salmonella addition samples, "Salmonella POST" stands for water-treated drain mixed biofilm with Salmonella addition group, "Control POST" stands for QAC-treated drain mixed biofilm without Salmonella addition, and "Salmonella POST" denotes for QAC-treated drain mixed biofilm with Salmonella addition group.The ellipses covered the samples belonging to each group.

FIG 5
FIG 5 Spearman correlation analysis.Correlation analysis of taxa at the genus level using sparse correlation coefficient.Each node represents a genus colored by (a) displaying correlation for the six most core genera and (b) displaying correlation for the remaining genera.Each node represents a genus by different groups: the water-treated drain mixed biofilm without Salmonella addition (green), water-treated drain mixed biofilm with Salmonella addition (orange), QAC-treated drain mixed biofilm without Salmonella addition (purple), and QAC-treated drain mixed biofilm with Salmonella addition (magenta).The lines indicate correlations (>|0.6|) between the genera, with blue lines representing negative correlations while red lines representing positive correlations.