Microbial Risk Assessment Across Diverse Environments Based on Metagenomic Absolute Quantification with Cellular Internal Standard

Microbial Risk Assessment Across Diverse Environments Based on Metagenomic Absolute Quantification with Cellular Internal Standard | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Microbial Risk Assessment Across Diverse Environments Based on Metagenomic Absolute Quantification with Cellular Internal Standard Tong Zhang, Xianghui Shi, Yu Yang, Chunxiao Wang, Xiaoqing Xu, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5150537/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 22 Apr, 2025 Read the published version in Nature Water → Version 1 posted You are reading this latest preprint version Abstract The risk posed by microorganisms in diverse environments has emerged as a significant concern. Despite this, existing microbial risk assessment frameworks often lack comprehensiveness and systematicness. To tackle this constraint, we developed a cellular spike-in (one Gram-positive and one Gram-negative bacteria) method that enables absolute quantification of microorganisms in various environmental compartments. This method was rigorously evaluated for reproducibility, accuracy, and applicability. Furthermore, we investigated biases that might arise from DNA extraction to sequencing under different cell lysis conditions for both types of bacteria, and importantly, demonstrated that this spike-in absolute quantification method could correct such biases. We then applied this method to a range of samples to determine the absolute abundance of various microorganisms, pathogens, and antibiotic resistance genes (ARGs) across eight different sample types, including influent, effluent, primary sludge, activated sludge, marine water, marine bathing beach water, marine fishery water, and river water. Based on the results, we evaluated and compared the treatment efficiencies in terms of pathogens and ARGs in five WWTPs of different operational modes. Finally, we integrated the absolute abundances of 1) total pathogens and key pathogens used for cumulative pathogenic possibility calculation in the framework of Quantitative Microbial Risk Assessment (QMRA); 2) Risk Rank1&2 ARGs and high-risk ARGs associated with ESKAPE ( Enterococcus faecium , Staphylococcus aureus , Klebsiella pneumoniae , Acinetobacter baumannii , Pseudomonas aeruginosa , and Enterobacter spp. ) + EV ( E.coli and Vibrio spp. ); 3) two most common fecal indicator bacteria (FIBs), namely Escherichia coli and Enterococci ; and 4) plasmids and other mobile genetic elements (MGEs), into an index to facilitate comprehensive microbial risk assessment and comparison across different environments. Earth and environmental sciences/Environmental sciences/Environmental impact Biological sciences/Biotechnology/Environmental biotechnology Biological sciences/Microbiology/Bacteria/Metagenomics Biological sciences/Genetics/Sequencing/DNA sequencing Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Within the context of One Health, humans, animals, and the environment are connected [ 1 ]. A more precise and comprehensive evaluation of microbial risk levels in environments can aid in minimizing human exposure to risks and enhancing health maintenance more effectively. Most existing frameworks for environmental microbial risk assessments predominantly depend on either qualitative evaluations or quantitative analyses using a single factor [ 2 , 3 ]. Qualitative risk assessments are characterized by their descriptive approaches [ 4 ], so they are constrained by subjectivity, ambiguity, and restricted decision-making support [ 5 ]. Single-factor microbial risk assessments are inherently susceptible to biases, as the outcome is directly influenced by one variable. In contrast, a multi-factor microbial risk assessment framework offers a more subtle approach, enabling comprehensive evaluation of environmental risks from diverse perspectives and mitigating biases associated with the limitations of specific factors. Various factors can be used to measure environmental risk, necessitating a holistic consideration. Fecal Indicator Bacteria (FIBs) have been used to signal the health hazards from human fecal contamination conventionally [ 6 ], but findings in research that FIBs can survive and replicate in natural environments unaffected by human activity have cast doubt on their indicative role [ 7 , 8 ]. Total number of pathogens in environments can quantify the overall threat they pose to human health, but it fails to assess the likelihood of various health outcomes—such as infections, illnesses, or deaths—caused by different doses of pathogens, as this would require specific dose-response models tailored to each pathogen. Quantitative microbial risk assessment (QMRA) can meet such requirement [ 9 , 10 ], but QMRA framework is constrained to a small subset of key pathogens [ 11 , 12 ] since only a limited number of pathogens currently have well-established dose-response models. Additionally, ARGs in the environment have become a significant threat to human health. In recent years, the misuse and overuse of antibiotics in medical, agricultural, and other environmental contexts have led to the widespread presence of ARGs in the environment [ 13 , 14 ]. These genes are persistent in environments [ 15 ] and can spread among different bacteria through horizontal gene transfer (HGT) [ 16 , 17 ], a process primarily facilitated by mobile genetic elements (MGEs). ARG transfer has been observed across phylogenetic barriers and ecological environments [ 18 , 19 ]. Under the selective pressure of antibiotics and co-selection factors, bacteria with ARGs are more likely to survive, thus facilitating the spread of these genes within bacterial communities. Simultaneously, studies showed that different ARGs posed different degrees of potential health risk [ 20 ]. Strategically concentrating on high-risk ARGs with priority is beneficial to improve the effectiveness of environmental management. It is also necessary to focus on the associations between pathogens and ARGs due to their impacts on potential risks. For instance, pathogens carrying ARGs can lead to infections that are more difficult to treat, resulting in increased mortality, extended hospital stays and elevated medical costs, which are serious concerns for public health [ 21 – 23 ]. Identifying the prevalence of ARGs associated with clinically significant pathogens can aid in assessing health risks and inform strategies for environmental management and pollution control. A risk assessment framework applicable across multiple environments relies on the absolute abundance of microorganisms to ensure comparability across different settings. Cultivation is regarded as the gold standard for determining microbial absolute abundance [ 24 , 25 ]. However, it presents several drawbacks, including time-consuming labor-intensive [ 26 ], limited specificity and sensitivity [ 25 ], and the significant challenge posed by many microbes being unculturable [ 27 ] or having stringent cultivation requirements [ 28 ]. Flow cytometry is a universally applicable and real-time method [ 29 ] for assessing microbial absolute abundance. Despite its advantages, it remains expensive and operationally complex, and is hindered by technical challenges, including difficulties in dispersing clumped cells and separating cells from other impurities. The qPCR technique often has issues of amplification bias [ 30 , 31 ]. The rapid advancement of high-throughput metagenomic sequencing offers a swift and detailed analysis of microbial composition in the environment. It can be combined with total biomass method [ 32 , 33 ] and internal standard method [ 34 – 36 ] to quantify the absolute abundance of various microorganism. Total biomass method still relies on flow cytometry or microscopy or dry weight to measure total microbial mass, and thus cannot circumvent the inherent technical shortcomings of these techniques. Conversely, internal standard (also known as “spike-in”) method, favored for its operational simplicity and wide applicability, has emerged as a pivotal approach for absolute quantification. This technique now enables quantitative analysis across diverse environments, including soil [ 36 ], sewage treatment facilities [ 37 ], swimming beach waters [ 38 ], and anaerobic reactors [ 35 ]. Initially, synthetic or exogenous DNA were used as spike-ins [ 39 ], then spike-in method has advanced from DNA-based to cellular-based approaches [ 38 ] to more accurately simulate environmental bacteria performance from pretreatment to sequencing. However, methods for absolute quantification that combine metagenomic sequencing with internal standards still need further refinement and enhancement. To simulate performance of bacteria more accurately during pretreatment, DNA extraction, and sequencing, it is crucial to develop a method using cellular internal standards of distinct cell structures (like G + and G- bacteria). Additionally, systematic assessment of the spike-in method is essential for standardizing and enhancing the adoption of cellular internal standards in absolute quantification. To address challenges mentioned above, we developed a metagenomic absolute quantification method (MAQ) using cellular internal standard (CIS). Two marine bacteria (G + and G-) were chosen as spike-ins to determine the absolute abundance of pathogens and ARGs in environments. We comprehensively evaluated the reproducibility, accuracy, applicability, and representativeness of selected spike-ins, examined the biases arising from different cell lysis conditions and evaluated the effectiveness of selected spike-ins for correcting these biases. Finally, based on the absolute abundance of pathogens, ARGs, FIBs and MGEs, we established a risk assessment framework for multiple environments. 2. Results 2.1 Accuracy, reproducibility and applicability of the MAQ methods using CIS With two spike-ins added to mock community, the absolute quantification results showed high consistency with theoretical values (Fig. 2 b). Absolute quantification of 15 species in the mock community revealed that 14 species (with genome copy numbers greater than 4×10 3 cell/mL) had an average deviation of 8.0% from their labeled values, only the species with the lowest genome copy number (3.6×10 3 cell/mL) showed a large discrepancy (90 ~ 466%) from the labeled value (Fig. 2 a, SI2 Table 4). For the three replicates, the r values between each pair of replicates calculated with absolute abundance were higher than 0.95, demonstrating high reproducibility in simple microbial communities (SI2 Table 4). This method also showed good reproducibility in complex communities. The same amounts of spike-ins were added to three replicate sub-samples of the same ST effluent samples, which were then extracted with the Powersoil Pro kit, and to three replicate sub-samples that were then extracted with the PowerWater kit, respectively. Relative standard deviation (RSD) of ALH and IMH ratio in total microorganism across three replicates extracted with the same kit ranged from 4%-14%, and the RSD of ALH/IMH across three replicates extracted with the same kit was 21% and 33% (SI2 Table 5). Using two different DNA extraction kits cause the variation of bacteria recovery percentages and their ratios. When using PowerSoil Pro kit and PowerWater kit, the ratio of ALH/IMH was 1.8 ± 0.2 and 3.1 ± 0.3 respectively (Fig. 2 c), and we chose PowerSoil Pro kit for subsequent experiments due to its better reproducibility and stability. ALH and IMH demonstrated broad applicability in various environments. Aligning the reference genomes of spike-ins to 119 metagenomes, two spike-ins accounted for low abundances of < 0.13% (Fig. 2 d, SI2 Table 6). The original amounts of ALH or IMH in these samples were significantly less than the targeted percentages of 3%~5%, making the impact of native amounts of these two spike-ins in the environmental samples negligible. 2.2 Correction of G + and G- bacteria composition using the two CIS spike-ins Technical bias could arise at any stage from DNA extraction to library construction and sequencing. Using eight aliquot sub-samples from the same ST influent sample spiked with two spike-ins, we evaluated the variations under four different DNA extraction conditions using two replicates in each condition. Ideally, the correlation coefficient R between different groups should be 1 if there is no extraction bias, however, observed R values for relative abundance of microorganism in eight aliquot sub-samples, which were subjected to different lysis conditions, ranged from 0.58 to 0.96 (Fig. 2 e, SI2 Table 7). After correction using recovery rates of the two spike-ins, R of species composition rooted at absolute abundance between pairs of samples under different lysis conditions ranged from 0.83 to 0.98 (Fig. 2 e, SI2 Table 7), indicating the biases were significantly corrected by using the two spike-ins to represent their corresponding bacteria groups and providing more accurate composition of the microbial communities. 2.3 Profiles and removal rates of pathogens and ARGs in WWTPs Across 14 samples of WWTPs, the absolute abundance of prokaryotic microorganisms ranged from 3.64×10 8 cells/L (in effluent treated by the MBR reactor) to 2.03×10 13 cells/L (in AS in the MBR reactor). AS had the highest average total cell abundance (TCA) of 3.32×10 12 cells/L (excluding AS from the MBR reactor). The average TCA of five influents was 6.31×10 11 cells/L. After primary, secondary, and tertiary wastewater treatment, TCA in effluent could be reduced to 1/7 (8.97×10 10 cells/L), 1/127 (4.95×10 9 cells/L), and 1/1700 (3.71×10 8 cells/L) of that in influent, roughly equivalent to 1 log, 2 log and 3 log reduction, respectively (SI2 Table 8). There were 60 pathogenic species across 27 genera being identified in the top 500 species (based on absolute abundances) in at least one out of the total 34 samples (including 14 WWTP samples and 20 environmental samples). The sum absolute abundance of these pathogens ranged from 1.32×10 2 cells/L to 7.03×10 10 cells/L, with WWTP influent having the highest average concentration of these pathogens at 2.33×10 10 cells/L. MBR AS and AS (excluding AS from the MBR reactor) showed pathogen abundances of 1.63×10 10 cells/L and 4.58×10 9 cells/L, respectively, and pathogen abundances of effluent from primary, secondary, and tertiary treatment were 4.07×10 9 , 1.43×10 7 and 9.88×10 6 cells/L, respectively (SI2 Table 10). There were 7 dominant pathogen genera (including Aeromonas , Acinetobacter , Aliarcobacter , Bacteroides , Mycobacterium , Pseudomonas_E , and Vibrio ) among these 60 pathogens, which accounted for 61–99% of the total pathogen abundance in WWTPs. Among these pathogenic species, Aliarcobacter cryaerophilus_A consistently exhibited higher abundance relative to other pathogens in 9 samples of total 14 WWTP samples, including all influent samples. Furthermore, Aeromonas caviae , Pseudomonas_E alcaligenes , and Vibrio fluvialis were also identified as high-abundance pathogens across WWTP samples (Fig. 3 a, SI2 Table 9, SI2 Table 10). Absolute abundance of ARGs ranged from 3.32 × 10 8 copies/L to 5.41 × 10 12 copies/L in WWTP samples. The highest average absolute abundance was observed in AS of MBR, at 1.8 × 10 12 copies/L, closely followed by influent at 1.42 × 10 12 copies/L. ARG abundance of primary, secondary and tertiary effluent stood at 1.85 × 10 11 , 4.39 × 10 9 and 3.56 × 10 8 copies/L, which was 0.88, 2.51 and 3.6 log/L reduction from influent, respectively (Fig. 3 b, SI2 Table 14). It was crucial to focus on enriched in human environments, mobile, and pathogenic ARGs classified as high-risk (Rank I and Rank II) ARGs to further refine the potential risks posed by ARGs [ 20 ]. High-risk ARGs constituted 23–33% of all ARGs in influent samples, with an average of 29%, which was the highest among all environmental types. In primary, secondary and tertiary effluent, the average proportions of high-risk ARGs were 18%, 26% and 10%, respectively. High-risk ARGs accounted for 18%, 12% and 20% in settle sludge, AS (not including MBR AS) and MBR reactor AS. Total absolute abundance of Rank I and II ARGs ranged from 1.18 × 10 7 copies/L (in MBR reactor effluent) to 1.68 × 10 12 copies/L (in influent) (SI2 Table 14, SI2 Table 16). Six high-risk ARG subtypes ( QnrS2 , ere(A) , OXA-10 , other class A beta-lactamases , AAC(6')-Ib9 , and aadA ) were present in all influent and effluent samples, with total absolute abundances of 1.92×10 10 − 2.79×10 11 copies/L in influent and 6.36×10 6 − 1.18×10 10 copies/L in effluent (Fig. 3 c, SI2 Table 13, SI2 Table 16). Employing the absolute abundance, the removal rate (expressed as log removal) was calculated and five WWTPs exhibited significant variations in pathogen removal rates due to differing wastewater treatment processes. HYW WWTP, employing the MBR process, achieved the highest average removal rate at 3.90 log/L, which was much more effective than the conventional secondary treatment plants YL and STK, with removal rates of 1.93 log/L and 1.91 log/L respectively. In contrast, Stonecutters WWTP lacked a biological treatment process and achieved only a modest average removal rate of 0.32 log/L. This WWTP method was not significantly effective for most pathogens, with log removals below 1 for 54 pathogens. Nevertheless, it achieved high removal rates for 11 Mycobacterium species, its removal rate for Mycobacterium chelonae at 2.65 log/L surpassed the average of 2.5 log/L across the five WWTPs (Fig. 3 a, SI2 Table 11). For all detected ARGs, the average removal rate across five WWTPs ranged from 0.5 log reduction in Stonecutters to 4.3 log reduction in HYW, closely matching their average removal rate for pathogens. Despite widespread presence and high abundance of MLS (macrolides-lincosamids-streptogramins) resistance genes in influents (ranging from 2×10 10 to 1.12×10 12 copies/L), this ARG type was consistently removed at rates exceeding that of average for all ARGs across five WWTPs. Aminoglycoside and sulfonamide resistance genes showed significant removal rates improvements under MBR treatment, with log reduction increasing from 2.51 ± 0.32 log and 2.21 ± 0.42 log in secondary treatment to 5.27 log and 5.47 log in MBR treatment, respectively (Fig. 3 b, SI2 Table 15). After treatment with MBR, 87% of high-risk ARG subtypes detected in influent were absent in the effluent, with no new high-risk ARG subtypes appearing. The results of tertiary treatment revealed that while 41% of high-risk ARG subtypes were not detectable in effluent, eight previously undetected subtypes emerged. The results of CEPT treatment showed that only 13% of high-risk ARG subtypes from influent was missing in the effluent, with the emergence of 12 new subtypes including in 7 ARG types (Fig. 3 d, SI2 Table 15). 2.4 Pathogens and ARGs profiles in environmental samples The absolute abundance of prokaryotic microorganisms in 20 environmental samples (including 5 RW, 5 BBW, 5 MW and 5 FW) ranged from 3.09 × 10 8 cells/L (in RW) to 1.12 × 10 10 cells/L (in MW). For these dominant pathogens (mentioned in Part 3.3), RW, BBW, MW, and FW exhibited similar total absolute quantification levels, ranging from 1.38×10 6 to 5.05×10 6 cells/L. These dominant pathogen genera in WWTP samples also dominated in three types of environmental samples with more human activity disturbance (RW, BBW and FW), accounting for 71–83%. In most WWTP samples, Aliarcobacter cryaerophilus_A , which has the highest abundance, remained the most abundant pathogen in some environmental samples, including all 5 FW, 2 RW and 1 MW. For BBW and MW, Mycobacterium chelonae and Vibrio alginolyticus were also the most dominant pathogens in 4 and 3 samples, separately (Fig. 4 a, SI2 Table 10). For absolute abundance of ARGs in 4 types of environmental samples, RW showed an average ARG absolute abundance of 9.18 × 10 8 copies/L, which was 1.6 to 3.1 times higher than that of the seawater system, including MW, BBW and FW samples (2.94 × 10 8 − 5.66 × 10 8 copies/L). The average proportion of high-risk ARGs in RW was similar to that in effluent, at 16%, with a range of 9.8%-25% across five RW samples. In MW, FW, and BBW, the average proportions of high-risk ARGs were much lower, at 3.2%, 1.7%, and 1.2%, respectively (SI2 Table 14). Risk Rank I and II ARGs included 67 ARG subtypes across 10 ARG types in 20 environmental samples, with total absolute abundance ranging from 0 copies/L in one BBW and one MW to 5.31×10 8 copies/L in RW (Fig. 4 b). Five ARG subtypes ( tet(A) , sul1 , ere(A) , OXA-10 , aadA ) were found in all RW samples, with the absolute abundance of 7.04 × 10 5 – 1.06 × 10 8 copies/L, including three belonging to the six high-risk ARG subtypes known for their existence in all influents and effluents of these five WWTPs. Meanwhile, two of these five ARG subtypes ( sul1 and aadA ) and five other ARG subtypes ( NPS-1 , APH (3”) -Ib , mef (C), mph (G) and VEB-3 ) were present in two or more seawater system samples, with the absolute abundance of 6.5 × 10 5 – 1.1 × 10 8 copies/L (Fig. 4 c). 2.5 Host tracking of ARGs We tracked the pathogenic hosts of all identified ARGs across 34 environmental samples from eight types of environments, identifying 17 ARG types carried by 97 pathogens (SI1 Fig. 1 ). We aggregated the pathogen-ARG relationships based on absolute abundance to determine eight ARG types closely associated with pathogens (with pathogen-related absolute abundances exceeding 1×10 10 copies/L), which included aminoglycoside, bacitracin, beta-lactam, MLS, polymyxin, rifamycin, sulfonamide, and tetracycline resistance genes (SI1 Fig. 1 ). More than half (55%) of aminoglycoside resistance genes carried by pathogens were located on the chromosome, which could be used for host tracking. Enterococcus_A raffinosus was identified to be the major carrier of aminoglycoside resistance genes, which all carried aad (6) and APH (3') -IIIa detected in influent (SI1 Fig. 2 a, SI2 Table 17). Beta-lactam resistance genes carried by chromosomes were linked to the broadest range of pathogens, involving 36 genes across 24 different pathogenic species. Aeromonas caviae , an emerging/re-emerging pathogen, was most closely associated, carrying five beta-lactam resistance genes totaling 9.16×10 9 copies/L in influent, effluent and RW (Fig. 5 a, SI2 Table 17). All bacitracin resistance genes associated with pathogens were bacA , predominantly carried by Aeromonas caviae and Klebsiella pneumoniae , which was detected in influent. Bacitracin carried by these two pathogens (1.00×10 10 copies/L) accounted for 66% of the total bacitracin resistance genes carried by pathogens in chromosome (1.52×10 10 copies/L) (SI1 Fig. 2 b, SI2 Table 17). Only 13% of total sulfonamide resistance genes associated with pathogens were carried by chromosome (detected in influent and effluent), which were sul1 (95%) and sul2 (5.2%), with Aeromonas caviae being the pathogens most closely linked to sul1 and all their relationship found in influent (SI1 Fig. 2 c, SI2 Table 17). Only 4.2% of rifamycin resistance genes carried by pathogens were identified in chromosome, most of these ARGs from AS sample and was found in Mycobacterium gordonae (SI1 Fig. 2 d, SI2 Table 17). For tetracycline resistance genes, pathogens belonging to Vibrio carried tet(34) and tet(35) in their chromosome, and a strong association also appeared between Prevotella disiens and tetracycline resistance genes of tet(Q) in influent and RW (SI1 Fig. 2 e, SI2 Table 17). The primary pathogen related to polymyxin resistance genes was Klebsiella pneumoniae , holding 90% of all such genes identified in chromosome of pathogens, specifically arnA and pmrF genes within polymyxin resistance genes (SI1 Fig. 2 f, SI2 Table 17). IDSA (The Infectious Diseases Society of America) has specifically highlighted the ESKAPE pathogens for their ability to evade multiple antibiotics, which were persistent in environments and frequently caused hospital-acquired infections [ 51 ]. Their transmission methods and antibiotic resistance mechanisms also merited close scrutiny [ 52 ]. Additionally, E.coli [ 51 , 53 ] and Vibrio [ 54 ] species were also recognized as significant pathogens and carriers of ARGs. The propensity for ARGs to undergo HGT within the same genus underscored the importance of monitoring ARGs in key pathogenic genera, particularly those ARGs that were both transferable and pathogenic. Across 34 environment samples, 316 ARG subtypes were carried by prokaryotes within the ESKAPE + EV related genera. Among these, tetracycline, quinolone, and multidrug were ARG types consistently associated with the ESKAPE + EV related genera across all sample types, with multidrug being unique as not all were linked to antibiotic resistance [ 55 ]. Important ARG subtypes carried by microorganism within ESKAPE + EV-related genera were selected using three criteria: being carried by three or more genera in ESKAPE + EV group, presence in three or more sample types, and relevance to ESKAPE + EV-related genera while not being classified as multidrug resistance genes. Following this selection, 23 ARG subtypes (SI1 Table 1) that met these standards were identified as ESKAPE + EV-related high-risk ARGs, serving as key indicators for risk assessment of environmental samples. In 6 ESKAPE + EV-related genera, the ARG copy number per cell was significantly higher than the average ARG copy number per cell across all microorganisms in the corresponding sample types (Fig. 5 b). And absolute abundance of ARGs carried by ESKAPE + EV related genus in influent, effluent, and RW was notably high, especially in effluent where the proportion reached 34% of all ARGs (Fig. 5 c). 2.6 Multiple environments risk index under the environmental risk assessment framework Based on absolute quantification of pathogens and ARGs in various types of environments, a comprehensive risk assessment framework was developed in this study. The framework was organized into four main sections: pathogens, fecal indicators, resistance, and mobilome, and each was divided into two sub-sections (Fig. 6 a). For pathogens, risk assessment involved evaluating total absolute abundance of pathogens (A1) and the cumulative likelihood of infection, which was calculated using dose-response models for 12 key pathogens within the QMRA framework (A2). The resistance section focused on the absolute abundance of Risk Rank I and Rank II ARGs (B1) and ESKAPE + EV-related high risk ARGs (B2). We have integrated commonly used methods for assessing environmental quality through the absolute abundance of fecal indicators (C) into our risk assessment system. The absolute abundance of plasmids (D1) and other MGEs (D2) was also considered within the risk assessment framework to evaluate the transfer and spread potential of ARGs. The final environmental risk index was derived from the arithmetic mean of the eight criteria indexes, where higher index signified greater environmental risk within this framework. Either the number of cells per 100 mL of water samples (A1, B1, B2, C1, C2, D1, D2) or the likelihood of infection per 100 mL of water samples (A2) was used as the scoring criteria (Fig. 6 a). The indexes from eight criteria demonstrated significant variability when different metrics were employed to assess the same sample (Fig. 6 b, SI2 Table 19). Except for A2, the results of the remaining risk assessment methods showed risk indexes for influent and AS were similar. The indexes for effluent are about 2 points lower than those for influent, which were slightly higher than those for RW (except for C2). A1, A2, C2, and D1 indicated that the risk indexes for RW, MW, BBW and FW were close, but the remaining four risk assessment methods showed that the risk level of RW was significantly higher than that of MW, BBW, and FW. Additionally, only B1 and B2 could clearly distinguish the risk levels of MW, BBW, and FW (Fig. 6 c). The final risk indexes across 34 samples ranged from 3.7 to 10.2. The indexes of influent, AS, and settle sludge were notably higher, varying between 7.9 and 10.2. Among the effluents, Stonecutters effluent had the highest risk index at 9.1, followed by YL effluent with the index of 6.6, moreover, effluent risk indexes for the other three WWTPs (HYW, NP, and STK), as well as most of the MW, BBW, FW, and RW samples studied, fell within the range of 4.6 to 5.7. The risk indexes for the two rivers, PR6 and KN1, both were recorded at 6.1, surpassing the average risk index of river water (5.7). Those for BBW_BB26 and MW_PM1 stood at 3.7 and 3.8, which were lower than the average indexes for their respective environmental samples, BBW and MW (Fig. 6 d). Using 34 samples to evaluate within the proposed risk assessment framework, the Pearson’s correlation coefficient (R) between the risk indexes obtained from the two assessment methods for the same module (A1&A2; B1&B2; C1&C2 and D1&D2) ranged from 0.84 to 0.97. The R value for A1/A2 and B1/B2 were relatively low (0.84 and 0.87), indicating that the two assessment methods in module A and B were more complementary. The correlation between D1/D2 was relatively high, reaching 0.97, suggesting that the two assessment methods in module D were more corroborative than complementary. Notably, the correlation between A1 and D1 was the strongest among all pairwise comparisons (R was 0.99). Similarly, the correlation between A1 and D2 was also very high at 0.96, suggesting a possible interaction between the total absolute abundance of pathogens and that of mobile elements. Additionally, the correlations between A2 and either B1 or B2 were relatively low (both R values were 0.62), indicating that there might be significant differences between the QMRA risk assessment method, which was constructed based on the pathogenicity of specific pathogens, and the risk assessment methods based on the absolute abundance of high-risk ARGs (SI2 Table 19). Therefore, these methods served as complementary approaches from different perspectives within the overall risk assessment framework. 3. Discussion The Pearson’s correlation coefficient (R) was used to characterize the variations in microbial community of the groups under different lysis conditions. The samples used in eight experimental groups were identical, theoretically, these differences should not exist. We attributed these discrepancies primarily to different sensitivities of G + and G- bacteria to cell lysis intensity and duration, which would lead to variations in community composition. In the process of absolute quantification, G + and G- bacterial spike-ins were used to simulate the performance of other G + and G- bacteria across the entire processes from DNA extraction to sequencing. The improved correlation coefficient (R) after absolute quantification confirmed the necessity of considering the differences between G + and G- bacteria in absolute quantification. This also indicated that there were limitations of past absolute quantification studies based on one spike-in [ 37 , 38 ]. Nevertheless, biases still existed in our research even after such correction (Fig. 2 e), possibly due to multiple factors. Firstly, different G + bacteria (or G- bacteria) may have varying sensitivities to lysis strength and duration, and this bias cannot be corrected in this study. Secondly, not every type of G + or G- bacteria can be better represented by the corresponding spike-in (G + or G- spike-ins), for example, with two spike-ins added into Zymo gut community mock, the real concentration of G- bacterium Salmonella enterica in this mock was consistently closer to the quantification results of the G + spike-in in three parallel experiments (SI2 Table 4). The differential lysis efficiencies and the corrective role of spike-ins for G + and G- bacteria underscored the importance of using cellular spike-ins throughout the entire process from pretreatment to sequencing in our study. This contrasted with previous studies which proposed negligible differences in extraction efficiencies between G + and G- bacteria allowed skipping the DNA extraction step when using spike-ins to simulate other bacterial behaviors [ 56 ]. Our experimental results revealed that the extraction efficiencies of two spike-ins only approximated each other under specific lysis intensities and durations, with the ratio of G + to G- spike-ins increasing as lysis intensity and time increase. We also found that even when using sufficiently intense and prolonged lysis conditions to fully release G + bacteria, as previously mentioned [ 38 , 57 ], it was difficult to ensure the recovery ratio of G + to G- bacteria remained stable. Under lysis conditions optimized to fully release G + bacteria, DNA of G- bacteria became fragmented due to excessively shearing, which leaded to greater loss of G- bacteria in Nanopore sequencing, resulting in lower recovery rate for G- bacteria compared to G + bacteria (SI2 Table 7). The absolute quantification units for microorganisms included DNA mass per sample mass (soil) or volume (water and air) [ 58 ], cells per sample mass or volume [ 32 ], and gene copies per sample mass or volume [ 36 , 39 ]. Absolute abundance could assign practical significance to microbial abundances, revealing the real composition and distribution of taxa in various environments, thereby enables inter-samples comparisons rather than merely intra-sample comparisons [ 35 , 37 ]. While previous metrics for describing ARG abundance included ARG copies/cell, ARG copies/genome, ARG density, ARG copies/16S rRNA gene, RPKM, coverage, PPM, the consensus has shifted towards using copies/cell to describe the average number of ARG copies carried by each cell [ 59 ]. Based on absolute quantification methods now, the unit of ARG copies/sample mass or volume was widely accepted and used. Different units for describing ARG abundance could help us understand ARG-related issues of various aspects. According to previous reports, the relative abundance of ARGs in AS was typically 0.2–0.4 copies/cell, significantly lower than 1-2.5 copies/cell typically found in influent of most WWTPs [ 60 ], indicating the higher prevalence of resistance genes in influent compared to AS. However, our studies showed that the absolute abundance of ARGs in AS was about an order of magnitude higher than that in influent, revealing more higher ARG level in AS per sample volume. Using the absolute quantification methods, we compared removal efficiencies of pathogens and ARGs in WWTPs of different treatment process and the results are highly expected, indicating tertiary treatment > secondary > primary. However, the specific removal efficiencies of each species or ARGs may be biased due to some limitations of this study, including a single sampling event and not considering the process of disinfection. Based on absolute quantification, a comprehensive and multi-environmental risk assessment system could enhance the comprehension of risk by 1) reducing bias and broaden information, better than a single assessment method which might rely too heavily on solely data sources, such as using E. coli as the single pathogen indicator in bathing beach water, overlooking other pathogenic populations; 2) improving the applicability, better than a single standard which might not be applicable for all environmental samples of high levels of complexity and matrix effect. We have constructed an environmental risk assessment system centered around pathogens, FIBs, ARGs, and mobilome in environments. For the pathogen part, we considered total abundance of pathogens without distinguishing varying pathogenicity of different pathogens. Additionally, we assessed the environmental risk of pathogens using QMRA framework, which calculated the cumulative likelihood of disease based on distinct dose-response models for 12 key pathogens. Each method had its strengths and weaknesses. The former offered a broad overview but lacked precision, while the latter allowed for individual analysis of each pathogen but was constrained by the lack of dose-response data for most pathogens. FIBs were commonly used to assess health risks in recreational waters due to their close association with human diseases [ 7 ], however, their delayed response in risk management decisions and the limitations in indicating threats were widely recognized [ 61 ]. We employed spike-in combined with Nanopore sequencing for quantitative measurement of FIBs in samples, replacing traditional culturing to improve the timeliness of risk assessment. Enterococci and E.coli were used as independent environmental risk criteria due to different suitability under different environmental conditions. For instance, Enterococci had a greater survival ability than E.coli in colder conditions [ 62 , 63 ], while E.coli was more suitable for freshwater environments compared to Enterococci [ 64 ]. The risk assessment of resistance focused on high-risk ARGs (Risk Rank 1&2) [ 20 ] and ARGs associated with the genera of key pathogenic bacteria, known as ESKAPE + EV (specifically described in SI1 S2). The former was based on the inherent characteristics of ARGs, and the latter focused on ARGs associated with multi-resistant key pathogens appeared in multiple environments. Additionally, HGT of most ARGs was related to the conjugation, transformation and transduction process of plasmids [ 65 ]. At the same time, MGEs like transposons, integrons, and insertion sequences, also promoted the spread of ARGs within bacterial populations and environment [ 66 ]. The absolute abundance of the mobilome (including plasmids and MGEs) could reflect the risk of antibiotic resistance dissemination to the same extent. The globally utilized Air Quality Index (AQI) converted measured concentrations of six major air pollutants (PM2.5, PM10, SO2, NO2, O3, CO) into individual air quality indexes (IAQI), and the highest IAQI among these pollutants determined the AQI [ 67 ]. In Hong Kong, the River and Stream Water Quality Index (WQI) comprised three assessment indicators: DO, BOD 5 , and ammonia nitrogen, each indicator was scored according to its specific scoring criteria (scores range from 1 to 5), with equal weighting given to all parameters [ 68 ]. The total score of these three parameters formed the monthly water quality index. Similarly, our risk assessment framework aimed to transform complex absolute abundance values into a simple index. In the conversion processes for A1, B1, B2, C1, C2, D1 and D2, the logarithm of absolute abundance is directly used as the final risk index, ranging from 0 to 13.22, and for A2, pathogenicity served as the scoring criterion, with scores ranging from 0.08-12 based on the absolute value of lg(1-cumulative pathogenic potential). Equal weight was given to the eight scores, and the average of these eight risk indexes constituted the final Microbial Risk Index (MRI). Furthermore, we explored the contribution of each risk assessment method to the final overall risk index by calculating the difference between the risk index after excluding certain assessment method and the final overall risk index. The results showed that after excluding part D, the average deviation between the risk index of the 34 environmental samples and the actual risk index was the largest, at 20%. Excluding the sub-standards D1 and D2 individually resulted in average deviations of 8.7% and 8.7%, respectively, indicating that part D had a significant contribution to the final result, while D1 and D2 contributing similarly. Excluding A2 resulted in an average deviation of 11% between the obtained environmental risk index and the actual risk index, which was the highest among the eight scoring methods and close to the average deviation when the whole part A was excluded (12%), indicating that A2 had a high contribution to the overall risk assessment system and was unique and irreplaceable within this framework. In contrast, the exclusion of A1 only resulted in an average deviation of 1.6%, indicating that other assessment methods within this risk assessment system produced results similar to those obtained by A1, so most environmental samples could still yield a similar risk index after excluding A1. However, the situation of some environmental samples may be different from the majority, such as BBW_BB26 and MW_PM1, where the deviations between the index obtained after excluding A1 and the final overall risk index reached 8.9% and 8.3%, far exceeding the average level (1.6%). Meanwhile, their dependency on part B and D (33%, 33% and 43%, 43%) also far exceeded the average level (6.5% and 20%). In contrast, the influence of C2 on the final risk index for these two environmental samples (0.79% and 0.76%) was much lower than the average level (4.6%) (SI2 Table 19). Our research has made several improvements compared to previous studies based on spike-in for absolute quantification. Initially, there was no need to use marker genes to differentiate post-added spike-ins from existing species within the community. Instead, we chose two tailored cellular spike-ins, reducing the data loss of bacteria related to marker genes [ 38 ]. Besides, our research detailed reported the lysis process significantly affected the extraction efficiency of G + and G- bacteria and using our spike-ins could be effective in correcting the proportional biases of G + and G- bacteria. Additionally, a multi-environmental risk assessment framework was built based on the absolute quantification results of pathogens, FIBs, ARGs and MGEs. However, we recognized that our study had certain limitations: 1) due to differences among various G + or G- bacteria, we could not completely rule out the possibility that selected spike-ins might not adequately represent certain bacteria, which could lead to significant biases in absolute quantification. 2) Both two spike-ins used in this study and previously used marker gene spike-ins [ 38 ] did not perform well in quantifying species of low abundance. In the future, we could consider customizing spike-ins tailored to the characteristics of bacteria from different phyla, in hopes of more closely simulating the recovery rates of various bacteria and striving to address the issue of inaccurate quantification of some low-abundance species. 3) We used Kraken2 to classify microorganism directly from raw reads obtained by Nanopore sequencing, despite the ongoing efforts to improve accuracy in Nanopore sequencing technologies, classifying closely related taxa still had challenges. Furthermore, Kraken2 always tended to overestimate species diversity. 4) This study did not involve research on eukaryotes and phages in environments, but their impact in microbial risk assessments should not be overlooked. 4. Methods 4.1 Spike-in selection, cultivation, counting, preservation and purity verification Two marine-derived strains were used as spike-in, Allobacillus halotolerans (ALH, Gram-positive) and Imtechella halotolerans (IMH, Gram-negative), because of their low abundance in environments and high distinguishability between each other. These two strains were purchased from the Belgian Co-ordinated Collections of Micro-organisms (BCCM) and cultured using the Marine Broth (MB) medium. Cells were collected at the early stationary phase (ALH 16h and IMH 24h, respectively) and quantified using both culture method and flow cytometry method (SI2 Table 1, SI2 Table 2). The results of flow cytometry could be used to validate the outcomes obtained from culture method. The culture method results were used to calculate the amount of spike-in (in volume) required for samples. Bacterial suspensions were aliquoted at a ratio of 500 µL of suspension to 250 µL of 60% glycerol and stored in 1.5 mL tubes at -80°C for use in different batches. After stored bacterial suspensions of the two spike-ins were thawed at room temperature, and subsequently centrifuged at 15,000 ×g for 3 minutes (Beckman Coulter Microfuge® 20R) to collect the pellet for DNA extraction. DNeasy PowerSoil Pro Kit (QIAGEN, Germany) was used for DNA extraction. Quality and concentration of the extracted DNA were assessed using NanoDrop™ One (Thermo Scientific™, Thermo Fisher Science) and Qubit® 2.0 Fluorometer (Invitrogen Life Technologies, NY, USA), respectively. DNA libraries were prepared using the SQK-RBK004 kit (Nanopore, UK), and sequencing was performed on GridION with R9.4.1 flow cells (FLO-MIN106) to confirm that both two spike-in bacteria were pure cultures and not contaminated by other species. 4.2 Accuracy assessment of MAQ method using CIS To validate the accuracy of MAQ method using CIS, we introduced two spike-ins into 90 µL of the ZymoBIOMICS® Gut Microbiome Standard mock community (Zymo Research, USA), which had 15 bacteria species with known absolute abundances, at an expected ratio of 4% of total bacterial count, and conducted in triplicates. DNA libraries were constructed using SQK-LSK114.24 kit (Nanopore, UK) and sequenced on PromethION with R10.4.1 (FLO- PRO114M) flow cells. The recovery rates of the six G + and eight G- bacteria in the mock were represented by ALH and IMH, respectively. For the bacterium could not be classified as either G + or G- (defined as G+/-), its recovery rate was represented by the average of ALH and IMH. The absolute abundances of the 15 bacteria in the mock were calculated based on the recovery rates of ALH and IMH and compared to the absolute abundances ranges given by the producer of the mock to assess the accuracy of the MAQ method (Fig. 1 a). 4.3 Reproducibility assessment of MAQ method using CIS To assess the reproducibility, 6 aliquots of an effluent sample from Shatin WWTP in Hong Kong were spiked with ALH and IMH and filtered using cellulose acetate membrane (pore size 0.45 µm). After being lysed under the same conditions, DNA was extracted using two different kits, three samples using PowerSoil Pro kit (QIAGEN, Germany) and another three using PowerWater kit (QIAGEN, Germany). DNA libraries were constructed using SQK-LSK114.24 kit (Nanopore, UK) and sequenced on PromethION with R10.4.1 (FLO- PRO114M) flow cells. Reproducibility of the method was evaluated by analyzing the variations in the proportions of ALH and IMH in relation to the total microorganism abundance, as well as the ratios of ALH/IMH across three replicates for the two DNA extraction kits. The kit that offered better reproducibility was selected for the downstream experiments (Fig. 1 b). 4.4 Applicability assessment of MAQ method using CIS To assess the applicability, the baseline of the two spike-ins as the cellular internal standards in different environmental samples were analyzed using minimap2 (V2.28) to survey the possible interference of native ALH and IMH on results of added spike-ins. We analyzed the metagenomic sequencing data of 119 samples from eight typical categories in Hong Kong, including influent, effluent and activated sludge (AS) of wastewater treatment plant (WWTP), river water (RW), marine water (MW), marine bathing beach water (BBW), marine fishery water (FW), and marine sediment, to survey the relative abundances of ALM and IMH. 4.5 Correction of G + and G- bacteria composition using the two CIS spike-ins Eight aliquots of 10 mL an influent sample from Shatin WWTP were spiked at a 4% ratio for both ALH and IMH. DNA was extracted using four different combinations of intensities and durations for cell lysis and sequenced using the same method as before (see section 2.3 for details), conducted in duplicates. The Pearson correlation coefficients (r) of the relative abundances were computed between these eight samples (pairwise). The variation in r values mainly reflects the impacts of lysis intensity and duration on the effectiveness of bacterial lysis during DNA extraction, as well as potential random variations during library construction and sequencing. We then obtained the absolute abundances of these samples and recalculated the r values by classifying reads into G+, G-, and G+/- according to their phylum-level taxonomic assignments (GTDB) (SI2 Table 3) and scaling the relative abundances with respect to ALH and IMH. After correction, r values were supposed to be higher if the two spike-ins could reduce the biases resulted from different cell lysis conditions (Fig. 1 c). 4.6 Applications of MAQ method in environmental samples Samples were collected from 8 different types of environments in August 2023, including influent (IN), effluent (EF) of five WWTPs in Hong Kong (HYW, NP, YL, STK, and Stone), activated sludge (AS) of HYW, NP and YL, settle sludge of Stone, as well as samples of five river water (RW), five marine bathing beach water (BBW), five marine water (MW), and five marine fishery water (FW). For HYW, AS is from a membrane bioreactor (MBR). Stone uses chemically enhanced primary treatment process. NP is a tertiary wastewater treatment plant (WWTP). Both YL and STK are secondary WWTPs. Adding an appropriate number of spike-ins is crucial for absolute quantification to avoid technical drawbacks, such as big variation if the added amount is too small compared to the total biomass of the sample, and low yield of data useful for microbial analysis of the sample if the opposite is true. Based on previous DNA extraction data of one ST influent sample in August 2021 collected in our lab (sample volume used for extraction and the extracted DNA mass), absolute abundance of which has been reported [ 37 ], we empirically estimated the total number of cells for each sample using their DNA extraction data, and calculated the required number of spike-ins in CFU to achieve an expected relative abundance of 4% for both ALH and IMH. After adding spike-ins, microbial cells of samples were collected on cellulose acetate membrane (pore size 0.45 µm) for IN, EF, RW, BBW, MW and FW, and those were collected by centrifugation for AS and settle sludge. DNAs were extracted using PowerSoil Pro kit (QIAGEN, Germany), quality and concentration of extracted DNA were measured by NanoDrop™ One (Thermo Scientific™, Thermo Fisher Science) and Qubit® 2.0 Fluorometer (Invitrogen Life Technologies, NY, USA), respectively. Purified DNAs (by 1X AMPure XP beads, Beckman Coulter, USA) were individually prepared for Nanopore sequencing using SQK-NBD114.24 following the manufacturer’s protocol, loaded onto R10.4.1 (FLO- PRO114M) flow cells on PromethION, and sequenced until the pores were completely utilized. 4.7 Statistical analyses and absolute quantification calculation methodology Raw sequencing signals were decoded by the built-in Guppy (v1.2.0) basecaller of NanoPore MinKNOW. Adapters were removed using Porechop (v0.2.4). NanoFilt (v2.8.0) was used for quality control [ 40 ], with a filtering criterion of reads shorter than 500 bp and quality scores lower than Q10. Taxonomic classification of reads was performed using Kraken2 [ 41 ] (v2.1.2) with GTDB [ 42 ] (v207). To minimize the interference from the spike-ins, all data related to these two species were excluded [ 37 ]. ARGs were annotated through DIAMOND (v2.1.6) blastx [ 43 , 44 ] using the SARG protein database [ 45 , 46 ] (v3.2). Genetic context of ARG-carrying reads were predicted using PlasFlow [ 47 ] (v1.1). Potential hosts of ARGs were determined based on Kraken2’s results, requiring an additional 1 kb of sequence beyond the ARG region [ 38 ]. Mobile genetic elements associated with ARG-carrying reads were identified by aligning them to the MobileGeneticElementDatabase [ 48 ] using minimap2 [ 49 , 50 ] (v2.24, cutoffs: identity ≥ 80%, subject coverage ≥ 50%). ARG_ranker [ 20 ] (v 2.0) was employed to assess the risk of ARGs. The methodology for metagenomic absolute quantification calculation was detailed in SI1 S1. 5. Conclusion In summary, we have developed a Nanopore-based cellular spike-in absolute quantification method that had been proven to be accurate, repeatable, and broadly applicable. This method used two spike-ins (one G + and one G-) to absolutely quantify G + and G- bacteria respectively, and the spike-ins used in this study have been confirmed to accurately represent the respective bacterial communities and effectively correct the microorganism ratio distortions. Based on the absolute quantification method, we quantified microorganism, pathogens, FIBs, ARGs, and MGEs in 34 samples across eight distinct sample types. We also identified the pathogenic hosts of high-risk ARGs based on Nanopore long-read sequences. Moreover, we assessed the efficiency of five different wastewater treatment processes in removing pathogens and ARGs. Finally, we constructed a microbial risk assessment framework applicable to multiple environments, covering four major aspects: pathogens, FIBs, resistance, and mobilome. Using this framework, the complex concentrations of these items could be converted into simple indexes to derive risk indices for various environments and provide guidance for formulating timely and effective risk management policies. Declarations Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgment The authors appreciate the substantial support from the Theme-based Research Scheme funded by the University Grants Committee of Hong Kong, China (Grant No. T21-705/20-N). Ms. Xianghui Shi, Dr. Yu Yang, Mr. Xi Chen, Ms. Jiahui Ding and Shuxian Li would like to thank the University of Hong Kong for the postgraduate studentship. Dr. Chunxiao Wang, Dr. Xiaoqing Xu and Dr. Xuemei Mao would like to thank the University of Hong Kong for the postdoctoral fellowship. The authors would also like to thank Hong Kong Agriculture, Fisheries and Conservation Department and Hong Kong Environmental Protection Department for sample collections. The computations were performed using research computing facilities offered by Information Technology Services at the University of Hong Kong. The authors would also like to thank the lab technician, Ms. Vicky Fung, for assisting with the experimental process. Data availability All raw sequencing data from the six AS viromes generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) database under BioProject ID: PRJNA1158533. References Essack, S.Y., Environment: the neglected component of the One Health triad. The Lancet Planetary Health, 2018. 2 (6): p. e238-e239. Federigi, I., et al., The application of quantitative microbial risk assessment to natural recreational waters: A review. Marine Pollution Bulletin, 2019. 144 : p. 334-350. Schoen, M.E., et al., Quantitative Microbial Risk Assessment of Antimicrobial Resistant and Susceptible Staphylococcus aureus in Reclaimed Wastewaters. Environmental Science & Technology, 2021. 55 (22): p. 15246-15255. Astles, K.L., et al., An ecological method for qualitative risk assessment and its use in the management of fisheries in New South Wales, Australia. Fisheries Research, 2006. 82 (1): p. 290-303. Hordyk, A.R. and T.R. Carruthers, A quantitative evaluation of a qualitative risk assessment framework: Examining the assumptions and predictions of the Productivity Susceptibility Analysis (PSA). PloS one, 2018. 13 (6): p. e0198298. USEPA 2012 Recreational Water Quality Criteria. The Health and Ecological Criteria Division, O.o.S.a.T. (ed), United States (U.S.) Environmental Protection Agency (EPA). Lim, K.-Y., et al., Evaluation of the dry and wet weather recreational health risks in a semi-enclosed marine embayment in Southern California. Water Research, 2017. 111 : p. 318-329. Tymensen, L.D., et al., Comparative accessory gene fingerprinting of surface water Escherichia coli reveals genetically diverse naturalized population. Journal of Applied Microbiology, 2015. 119 (1): p. 263-277. Mbanga, J., et al., Quantitative microbial risk assessment for waterborne pathogens in a wastewater treatment plant and its receiving surface water body. BMC microbiology, 2020. 20 : p. 1-12. Owens, C.E., et al., Implementation of quantitative microbial risk assessment (QMRA) for public drinking water supplies: systematic review. Water Research, 2020. 174 : p. 115614. Miliotis, M., et al., Role of epidemiology in microbial risk assessment. Food additives and contaminants, 2008. 25 (9): p. 1052-1057. Goh, S.G., et al., A new modelling framework for assessing the relative burden of antimicrobial resistance in aquatic environments. Journal of Hazardous Materials, 2022. 424 : p. 127621. Shao, Y., et al., A systematic review on antibiotics misuse in livestock and aquaculture and regulation implications in China. Science of The Total Environment, 2021. 798 : p. 149205. Shao, S., et al., Research progress on distribution, migration, transformation of antibiotics and antibiotic resistance genes (ARGs) in aquatic environment. Critical reviews in biotechnology, 2018. 38 (8): p. 1195-1208. Huemer, M., et al., Antibiotic resistance and persistence—Implications for human health and treatment perspectives. EMBO reports, 2020. 21 (12): p. e51034. Martínez, J.L., Ecology and evolution of chromosomal gene transfer between environmental microorganisms and pathogens. Microbiology spectrum, 2018. 6 (1): p. 10.1128/microbiolspec. mtbp-0006-2016. Che, Y., et al., Mobile antibiotic resistome in wastewater treatment plants revealed by Nanopore metagenomic sequencing. Microbiome, 2019. 7 : p. 1-13. Forster, S.C., et al., Strain-level characterization of broad host range mobile genetic elements transferring antibiotic resistance from the human microbiome. Nature Communications, 2022. 13 (1): p. 1445. Larsson, D. and C.-F. Flach, Antibiotic resistance in the environment. Nature Reviews Microbiology, 2022. 20 (5): p. 257-269. Zhang, A.-N., et al., An omics-based framework for assessing the health risk of antimicrobial resistance genes. Nature communications, 2021. 12 (1): p. 4765. De Oliveira, D.M., et al., Antimicrobial resistance in ESKAPE pathogens. Clinical microbiology reviews, 2020. 33 (3): p. 10.1128/cmr. 00181-19. Santajit, S. and N. Indrawattana, Mechanisms of antimicrobial resistance in ESKAPE pathogens. BioMed research international, 2016. 2016 . Zhen, X., et al., Economic burden of antibiotic resistance in ESKAPE organisms: a systematic review. Antimicrobial Resistance & Infection Control, 2019. 8 : p. 1-23. Reyneke, B., et al., Comparison of EMA-, PMA-and DNase qPCR for the determination of microbial cell viability. Applied microbiology and biotechnology, 2017. 101 : p. 7371-7383. McLain, J.E., et al., Culture ‐ based methods for detection of antibiotic resistance in agroecosystems: Advantages, challenges, and gaps in knowledge. Journal of environmental quality, 2016. 45 (2): p. 432-440. Ko, K.K., K.R. Chng, and N. Nagarajan, Metagenomics-enabled microbial surveillance. Nature Microbiology, 2022. 7 (4): p. 486-496. Xu, H.-S., et al., Survival and viability of nonculturable Escherichia coli and Vibrio cholerae in the estuarine and marine environment. Microbial ecology, 1982. 8 : p. 313-323. Li, L., et al., The importance of the viable but non-culturable state in human bacterial pathogens. Frontiers in microbiology, 2014. 5 : p. 88860. Frossard, A., F. Hammes, and M.O. Gessner, Flow cytometric assessment of bacterial abundance in soils, sediments and sludge. Frontiers in microbiology, 2016. 7 : p. 195298. Ruijter, J., et al., Amplification efficiency: linking baseline and bias in the analysis of quantitative PCR data. Nucleic acids research, 2009. 37 (6): p. e45-e45. Krehenwinkel, H., et al., Estimating and mitigating amplification bias in qualitative and quantitative arthropod metabarcoding. Scientific reports, 2017. 7 (1): p. 17668. Vandeputte, D., et al., Quantitative microbiome profiling links gut community variation to microbial load. Nature, 2017. 551 (7681): p. 507-511. Zhang, Z., et al., Soil bacterial quantification approaches coupling with relative abundances reflecting the changes of taxa. Scientific reports, 2017. 7 (1): p. 4837. Ji, B.W., et al., Quantifying spatiotemporal variability and noise in absolute microbiota abundances using replicate sampling. Nature Methods, 2019. 16 (8): p. 731-736. Wang, C., et al., Absolute quantification and genome-centric analyses elucidate the dynamics of microbial populations in anaerobic digesters. Water Research, 2022. 224 : p. 119049. Tkacz, A., M. Hortala, and P.S. Poole, Absolute quantitation of microbiota abundance in environmental samples. Microbiome, 2018. 6 : p. 1-13. Yang, Y., et al., Rapid absolute quantification of pathogens and ARGs by nanopore sequencing. Science of The Total Environment, 2022. 809 : p. 152190. Yang, Y., et al., QMRA of beach water by Nanopore sequencing-based viability-metagenomics absolute quantification. Water Research, 2023. 235 : p. 119858. Smets, W., et al., A method for simultaneous measurement of soil bacterial abundances and community composition via 16S rRNA gene sequencing. Soil Biology and Biochemistry, 2016. 96 : p. 145-151. De Coster, W., et al., NanoPack: visualizing and processing long-read sequencing data. Bioinformatics, 2018. 34 (15): p. 2666-2669. Wood, D.E., J. Lu, and B. Langmead, Improved metagenomic analysis with Kraken 2. Genome biology, 2019. 20 : p. 1-13. Parks, D.H., et al., GTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy. Nucleic acids research, 2022. 50 (D1): p. D785-D794. Buchfink, B., K. Reuter, and H.-G. Drost, Sensitive protein alignments at tree-of-life scale using DIAMOND. Nature methods, 2021. 18 (4): p. 366-368. Buchfink, B., C. Xie, and D.H. Huson, Fast and sensitive protein alignment using DIAMOND. Nature methods, 2015. 12 (1): p. 59-60. Yin, X., et al., ARGs-OAP v2. 0 with an expanded SARG database and Hidden Markov Models for enhancement characterization and quantification of antibiotic resistance genes in environmental metagenomes. Bioinformatics, 2018. 34 (13): p. 2263-2270. Yin, X., et al., ARGs-OAP v3. 0: Antibiotic-resistance gene database curation and analysis pipeline optimization. Engineering, 2022. Krawczyk, P.S., L. Lipinski, and A. Dziembowski, PlasFlow: predicting plasmid sequences in metagenomic data using genome signatures. Nucleic acids research, 2018. 46 (6): p. e35-e35. Pärnänen, K., et al., Maternal gut and breast milk microbiota affect infant gut antibiotic resistome and mobile genetic elements. Nature Communications, 2018. 9 (1): p. 3891. Li, H., Minimap2: pairwise alignment for nucleotide sequences. Bioinformatics, 2018. 34 (18): p. 3094-3100. Li, H., New strategies to improve minimap2 alignment accuracy. Bioinformatics, 2021. 37 (23): p. 4572-4574. Pendleton, J.N., S.P. Gorman, and B.F. Gilmore, Clinical relevance of the ESKAPE pathogens. Expert review of anti-infective therapy, 2013. 11 (3): p. 297-308. Rice, L.B., Federal funding for the study of antimicrobial resistance in nosocomial pathogens: no ESKAPE . 2008, The University of Chicago Press. p. 1079-1081. Zhang, S., et al., Dissemination of antibiotic resistance genes (ARGs) via integrons in Escherichia coli: a risk to human health. Environmental Pollution, 2020. 266 : p. 115260. Hirshfeld, B., et al., Prevalence and antimicrobial resistance profiles of Vibrio spp. and Enterococcus spp. in retail shrimp in Northern California. Frontiers in Microbiology, 2023. 14 : p. 1192769. Martinez, J.L., et al., Functional role of bacterial multidrug efflux pumps in microbial natural ecosystems. FEMS microbiology reviews, 2009. 33 (2): p. 430-449. Crossette, E., et al., Metagenomic quantification of genes with internal standards. MBio, 2021. 12 (1): p. 10.1128/mbio. 03173-20. Tourlousse, D.M., et al., Validation and standardization of DNA extraction and library construction methods for metagenomics-based human fecal microbiome measurements. Microbiome, 2021. 9 : p. 1-19. Contijoch, E.J., et al., Gut microbiota density influences host physiology and is shaped by host and microbial factors. Elife, 2019. 8 : p. e40553. Yin, X., et al., Toward a universal unit for quantification of antibiotic resistance genes in environmental samples. Environmental Science & Technology, 2023. 57 (26): p. 9713-9721. Yin, X., et al., An assessment of resistome and mobilome in wastewater treatment plants through temporal and spatial metagenomic analysis. Water Research, 2022. 209 : p. 117885. Tymensen, L., et al., Comparative accessory gene fingerprinting of surface water Escherichia coli reveals genetically diverse naturalized population. Journal of Applied Microbiology, 2015. 119 (1): p. 263-277. Gilmore, M.S., et al., Enterococci: From commensals to leading causes of drug resistant infection [Internet]. 2014. Offenbaume, K.L., E. Bertone, and R.A. Stewart, Monitoring approaches for faecal indicator bacteria in water: Visioning a remote real-time sensor for e. coli and enterococci. Water, 2020. 12 (9): p. 2591. Sharpe, T.J., Assessing a Fluorescence Spectroscopy Method for In-Situ Microbial Drinking Water Quality . 2017, Portland State University. Castañeda-Barba, S., E.M. Top, and T. Stalder, Plasmids, a molecular cornerstone of antimicrobial resistance in the One Health era. Nature Reviews Microbiology, 2024. 22 (1): p. 18-32. Ellabaan, M.M., et al., Forecasting the dissemination of antibiotic resistance genes across bacterial genomes. Nature communications, 2021. 12 (1): p. 2435. https://www.airnow.gov/aqi/ https://www.epd.gov.hk/epd/sites/default/files/epd/sc_chi/environmentinhk/water/hkwqrc/files/waterquality/annual-report/riverreport2020.pdf Additional Declarations There is NO Competing Interest. Supplementary Files GA.png Graphical abstract SI1.docx SI2.xlsx Supplementary Table SI2 Cite Share Download PDF Status: Published Journal Publication published 22 Apr, 2025 Read the published version in Nature Water → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5150537","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":366427226,"identity":"9d5805dd-1cb9-40a0-86e1-faa203552f18","order_by":0,"name":"Tong Zhang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAqklEQVRIiWNgGAWjYDACCTBpkwCk2RgbiNeSkEa6lsMkaJGf3fzs4dcf5/MkZySwPZxBjBbGOcfMjWUSbhdLSySwG24gRguzRIIZUPXtxHkSCWySD4jRwiaR/g2o5RwJWngkcswkPyQcSJwN0kKUwyQkcsqkGdKSE2f2PGyTJMr78jPSt0n+sLFLnHE8+ZhkDzFaQICZB0wRGZEQtT+IVzsKRsEoGAUjEQAAiQMwXj1QfccAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-1148-4322","institution":"The University of Hong Kong","correspondingAuthor":true,"prefix":"","firstName":"Tong","middleName":"","lastName":"Zhang","suffix":""},{"id":366427227,"identity":"17d607c7-ac3f-40f2-89e0-baa5d650d80d","order_by":1,"name":"Xianghui Shi","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Xianghui","middleName":"","lastName":"Shi","suffix":""},{"id":366427228,"identity":"abdaa702-9dfd-4dcc-8185-4d2a7ca5cf0f","order_by":2,"name":"Yu Yang","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Yu","middleName":"","lastName":"Yang","suffix":""},{"id":366427229,"identity":"0f5657be-9fa8-44a8-9f7d-cfef860adefa","order_by":3,"name":"Chunxiao Wang","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Chunxiao","middleName":"","lastName":"Wang","suffix":""},{"id":366427230,"identity":"78f946d6-7b95-4d95-9c76-6fbfc864f73a","order_by":4,"name":"Xiaoqing Xu","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Xiaoqing","middleName":"","lastName":"Xu","suffix":""},{"id":366427231,"identity":"5d80127e-5c96-449d-a1e6-34aae0a77df4","order_by":5,"name":"Xuemei Mao","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Xuemei","middleName":"","lastName":"Mao","suffix":""},{"id":366427232,"identity":"1f30f9df-ab6e-4c48-8a3b-a7f5c664d0e3","order_by":6,"name":"Xi Chen","email":"","orcid":"https://orcid.org/0009-0008-5001-7157","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Xi","middleName":"","lastName":"Chen","suffix":""},{"id":366427233,"identity":"f618f595-3c41-43df-b3ef-3403045647a8","order_by":7,"name":"Jiahui Ding","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Jiahui","middleName":"","lastName":"Ding","suffix":""},{"id":366427234,"identity":"73932f83-409b-4249-8fe0-3ad087dde12f","order_by":8,"name":"Shuxian Li","email":"","orcid":"","institution":"The University of Hong Kong","correspondingAuthor":false,"prefix":"","firstName":"Shuxian","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2024-09-25 09:00:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5150537/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5150537/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s44221-025-00421-y","type":"published","date":"2025-04-22T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":66756291,"identity":"a4b57574-7b4b-451f-a7f4-1d682cca1385","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":208262,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMethod validation. (a)\u003c/strong\u003eEvaluation of the accuracy and reproducibility in simple biological community for cellular spike-in method; \u003cstrong\u003e(b)\u003c/strong\u003e Evaluation of the reproducibility in complex biological community for cellular spike-in method; \u003cstrong\u003e(c)\u003c/strong\u003eEvaluation of the representativeness and effectiveness in correction the proportional biases caused by different lysis efficiencies of G+ and G- from DNA extraction to sequencing.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/67f84ac08eb877f5b6dc86e5.png"},{"id":66756292,"identity":"2088e06d-673e-4712-a3ca-f0cd889205ee","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":167322,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMethod validation results. (a) \u003c/strong\u003eAQ of mock community using spike-ins, rectangular range represent the theoretical absolute abundance of each bacterium, while black dots indicate the absolute abundance of each bacterium calculated through spike-in method in three repeated experiments.\u003cstrong\u003e (b) \u003c/strong\u003eThe correlation between theoretical values and calculated values of bacteria in three repeated experiments with spike-ins added to mock.\u003cstrong\u003e (c) \u003c/strong\u003eError analysis on the proportion of G+ and G- spike-ins in total classified microbial in three ST effluent replicates.\u003cstrong\u003e(d) \u003c/strong\u003eG+ and G- spike-ins original proportion in 119 samples from 8 different environmental sample types without added with spike-ins.\u003cstrong\u003e (e) \u003c/strong\u003eThe differences in correlation among community compositions before and after correction with G+ and G- spike-ins of the same eight ST influent samples added with spike-ins using different combinations of lysis strength and time after sequencing.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/73bd72fed1a469844a317b5d.png"},{"id":66757957,"identity":"4d5bf5b6-a5af-4afd-b26e-ca7224781ee8","added_by":"auto","created_at":"2024-10-16 08:27:26","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":198995,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAQ of pathogens and ARGs in WWTPs. (a) \u003c/strong\u003eAQ of primary pathogens and the removal rate of these pathogens in five WWTPs. \u003cstrong\u003e(b) \u003c/strong\u003eAQ of detected ARG types and the removal rate of these ARG types in five WWTPs.\u003cstrong\u003e (c) \u003c/strong\u003eAQ of six high risk ARG subtypes in five WWTPs.\u003cstrong\u003e(d) \u003c/strong\u003eThe number of ARG subtypes in influent and effluent of these five WWTPs.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/1d124a0960c37c94430829db.png"},{"id":66756298,"identity":"5edd45b5-e57d-47aa-86de-dd2d18ad9c95","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":225268,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAQ of pathogens and ARGs in RW, BBW, MW and FW.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003e AQ of primary pathogens and the removal rate of these pathogens in RW, BBW, MW and FW. \u003cstrong\u003e(b)\u003c/strong\u003e AQ of detected ARG types and the removal rate of these ARG types in RW, BBW, MW and FW. \u003cstrong\u003e(c)\u003c/strong\u003e AQ of high risk ARG subtypes in all RW samples or two or more seawater samples.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/0405fb107c0cfbb1b19f1dca.png"},{"id":66756294,"identity":"94a6f73c-f200-44b0-b76e-618d7678055b","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":216346,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHost tracking of ARGs in different sample types. (a)\u003c/strong\u003e Host tracking of beta-lactam resistance genes in chromosome. (Host tracking of other seven important ARG types was shown in \u003cstrong\u003eSI1 Figure 2\u003c/strong\u003e) \u003cstrong\u003e(b)\u003c/strong\u003e Lg value of ARG copy number per cell of ESKAPE+EV in different habitats. \u003cstrong\u003e(c)\u003c/strong\u003e Lg value of ARG copy number per liter samples of ESKAPE+EV in different habitats.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/7e93af60197edfc37c8760a0.png"},{"id":66756296,"identity":"e5db3c8d-f4da-4c32-8162-1dcbe5020186","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":146204,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRisk assessment of 34 environmental samples. (a)\u003c/strong\u003e Risk assessment framework, including factors of risk assessment and scoring methods. \u003cstrong\u003e(b)\u003c/strong\u003e Risk indexes of 34 samples for 8 scoring methods. \u003cstrong\u003e(c)\u003c/strong\u003e Average risk indexes of 8 scoring methods for 7 environmental sample types. \u003cstrong\u003e(d)\u003c/strong\u003e Final risk indexes of 34 samples. \u003cstrong\u003e(e)\u003c/strong\u003e Average risk indexes of seven environmental sample type.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/836331a736a38c46c09050a5.png"},{"id":81180002,"identity":"8082aa8b-7a40-4f41-be7b-34259f602066","added_by":"auto","created_at":"2025-04-23 07:11:49","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2400078,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/bccc17ca-f7f2-48b5-bbae-a777a1feb40f.pdf"},{"id":66758527,"identity":"c55541ea-24aa-44ea-aa5f-5a8188d484e6","added_by":"auto","created_at":"2024-10-16 08:35:26","extension":"png","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":204560,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGraphical abstract\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"GA.png","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/4fb3cf9d8e5416073bdb1ea6.png"},{"id":66756299,"identity":"ffd13ae3-9f4c-4d65-ad52-9011a35a3cfe","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":3081856,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SI1.docx","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/49a1cb44b2743e9d7e3bb01f.docx"},{"id":66756300,"identity":"e29532a7-6cdd-4a0a-ad3b-826ae16b88f6","added_by":"auto","created_at":"2024-10-16 08:19:26","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":5392696,"visible":true,"origin":"","legend":"\u003cp\u003eSupplementary Table SI2\u003c/p\u003e","description":"","filename":"SI2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-5150537/v1/d787c324cba09821b9a545fc.xlsx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Microbial Risk Assessment Across Diverse Environments Based on Metagenomic Absolute Quantification with Cellular Internal Standard","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eWithin the context of One Health, humans, animals, and the environment are connected [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. A more precise and comprehensive evaluation of microbial risk levels in environments can aid in minimizing human exposure to risks and enhancing health maintenance more effectively. Most existing frameworks for environmental microbial risk assessments predominantly depend on either qualitative evaluations or quantitative analyses using a single factor [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. Qualitative risk assessments are characterized by their descriptive approaches [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e], so they are constrained by subjectivity, ambiguity, and restricted decision-making support [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Single-factor microbial risk assessments are inherently susceptible to biases, as the outcome is directly influenced by one variable. In contrast, a multi-factor microbial risk assessment framework offers a more subtle approach, enabling comprehensive evaluation of environmental risks from diverse perspectives and mitigating biases associated with the limitations of specific factors.\u003c/p\u003e \u003cp\u003eVarious factors can be used to measure environmental risk, necessitating a holistic consideration. Fecal Indicator Bacteria (FIBs) have been used to signal the health hazards from human fecal contamination conventionally [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], but findings in research that FIBs can survive and replicate in natural environments unaffected by human activity have cast doubt on their indicative role [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Total number of pathogens in environments can quantify the overall threat they pose to human health, but it fails to assess the likelihood of various health outcomes\u0026mdash;such as infections, illnesses, or deaths\u0026mdash;caused by different doses of pathogens, as this would require specific dose-response models tailored to each pathogen. Quantitative microbial risk assessment (QMRA) can meet such requirement [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], but QMRA framework is constrained to a small subset of key pathogens [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] since only a limited number of pathogens currently have well-established dose-response models. Additionally, ARGs in the environment have become a significant threat to human health. In recent years, the misuse and overuse of antibiotics in medical, agricultural, and other environmental contexts have led to the widespread presence of ARGs in the environment [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. These genes are persistent in environments [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] and can spread among different bacteria through horizontal gene transfer (HGT) [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], a process primarily facilitated by mobile genetic elements (MGEs). ARG transfer has been observed across phylogenetic barriers and ecological environments [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. Under the selective pressure of antibiotics and co-selection factors, bacteria with ARGs are more likely to survive, thus facilitating the spread of these genes within bacterial communities. Simultaneously, studies showed that different ARGs posed different degrees of potential health risk [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Strategically concentrating on high-risk ARGs with priority is beneficial to improve the effectiveness of environmental management. It is also necessary to focus on the associations between pathogens and ARGs due to their impacts on potential risks. For instance, pathogens carrying ARGs can lead to infections that are more difficult to treat, resulting in increased mortality, extended hospital stays and elevated medical costs, which are serious concerns for public health [\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. Identifying the prevalence of ARGs associated with clinically significant pathogens can aid in assessing health risks and inform strategies for environmental management and pollution control.\u003c/p\u003e \u003cp\u003eA risk assessment framework applicable across multiple environments relies on the absolute abundance of microorganisms to ensure comparability across different settings. Cultivation is regarded as the gold standard for determining microbial absolute abundance [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. However, it presents several drawbacks, including time-consuming labor-intensive [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e], limited specificity and sensitivity [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e], and the significant challenge posed by many microbes being unculturable [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] or having stringent cultivation requirements [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Flow cytometry is a universally applicable and real-time method [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e] for assessing microbial absolute abundance. Despite its advantages, it remains expensive and operationally complex, and is hindered by technical challenges, including difficulties in dispersing clumped cells and separating cells from other impurities. The qPCR technique often has issues of amplification bias [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. The rapid advancement of high-throughput metagenomic sequencing offers a swift and detailed analysis of microbial composition in the environment. It can be combined with total biomass method [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] and internal standard method [\u003cspan additionalcitationids=\"CR35\" citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] to quantify the absolute abundance of various microorganism. Total biomass method still relies on flow cytometry or microscopy or dry weight to measure total microbial mass, and thus cannot circumvent the inherent technical shortcomings of these techniques. Conversely, internal standard (also known as \u0026ldquo;spike-in\u0026rdquo;) method, favored for its operational simplicity and wide applicability, has emerged as a pivotal approach for absolute quantification. This technique now enables quantitative analysis across diverse environments, including soil [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e], sewage treatment facilities [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e], swimming beach waters [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e], and anaerobic reactors [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]. Initially, synthetic or exogenous DNA were used as spike-ins [\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e], then spike-in method has advanced from DNA-based to cellular-based approaches [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] to more accurately simulate environmental bacteria performance from pretreatment to sequencing. However, methods for absolute quantification that combine metagenomic sequencing with internal standards still need further refinement and enhancement. To simulate performance of bacteria more accurately during pretreatment, DNA extraction, and sequencing, it is crucial to develop a method using cellular internal standards of distinct cell structures (like G\u0026thinsp;+\u0026thinsp;and G- bacteria). Additionally, systematic assessment of the spike-in method is essential for standardizing and enhancing the adoption of cellular internal standards in absolute quantification.\u003c/p\u003e \u003cp\u003eTo address challenges mentioned above, we developed a metagenomic absolute quantification method (MAQ) using cellular internal standard (CIS). Two marine bacteria (G\u0026thinsp;+\u0026thinsp;and G-) were chosen as spike-ins to determine the absolute abundance of pathogens and ARGs in environments. We comprehensively evaluated the reproducibility, accuracy, applicability, and representativeness of selected spike-ins, examined the biases arising from different cell lysis conditions and evaluated the effectiveness of selected spike-ins for correcting these biases. Finally, based on the absolute abundance of pathogens, ARGs, FIBs and MGEs, we established a risk assessment framework for multiple environments.\u003c/p\u003e"},{"header":"2. Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Accuracy, reproducibility and applicability of the MAQ methods using CIS\u003c/h2\u003e \u003cp\u003eWith two spike-ins added to mock community, the absolute quantification results showed high consistency with theoretical values (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Absolute quantification of 15 species in the mock community revealed that 14 species (with genome copy numbers greater than 4\u0026times;10\u003csup\u003e3\u003c/sup\u003e cell/mL) had an average deviation of 8.0% from their labeled values, only the species with the lowest genome copy number (3.6\u0026times;10\u003csup\u003e3\u003c/sup\u003e cell/mL) showed a large discrepancy (90\u0026thinsp;~\u0026thinsp;466%) from the labeled value (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, SI2 Table\u0026nbsp;4). For the three replicates, the \u003cem\u003er\u003c/em\u003e values between each pair of replicates calculated with absolute abundance were higher than 0.95, demonstrating high reproducibility in simple microbial communities (SI2 Table\u0026nbsp;4).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis method also showed good reproducibility in complex communities. The same amounts of spike-ins were added to three replicate sub-samples of the same ST effluent samples, which were then extracted with the Powersoil Pro kit, and to three replicate sub-samples that were then extracted with the PowerWater kit, respectively. Relative standard deviation (RSD) of ALH and IMH ratio in total microorganism across three replicates extracted with the same kit ranged from 4%-14%, and the RSD of ALH/IMH across three replicates extracted with the same kit was 21% and 33% (SI2 Table\u0026nbsp;5). Using two different DNA extraction kits cause the variation of bacteria recovery percentages and their ratios. When using PowerSoil Pro kit and PowerWater kit, the ratio of ALH/IMH was 1.8\u0026thinsp;\u0026plusmn;\u0026thinsp;0.2 and 3.1\u0026thinsp;\u0026plusmn;\u0026thinsp;0.3 respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ec), and we chose PowerSoil Pro kit for subsequent experiments due to its better reproducibility and stability.\u003c/p\u003e \u003cp\u003eALH and IMH demonstrated broad applicability in various environments. Aligning the reference genomes of spike-ins to 119 metagenomes, two spike-ins accounted for low abundances of \u0026lt;\u0026thinsp;0.13% (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ed, SI2 Table\u0026nbsp;6). The original amounts of ALH or IMH in these samples were significantly less than the targeted percentages of 3%~5%, making the impact of native amounts of these two spike-ins in the environmental samples negligible.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Correction of G\u0026thinsp;+\u0026thinsp;and G- bacteria composition using the two CIS spike-ins\u003c/h2\u003e \u003cp\u003eTechnical bias could arise at any stage from DNA extraction to library construction and sequencing. Using eight aliquot sub-samples from the same ST influent sample spiked with two spike-ins, we evaluated the variations under four different DNA extraction conditions using two replicates in each condition. Ideally, the correlation coefficient R between different groups should be 1 if there is no extraction bias, however, observed R values for relative abundance of microorganism in eight aliquot sub-samples, which were subjected to different lysis conditions, ranged from 0.58 to 0.96 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ee, SI2 Table\u0026nbsp;7). After correction using recovery rates of the two spike-ins, R of species composition rooted at absolute abundance between pairs of samples under different lysis conditions ranged from 0.83 to 0.98 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ee, SI2 Table\u0026nbsp;7), indicating the biases were significantly corrected by using the two spike-ins to represent their corresponding bacteria groups and providing more accurate composition of the microbial communities.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Profiles and removal rates of pathogens and ARGs in WWTPs\u003c/h2\u003e \u003cp\u003eAcross 14 samples of WWTPs, the absolute abundance of prokaryotic microorganisms ranged from 3.64\u0026times;10\u003csup\u003e8\u003c/sup\u003e cells/L (in effluent treated by the MBR reactor) to 2.03\u0026times;10\u003csup\u003e13\u003c/sup\u003e cells/L (in AS in the MBR reactor). AS had the highest average total cell abundance (TCA) of 3.32\u0026times;10\u003csup\u003e12\u003c/sup\u003e cells/L (excluding AS from the MBR reactor). The average TCA of five influents was 6.31\u0026times;10\u003csup\u003e11\u003c/sup\u003e cells/L. After primary, secondary, and tertiary wastewater treatment, TCA in effluent could be reduced to 1/7 (8.97\u0026times;10\u003csup\u003e10\u003c/sup\u003e cells/L), 1/127 (4.95\u0026times;10\u003csup\u003e9\u003c/sup\u003e cells/L), and 1/1700 (3.71\u0026times;10\u003csup\u003e8\u003c/sup\u003e cells/L) of that in influent, roughly equivalent to 1 log, 2 log and 3 log reduction, respectively (SI2 Table\u0026nbsp;8).\u003c/p\u003e \u003cp\u003eThere were 60 pathogenic species across 27 genera being identified in the top 500 species (based on absolute abundances) in at least one out of the total 34 samples (including 14 WWTP samples and 20 environmental samples). The sum absolute abundance of these pathogens ranged from 1.32\u0026times;10\u003csup\u003e2\u003c/sup\u003e cells/L to 7.03\u0026times;10\u003csup\u003e10\u003c/sup\u003e cells/L, with WWTP influent having the highest average concentration of these pathogens at 2.33\u0026times;10\u003csup\u003e10\u003c/sup\u003e cells/L. MBR AS and AS (excluding AS from the MBR reactor) showed pathogen abundances of 1.63\u0026times;10\u003csup\u003e10\u003c/sup\u003e cells/L and 4.58\u0026times;10\u003csup\u003e9\u003c/sup\u003e cells/L, respectively, and pathogen abundances of effluent from primary, secondary, and tertiary treatment were 4.07\u0026times;10\u003csup\u003e9\u003c/sup\u003e, 1.43\u0026times;10\u003csup\u003e7\u003c/sup\u003e and 9.88\u0026times;10\u003csup\u003e6\u003c/sup\u003e cells/L, respectively (SI2 Table\u0026nbsp;10). There were 7 dominant pathogen genera (including \u003cem\u003eAeromonas\u003c/em\u003e, \u003cem\u003eAcinetobacter\u003c/em\u003e, \u003cem\u003eAliarcobacter\u003c/em\u003e, \u003cem\u003eBacteroides\u003c/em\u003e, \u003cem\u003eMycobacterium\u003c/em\u003e, \u003cem\u003ePseudomonas_E\u003c/em\u003e, and \u003cem\u003eVibrio\u003c/em\u003e) among these 60 pathogens, which accounted for 61\u0026ndash;99% of the total pathogen abundance in WWTPs. Among these pathogenic species, \u003cem\u003eAliarcobacter cryaerophilus_A\u003c/em\u003e consistently exhibited higher abundance relative to other pathogens in 9 samples of total 14 WWTP samples, including all influent samples. Furthermore, \u003cem\u003eAeromonas caviae\u003c/em\u003e, \u003cem\u003ePseudomonas_E alcaligenes\u003c/em\u003e, and \u003cem\u003eVibrio fluvialis\u003c/em\u003e were also identified as high-abundance pathogens across WWTP samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ea, SI2 Table\u0026nbsp;9, SI2 Table\u0026nbsp;10).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAbsolute abundance of ARGs ranged from 3.32 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L to 5.41 \u0026times; 10\u003csup\u003e12\u003c/sup\u003e copies/L in WWTP samples. The highest average absolute abundance was observed in AS of MBR, at 1.8 \u0026times; 10\u003csup\u003e12\u003c/sup\u003e copies/L, closely followed by influent at 1.42 \u0026times; 10\u003csup\u003e12\u003c/sup\u003e copies/L. ARG abundance of primary, secondary and tertiary effluent stood at 1.85 \u0026times; 10\u003csup\u003e11\u003c/sup\u003e, 4.39 \u0026times; 10\u003csup\u003e9\u003c/sup\u003e and 3.56 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L, which was 0.88, 2.51 and 3.6 log/L reduction from influent, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003eb, SI2 Table\u0026nbsp;14). It was crucial to focus on enriched in human environments, mobile, and pathogenic ARGs classified as high-risk (Rank I and Rank II) ARGs to further refine the potential risks posed by ARGs [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. High-risk ARGs constituted 23\u0026ndash;33% of all ARGs in influent samples, with an average of 29%, which was the highest among all environmental types. In primary, secondary and tertiary effluent, the average proportions of high-risk ARGs were 18%, 26% and 10%, respectively. High-risk ARGs accounted for 18%, 12% and 20% in settle sludge, AS (not including MBR AS) and MBR reactor AS. Total absolute abundance of Rank I and II ARGs ranged from 1.18 \u0026times; 10\u003csup\u003e7\u003c/sup\u003e copies/L (in MBR reactor effluent) to 1.68 \u0026times; 10\u003csup\u003e12\u003c/sup\u003e copies/L (in influent) (SI2 Table\u0026nbsp;14, SI2 Table\u0026nbsp;16). Six high-risk ARG subtypes (\u003cem\u003eQnrS2\u003c/em\u003e, \u003cem\u003eere(A)\u003c/em\u003e, \u003cem\u003eOXA-10\u003c/em\u003e, \u003cem\u003eother class A beta-lactamases\u003c/em\u003e, \u003cem\u003eAAC(6')-Ib9\u003c/em\u003e, and \u003cem\u003eaadA\u003c/em\u003e) were present in all influent and effluent samples, with total absolute abundances of 1.92\u0026times;10\u003csup\u003e10\u003c/sup\u003e \u0026minus;\u0026thinsp;2.79\u0026times;10\u003csup\u003e11\u003c/sup\u003e copies/L in influent and 6.36\u0026times;10\u003csup\u003e6\u003c/sup\u003e \u0026minus;\u0026thinsp;1.18\u0026times;10\u003csup\u003e10\u003c/sup\u003e copies/L in effluent (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, SI2 Table\u0026nbsp;13, SI2 Table\u0026nbsp;16).\u003c/p\u003e \u003cp\u003eEmploying the absolute abundance, the removal rate (expressed as log removal) was calculated and five WWTPs exhibited significant variations in pathogen removal rates due to differing wastewater treatment processes. HYW WWTP, employing the MBR process, achieved the highest average removal rate at 3.90 log/L, which was much more effective than the conventional secondary treatment plants YL and STK, with removal rates of 1.93 log/L and 1.91 log/L respectively. In contrast, Stonecutters WWTP lacked a biological treatment process and achieved only a modest average removal rate of 0.32 log/L. This WWTP method was not significantly effective for most pathogens, with log removals below 1 for 54 pathogens. Nevertheless, it achieved high removal rates for 11 \u003cem\u003eMycobacterium\u003c/em\u003e species, its removal rate for \u003cem\u003eMycobacterium chelonae\u003c/em\u003e at 2.65 log/L surpassed the average of 2.5 log/L across the five WWTPs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ea, SI2 Table\u0026nbsp;11).\u003c/p\u003e \u003cp\u003eFor all detected ARGs, the average removal rate across five WWTPs ranged from 0.5 log reduction in Stonecutters to 4.3 log reduction in HYW, closely matching their average removal rate for pathogens. Despite widespread presence and high abundance of MLS (macrolides-lincosamids-streptogramins) resistance genes in influents (ranging from 2\u0026times;10\u003csup\u003e10\u003c/sup\u003e to 1.12\u0026times;10\u003csup\u003e12\u003c/sup\u003e copies/L), this ARG type was consistently removed at rates exceeding that of average for all ARGs across five WWTPs. Aminoglycoside and sulfonamide resistance genes showed significant removal rates improvements under MBR treatment, with log reduction increasing from 2.51\u0026thinsp;\u0026plusmn;\u0026thinsp;0.32 log and 2.21\u0026thinsp;\u0026plusmn;\u0026thinsp;0.42 log in secondary treatment to 5.27 log and 5.47 log in MBR treatment, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003eb, SI2 Table\u0026nbsp;15). After treatment with MBR, 87% of high-risk ARG subtypes detected in influent were absent in the effluent, with no new high-risk ARG subtypes appearing. The results of tertiary treatment revealed that while 41% of high-risk ARG subtypes were not detectable in effluent, eight previously undetected subtypes emerged. The results of CEPT treatment showed that only 13% of high-risk ARG subtypes from influent was missing in the effluent, with the emergence of 12 new subtypes including in 7 ARG types (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003ed, SI2 Table\u0026nbsp;15).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e2.4 Pathogens and ARGs profiles in environmental samples\u003c/h2\u003e \u003cp\u003eThe absolute abundance of prokaryotic microorganisms in 20 environmental samples (including 5 RW, 5 BBW, 5 MW and 5 FW) ranged from 3.09 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e cells/L (in RW) to 1.12 \u0026times; 10\u003csup\u003e10\u003c/sup\u003e cells/L (in MW). For these dominant pathogens (mentioned in Part 3.3), RW, BBW, MW, and FW exhibited similar total absolute quantification levels, ranging from 1.38\u0026times;10\u003csup\u003e6\u003c/sup\u003e to 5.05\u0026times;10\u003csup\u003e6\u003c/sup\u003e cells/L. These dominant pathogen genera in WWTP samples also dominated in three types of environmental samples with more human activity disturbance (RW, BBW and FW), accounting for 71\u0026ndash;83%. In most WWTP samples, \u003cem\u003eAliarcobacter cryaerophilus_A\u003c/em\u003e, which has the highest abundance, remained the most abundant pathogen in some environmental samples, including all 5 FW, 2 RW and 1 MW. For BBW and MW, \u003cem\u003eMycobacterium chelonae\u003c/em\u003e and \u003cem\u003eVibrio alginolyticus\u003c/em\u003e were also the most dominant pathogens in 4 and 3 samples, separately (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003ea, SI2 Table\u0026nbsp;10).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFor absolute abundance of ARGs in 4 types of environmental samples, RW showed an average ARG absolute abundance of 9.18 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L, which was 1.6 to 3.1 times higher than that of the seawater system, including MW, BBW and FW samples (2.94 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e \u0026minus;\u0026thinsp;5.66 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L). The average proportion of high-risk ARGs in RW was similar to that in effluent, at 16%, with a range of 9.8%-25% across five RW samples. In MW, FW, and BBW, the average proportions of high-risk ARGs were much lower, at 3.2%, 1.7%, and 1.2%, respectively (SI2 Table\u0026nbsp;14). Risk Rank I and II ARGs included 67 ARG subtypes across 10 ARG types in 20 environmental samples, with total absolute abundance ranging from 0 copies/L in one BBW and one MW to 5.31\u0026times;10\u003csup\u003e8\u003c/sup\u003e copies/L in RW (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). Five ARG subtypes (\u003cem\u003etet(A)\u003c/em\u003e, \u003cem\u003esul1\u003c/em\u003e, \u003cem\u003eere(A)\u003c/em\u003e, \u003cem\u003eOXA-10\u003c/em\u003e, \u003cem\u003eaadA\u003c/em\u003e) were found in all RW samples, with the absolute abundance of 7.04 \u0026times; 10\u003csup\u003e5\u003c/sup\u003e \u0026ndash; 1.06 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L, including three belonging to the six high-risk ARG subtypes known for their existence in all influents and effluents of these five WWTPs. Meanwhile, two of these five ARG subtypes (\u003cem\u003esul1\u003c/em\u003e and \u003cem\u003eaadA\u003c/em\u003e) and five other ARG subtypes (\u003cem\u003eNPS-1\u003c/em\u003e, \u003cem\u003eAPH\u003c/em\u003e(3\u0026rdquo;)\u003cem\u003e-Ib\u003c/em\u003e, \u003cem\u003emef\u003c/em\u003e(C), \u003cem\u003emph\u003c/em\u003e(G) and \u003cem\u003eVEB-3\u003c/em\u003e) were present in two or more seawater system samples, with the absolute abundance of 6.5 \u0026times; 10\u003csup\u003e5\u003c/sup\u003e \u0026ndash; 1.1 \u0026times; 10\u003csup\u003e8\u003c/sup\u003e copies/L (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e4\u003c/span\u003ec).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e2.5 Host tracking of ARGs\u003c/h2\u003e \u003cp\u003eWe tracked the pathogenic hosts of all identified ARGs across 34 environmental samples from eight types of environments, identifying 17 ARG types carried by 97 pathogens (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003e). We aggregated the pathogen-ARG relationships based on absolute abundance to determine eight ARG types closely associated with pathogens (with pathogen-related absolute abundances exceeding 1\u0026times;10\u003csup\u003e10\u003c/sup\u003e copies/L), which included aminoglycoside, bacitracin, beta-lactam, MLS, polymyxin, rifamycin, sulfonamide, and tetracycline resistance genes (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMore than half (55%) of aminoglycoside resistance genes carried by pathogens were located on the chromosome, which could be used for host tracking. \u003cem\u003eEnterococcus_A raffinosus\u003c/em\u003e was identified to be the major carrier of aminoglycoside resistance genes, which all carried \u003cem\u003eaad\u003c/em\u003e(6) and \u003cem\u003eAPH\u003c/em\u003e(3')\u003cem\u003e-IIIa\u003c/em\u003e detected in influent (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ea, SI2 Table\u0026nbsp;17). Beta-lactam resistance genes carried by chromosomes were linked to the broadest range of pathogens, involving 36 genes across 24 different pathogenic species. \u003cem\u003eAeromonas caviae\u003c/em\u003e, an emerging/re-emerging pathogen, was most closely associated, carrying five beta-lactam resistance genes totaling 9.16\u0026times;10\u003csup\u003e9\u003c/sup\u003e copies/L in influent, effluent and RW (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea, SI2 Table\u0026nbsp;17). All bacitracin resistance genes associated with pathogens were \u003cem\u003ebacA\u003c/em\u003e, predominantly carried by \u003cem\u003eAeromonas caviae\u003c/em\u003e and \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, which was detected in influent. Bacitracin carried by these two pathogens (1.00\u0026times;10\u003csup\u003e10\u003c/sup\u003e copies/L) accounted for 66% of the total bacitracin resistance genes carried by pathogens in chromosome (1.52\u0026times;10\u003csup\u003e10\u003c/sup\u003e copies/L) (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, SI2 Table\u0026nbsp;17). Only 13% of total sulfonamide resistance genes associated with pathogens were carried by chromosome (detected in influent and effluent), which were \u003cem\u003esul1\u003c/em\u003e (95%) and \u003cem\u003esul2\u003c/em\u003e (5.2%), with \u003cem\u003eAeromonas caviae\u003c/em\u003e being the pathogens most closely linked to \u003cem\u003esul1\u003c/em\u003e and all their relationship found in influent (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ec, SI2 Table\u0026nbsp;17). Only 4.2% of rifamycin resistance genes carried by pathogens were identified in chromosome, most of these ARGs from AS sample and was found in \u003cem\u003eMycobacterium gordonae\u003c/em\u003e (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ed, SI2 Table\u0026nbsp;17). For tetracycline resistance genes, pathogens belonging to \u003cem\u003eVibrio\u003c/em\u003e carried \u003cem\u003etet(34)\u003c/em\u003e and \u003cem\u003etet(35)\u003c/em\u003e in their chromosome, and a strong association also appeared between \u003cem\u003ePrevotella disiens\u003c/em\u003e and tetracycline resistance genes of \u003cem\u003etet(Q)\u003c/em\u003e in influent and RW (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ee, SI2 Table\u0026nbsp;17). The primary pathogen related to polymyxin resistance genes was \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, holding 90% of all such genes identified in chromosome of pathogens, specifically \u003cem\u003earnA\u003c/em\u003e and \u003cem\u003epmrF\u003c/em\u003e genes within polymyxin resistance genes (SI1 Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ef, SI2 Table\u0026nbsp;17).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIDSA (The Infectious Diseases Society of America) has specifically highlighted the ESKAPE pathogens for their ability to evade multiple antibiotics, which were persistent in environments and frequently caused hospital-acquired infections [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]. Their transmission methods and antibiotic resistance mechanisms also merited close scrutiny [\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]. Additionally, \u003cem\u003eE.coli\u003c/em\u003e [\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e] and \u003cem\u003eVibrio\u003c/em\u003e [\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e] species were also recognized as significant pathogens and carriers of ARGs. The propensity for ARGs to undergo HGT within the same genus underscored the importance of monitoring ARGs in key pathogenic genera, particularly those ARGs that were both transferable and pathogenic. Across 34 environment samples, 316 ARG subtypes were carried by prokaryotes within the ESKAPE\u0026thinsp;+\u0026thinsp;EV related genera. Among these, tetracycline, quinolone, and multidrug were ARG types consistently associated with the ESKAPE\u0026thinsp;+\u0026thinsp;EV related genera across all sample types, with multidrug being unique as not all were linked to antibiotic resistance [\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e]. Important ARG subtypes carried by microorganism within ESKAPE\u0026thinsp;+\u0026thinsp;EV-related genera were selected using three criteria: being carried by three or more genera in ESKAPE\u0026thinsp;+\u0026thinsp;EV group, presence in three or more sample types, and relevance to ESKAPE\u0026thinsp;+\u0026thinsp;EV-related genera while not being classified as multidrug resistance genes. Following this selection, 23 ARG subtypes (SI1 Table\u0026nbsp;1) that met these standards were identified as ESKAPE\u0026thinsp;+\u0026thinsp;EV-related high-risk ARGs, serving as key indicators for risk assessment of environmental samples. In 6 ESKAPE\u0026thinsp;+\u0026thinsp;EV-related genera, the ARG copy number per cell was significantly higher than the average ARG copy number per cell across all microorganisms in the corresponding sample types (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb). And absolute abundance of ARGs carried by ESKAPE\u0026thinsp;+\u0026thinsp;EV related genus in influent, effluent, and RW was notably high, especially in effluent where the proportion reached 34% of all ARGs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e2.6 Multiple environments risk index under the environmental risk assessment framework\u003c/h2\u003e \u003cp\u003eBased on absolute quantification of pathogens and ARGs in various types of environments, a comprehensive risk assessment framework was developed in this study. The framework was organized into four main sections: pathogens, fecal indicators, resistance, and mobilome, and each was divided into two sub-sections (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). For pathogens, risk assessment involved evaluating total absolute abundance of pathogens (A1) and the cumulative likelihood of infection, which was calculated using dose-response models for 12 key pathogens within the QMRA framework (A2). The resistance section focused on the absolute abundance of Risk Rank I and Rank II ARGs (B1) and ESKAPE\u0026thinsp;+\u0026thinsp;EV-related high risk ARGs (B2). We have integrated commonly used methods for assessing environmental quality through the absolute abundance of fecal indicators (C) into our risk assessment system. The absolute abundance of plasmids (D1) and other MGEs (D2) was also considered within the risk assessment framework to evaluate the transfer and spread potential of ARGs. The final environmental risk index was derived from the arithmetic mean of the eight criteria indexes, where higher index signified greater environmental risk within this framework.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eEither the number of cells per 100 mL of water samples (A1, B1, B2, C1, C2, D1, D2) or the likelihood of infection per 100 mL of water samples (A2) was used as the scoring criteria (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). The indexes from eight criteria demonstrated significant variability when different metrics were employed to assess the same sample (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb, SI2 Table\u0026nbsp;19). Except for A2, the results of the remaining risk assessment methods showed risk indexes for influent and AS were similar. The indexes for effluent are about 2 points lower than those for influent, which were slightly higher than those for RW (except for C2). A1, A2, C2, and D1 indicated that the risk indexes for RW, MW, BBW and FW were close, but the remaining four risk assessment methods showed that the risk level of RW was significantly higher than that of MW, BBW, and FW. Additionally, only B1 and B2 could clearly distinguish the risk levels of MW, BBW, and FW (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec). The final risk indexes across 34 samples ranged from 3.7 to 10.2. The indexes of influent, AS, and settle sludge were notably higher, varying between 7.9 and 10.2. Among the effluents, Stonecutters effluent had the highest risk index at 9.1, followed by YL effluent with the index of 6.6, moreover, effluent risk indexes for the other three WWTPs (HYW, NP, and STK), as well as most of the MW, BBW, FW, and RW samples studied, fell within the range of 4.6 to 5.7. The risk indexes for the two rivers, PR6 and KN1, both were recorded at 6.1, surpassing the average risk index of river water (5.7). Those for BBW_BB26 and MW_PM1 stood at 3.7 and 3.8, which were lower than the average indexes for their respective environmental samples, BBW and MW (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ed).\u003c/p\u003e \u003cp\u003eUsing 34 samples to evaluate within the proposed risk assessment framework, the Pearson\u0026rsquo;s correlation coefficient (R) between the risk indexes obtained from the two assessment methods for the same module (A1\u0026amp;A2; B1\u0026amp;B2; C1\u0026amp;C2 and D1\u0026amp;D2) ranged from 0.84 to 0.97. The R value for A1/A2 and B1/B2 were relatively low (0.84 and 0.87), indicating that the two assessment methods in module A and B were more complementary. The correlation between D1/D2 was relatively high, reaching 0.97, suggesting that the two assessment methods in module D were more corroborative than complementary. Notably, the correlation between A1 and D1 was the strongest among all pairwise comparisons (R was 0.99). Similarly, the correlation between A1 and D2 was also very high at 0.96, suggesting a possible interaction between the total absolute abundance of pathogens and that of mobile elements. Additionally, the correlations between A2 and either B1 or B2 were relatively low (both R values were 0.62), indicating that there might be significant differences between the QMRA risk assessment method, which was constructed based on the pathogenicity of specific pathogens, and the risk assessment methods based on the absolute abundance of high-risk ARGs (SI2 Table\u0026nbsp;19). Therefore, these methods served as complementary approaches from different perspectives within the overall risk assessment framework.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Discussion","content":"\u003cp\u003eThe Pearson\u0026rsquo;s correlation coefficient (R) was used to characterize the variations in microbial community of the groups under different lysis conditions. The samples used in eight experimental groups were identical, theoretically, these differences should not exist. We attributed these discrepancies primarily to different sensitivities of G\u0026thinsp;+\u0026thinsp;and G- bacteria to cell lysis intensity and duration, which would lead to variations in community composition. In the process of absolute quantification, G\u0026thinsp;+\u0026thinsp;and G- bacterial spike-ins were used to simulate the performance of other G\u0026thinsp;+\u0026thinsp;and G- bacteria across the entire processes from DNA extraction to sequencing. The improved correlation coefficient (R) after absolute quantification confirmed the necessity of considering the differences between G\u0026thinsp;+\u0026thinsp;and G- bacteria in absolute quantification. This also indicated that there were limitations of past absolute quantification studies based on one spike-in [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Nevertheless, biases still existed in our research even after such correction (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003ee), possibly due to multiple factors. Firstly, different G\u0026thinsp;+\u0026thinsp;bacteria (or G- bacteria) may have varying sensitivities to lysis strength and duration, and this bias cannot be corrected in this study. Secondly, not every type of G\u0026thinsp;+\u0026thinsp;or G- bacteria can be better represented by the corresponding spike-in (G\u0026thinsp;+\u0026thinsp;or G- spike-ins), for example, with two spike-ins added into Zymo gut community mock, the real concentration of G- bacterium \u003cem\u003eSalmonella enterica\u003c/em\u003e in this mock was consistently closer to the quantification results of the G\u0026thinsp;+\u0026thinsp;spike-in in three parallel experiments (SI2 Table\u0026nbsp;4).\u003c/p\u003e \u003cp\u003eThe differential lysis efficiencies and the corrective role of spike-ins for G\u0026thinsp;+\u0026thinsp;and G- bacteria underscored the importance of using cellular spike-ins throughout the entire process from pretreatment to sequencing in our study. This contrasted with previous studies which proposed negligible differences in extraction efficiencies between G\u0026thinsp;+\u0026thinsp;and G- bacteria allowed skipping the DNA extraction step when using spike-ins to simulate other bacterial behaviors [\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e]. Our experimental results revealed that the extraction efficiencies of two spike-ins only approximated each other under specific lysis intensities and durations, with the ratio of G\u0026thinsp;+\u0026thinsp;to G- spike-ins increasing as lysis intensity and time increase. We also found that even when using sufficiently intense and prolonged lysis conditions to fully release G\u0026thinsp;+\u0026thinsp;bacteria, as previously mentioned [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e], it was difficult to ensure the recovery ratio of G\u0026thinsp;+\u0026thinsp;to G- bacteria remained stable. Under lysis conditions optimized to fully release G\u0026thinsp;+\u0026thinsp;bacteria, DNA of G- bacteria became fragmented due to excessively shearing, which leaded to greater loss of G- bacteria in Nanopore sequencing, resulting in lower recovery rate for G- bacteria compared to G\u0026thinsp;+\u0026thinsp;bacteria (SI2 Table\u0026nbsp;7).\u003c/p\u003e \u003cp\u003eThe absolute quantification units for microorganisms included DNA mass per sample mass (soil) or volume (water and air) [\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e], cells per sample mass or volume [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e], and gene copies per sample mass or volume [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. Absolute abundance could assign practical significance to microbial abundances, revealing the real composition and distribution of taxa in various environments, thereby enables inter-samples comparisons rather than merely intra-sample comparisons [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. While previous metrics for describing ARG abundance included ARG copies/cell, ARG copies/genome, ARG density, ARG copies/16S rRNA gene, RPKM, coverage, PPM, the consensus has shifted towards using copies/cell to describe the average number of ARG copies carried by each cell [\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e]. Based on absolute quantification methods now, the unit of ARG copies/sample mass or volume was widely accepted and used. Different units for describing ARG abundance could help us understand ARG-related issues of various aspects. According to previous reports, the relative abundance of ARGs in AS was typically 0.2\u0026ndash;0.4 copies/cell, significantly lower than 1-2.5 copies/cell typically found in influent of most WWTPs [\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e], indicating the higher prevalence of resistance genes in influent compared to AS. However, our studies showed that the absolute abundance of ARGs in AS was about an order of magnitude higher than that in influent, revealing more higher ARG level in AS per sample volume.\u003c/p\u003e \u003cp\u003eUsing the absolute quantification methods, we compared removal efficiencies of pathogens and ARGs in WWTPs of different treatment process and the results are highly expected, indicating tertiary treatment\u0026thinsp;\u0026gt;\u0026thinsp;secondary\u0026thinsp;\u0026gt;\u0026thinsp;primary. However, the specific removal efficiencies of each species or ARGs may be biased due to some limitations of this study, including a single sampling event and not considering the process of disinfection.\u003c/p\u003e \u003cp\u003eBased on absolute quantification, a comprehensive and multi-environmental risk assessment system could enhance the comprehension of risk by 1) reducing bias and broaden information, better than a single assessment method which might rely too heavily on solely data sources, such as using \u003cem\u003eE. coli\u003c/em\u003e as the single pathogen indicator in bathing beach water, overlooking other pathogenic populations; 2) improving the applicability, better than a single standard which might not be applicable for all environmental samples of high levels of complexity and matrix effect. We have constructed an environmental risk assessment system centered around pathogens, FIBs, ARGs, and mobilome in environments. For the pathogen part, we considered total abundance of pathogens without distinguishing varying pathogenicity of different pathogens. Additionally, we assessed the environmental risk of pathogens using QMRA framework, which calculated the cumulative likelihood of disease based on distinct dose-response models for 12 key pathogens. Each method had its strengths and weaknesses. The former offered a broad overview but lacked precision, while the latter allowed for individual analysis of each pathogen but was constrained by the lack of dose-response data for most pathogens. FIBs were commonly used to assess health risks in recreational waters due to their close association with human diseases [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], however, their delayed response in risk management decisions and the limitations in indicating threats were widely recognized [\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e]. We employed spike-in combined with Nanopore sequencing for quantitative measurement of FIBs in samples, replacing traditional culturing to improve the timeliness of risk assessment. \u003cem\u003eEnterococci\u003c/em\u003e and \u003cem\u003eE.coli\u003c/em\u003e were used as independent environmental risk criteria due to different suitability under different environmental conditions. For instance, \u003cem\u003eEnterococci\u003c/em\u003e had a greater survival ability than \u003cem\u003eE.coli\u003c/em\u003e in colder conditions [\u003cspan citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e, \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e63\u003c/span\u003e], while \u003cem\u003eE.coli\u003c/em\u003e was more suitable for freshwater environments compared to \u003cem\u003eEnterococci\u003c/em\u003e [\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e]. The risk assessment of resistance focused on high-risk ARGs (Risk Rank 1\u0026amp;2) [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] and ARGs associated with the genera of key pathogenic bacteria, known as ESKAPE\u0026thinsp;+\u0026thinsp;EV (specifically described in SI1 S2). The former was based on the inherent characteristics of ARGs, and the latter focused on ARGs associated with multi-resistant key pathogens appeared in multiple environments. Additionally, HGT of most ARGs was related to the conjugation, transformation and transduction process of plasmids [\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e]. At the same time, MGEs like transposons, integrons, and insertion sequences, also promoted the spread of ARGs within bacterial populations and environment [\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e]. The absolute abundance of the mobilome (including plasmids and MGEs) could reflect the risk of antibiotic resistance dissemination to the same extent.\u003c/p\u003e \u003cp\u003eThe globally utilized Air Quality Index (AQI) converted measured concentrations of six major air pollutants (PM2.5, PM10, SO2, NO2, O3, CO) into individual air quality indexes (IAQI), and the highest IAQI among these pollutants determined the AQI [\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e]. In Hong Kong, the River and Stream Water Quality Index (WQI) comprised three assessment indicators: DO, BOD\u003csub\u003e5\u003c/sub\u003e, and ammonia nitrogen, each indicator was scored according to its specific scoring criteria (scores range from 1 to 5), with equal weighting given to all parameters [\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e]. The total score of these three parameters formed the monthly water quality index. Similarly, our risk assessment framework aimed to transform complex absolute abundance values into a simple index. In the conversion processes for A1, B1, B2, C1, C2, D1 and D2, the logarithm of absolute abundance is directly used as the final risk index, ranging from 0 to 13.22, and for A2, pathogenicity served as the scoring criterion, with scores ranging from 0.08-12 based on the absolute value of lg(1-cumulative pathogenic potential). Equal weight was given to the eight scores, and the average of these eight risk indexes constituted the final Microbial Risk Index (MRI).\u003c/p\u003e \u003cp\u003eFurthermore, we explored the contribution of each risk assessment method to the final overall risk index by calculating the difference between the risk index after excluding certain assessment method and the final overall risk index. The results showed that after excluding part D, the average deviation between the risk index of the 34 environmental samples and the actual risk index was the largest, at 20%. Excluding the sub-standards D1 and D2 individually resulted in average deviations of 8.7% and 8.7%, respectively, indicating that part D had a significant contribution to the final result, while D1 and D2 contributing similarly. Excluding A2 resulted in an average deviation of 11% between the obtained environmental risk index and the actual risk index, which was the highest among the eight scoring methods and close to the average deviation when the whole part A was excluded (12%), indicating that A2 had a high contribution to the overall risk assessment system and was unique and irreplaceable within this framework. In contrast, the exclusion of A1 only resulted in an average deviation of 1.6%, indicating that other assessment methods within this risk assessment system produced results similar to those obtained by A1, so most environmental samples could still yield a similar risk index after excluding A1. However, the situation of some environmental samples may be different from the majority, such as BBW_BB26 and MW_PM1, where the deviations between the index obtained after excluding A1 and the final overall risk index reached 8.9% and 8.3%, far exceeding the average level (1.6%). Meanwhile, their dependency on part B and D (33%, 33% and 43%, 43%) also far exceeded the average level (6.5% and 20%). In contrast, the influence of C2 on the final risk index for these two environmental samples (0.79% and 0.76%) was much lower than the average level (4.6%) (SI2 Table\u0026nbsp;19).\u003c/p\u003e \u003cp\u003eOur research has made several improvements compared to previous studies based on spike-in for absolute quantification. Initially, there was no need to use marker genes to differentiate post-added spike-ins from existing species within the community. Instead, we chose two tailored cellular spike-ins, reducing the data loss of bacteria related to marker genes [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Besides, our research detailed reported the lysis process significantly affected the extraction efficiency of G\u0026thinsp;+\u0026thinsp;and G- bacteria and using our spike-ins could be effective in correcting the proportional biases of G\u0026thinsp;+\u0026thinsp;and G- bacteria. Additionally, a multi-environmental risk assessment framework was built based on the absolute quantification results of pathogens, FIBs, ARGs and MGEs. However, we recognized that our study had certain limitations: 1) due to differences among various G\u0026thinsp;+\u0026thinsp;or G- bacteria, we could not completely rule out the possibility that selected spike-ins might not adequately represent certain bacteria, which could lead to significant biases in absolute quantification. 2) Both two spike-ins used in this study and previously used marker gene spike-ins [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] did not perform well in quantifying species of low abundance. In the future, we could consider customizing spike-ins tailored to the characteristics of bacteria from different phyla, in hopes of more closely simulating the recovery rates of various bacteria and striving to address the issue of inaccurate quantification of some low-abundance species. 3) We used Kraken2 to classify microorganism directly from raw reads obtained by Nanopore sequencing, despite the ongoing efforts to improve accuracy in Nanopore sequencing technologies, classifying closely related taxa still had challenges. Furthermore, Kraken2 always tended to overestimate species diversity. 4) This study did not involve research on eukaryotes and phages in environments, but their impact in microbial risk assessments should not be overlooked.\u003c/p\u003e"},{"header":"4. Methods","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Spike-in selection, cultivation, counting, preservation and purity verification\u003c/h2\u003e \u003cp\u003eTwo marine-derived strains were used as spike-in, \u003cem\u003eAllobacillus halotolerans\u003c/em\u003e (ALH, Gram-positive) and \u003cem\u003eImtechella halotolerans\u003c/em\u003e (IMH, Gram-negative), because of their low abundance in environments and high distinguishability between each other. These two strains were purchased from the Belgian Co-ordinated Collections of Micro-organisms (BCCM) and cultured using the Marine Broth (MB) medium. Cells were collected at the early stationary phase (ALH 16h and IMH 24h, respectively) and quantified using both culture method and flow cytometry method (SI2 Table\u0026nbsp;1, SI2 Table\u0026nbsp;2). The results of flow cytometry could be used to validate the outcomes obtained from culture method. The culture method results were used to calculate the amount of spike-in (in volume) required for samples. Bacterial suspensions were aliquoted at a ratio of 500 \u0026micro;L of suspension to 250 \u0026micro;L of 60% glycerol and stored in 1.5 mL tubes at -80\u0026deg;C for use in different batches.\u003c/p\u003e \u003cp\u003eAfter stored bacterial suspensions of the two spike-ins were thawed at room temperature, and subsequently centrifuged at 15,000 \u0026times;g for 3 minutes (Beckman Coulter Microfuge\u0026reg; 20R) to collect the pellet for DNA extraction. DNeasy PowerSoil Pro Kit (QIAGEN, Germany) was used for DNA extraction. Quality and concentration of the extracted DNA were assessed using NanoDrop\u0026trade; One (Thermo Scientific\u0026trade;, Thermo Fisher Science) and Qubit\u0026reg; 2.0 Fluorometer (Invitrogen Life Technologies, NY, USA), respectively. DNA libraries were prepared using the SQK-RBK004 kit (Nanopore, UK), and sequencing was performed on GridION with R9.4.1 flow cells (FLO-MIN106) to confirm that both two spike-in bacteria were pure cultures and not contaminated by other species.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Accuracy assessment of MAQ method using CIS\u003c/h2\u003e \u003cp\u003eTo validate the accuracy of MAQ method using CIS, we introduced two spike-ins into 90 \u0026micro;L of the ZymoBIOMICS\u0026reg; Gut Microbiome Standard mock community (Zymo Research, USA), which had 15 bacteria species with known absolute abundances, at an expected ratio of 4% of total bacterial count, and conducted in triplicates. DNA libraries were constructed using SQK-LSK114.24 kit (Nanopore, UK) and sequenced on PromethION with R10.4.1 (FLO- PRO114M) flow cells. The recovery rates of the six G\u0026thinsp;+\u0026thinsp;and eight G- bacteria in the mock were represented by ALH and IMH, respectively. For the bacterium could not be classified as either G\u0026thinsp;+\u0026thinsp;or G- (defined as G+/-), its recovery rate was represented by the average of ALH and IMH. The absolute abundances of the 15 bacteria in the mock were calculated based on the recovery rates of ALH and IMH and compared to the absolute abundances ranges given by the producer of the mock to assess the accuracy of the MAQ method (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003ea).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Reproducibility assessment of MAQ method using CIS\u003c/h2\u003e \u003cp\u003eTo assess the reproducibility, 6 aliquots of an effluent sample from Shatin WWTP in Hong Kong were spiked with ALH and IMH and filtered using cellulose acetate membrane (pore size 0.45 \u0026micro;m). After being lysed under the same conditions, DNA was extracted using two different kits, three samples using PowerSoil Pro kit (QIAGEN, Germany) and another three using PowerWater kit (QIAGEN, Germany). DNA libraries were constructed using SQK-LSK114.24 kit (Nanopore, UK) and sequenced on PromethION with R10.4.1 (FLO- PRO114M) flow cells. Reproducibility of the method was evaluated by analyzing the variations in the proportions of ALH and IMH in relation to the total microorganism abundance, as well as the ratios of ALH/IMH across three replicates for the two DNA extraction kits. The kit that offered better reproducibility was selected for the downstream experiments (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003eb).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e4.4 Applicability assessment of MAQ method using CIS\u003c/h2\u003e \u003cp\u003eTo assess the applicability, the baseline of the two spike-ins as the cellular internal standards in different environmental samples were analyzed using minimap2 (V2.28) to survey the possible interference of native ALH and IMH on results of added spike-ins. We analyzed the metagenomic sequencing data of 119 samples from eight typical categories in Hong Kong, including influent, effluent and activated sludge (AS) of wastewater treatment plant (WWTP), river water (RW), marine water (MW), marine bathing beach water (BBW), marine fishery water (FW), and marine sediment, to survey the relative abundances of ALM and IMH.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e4.5 Correction of G\u0026thinsp;+\u0026thinsp;and G- bacteria composition using the two CIS spike-ins\u003c/h2\u003e \u003cp\u003eEight aliquots of 10 mL an influent sample from Shatin WWTP were spiked at a 4% ratio for both ALH and IMH. DNA was extracted using four different combinations of intensities and durations for cell lysis and sequenced using the same method as before (see section \u003cspan refid=\"Sec5\" class=\"InternalRef\"\u003e2.3\u003c/span\u003e for details), conducted in duplicates. The Pearson correlation coefficients (r) of the relative abundances were computed between these eight samples (pairwise). The variation in r values mainly reflects the impacts of lysis intensity and duration on the effectiveness of bacterial lysis during DNA extraction, as well as potential random variations during library construction and sequencing. We then obtained the absolute abundances of these samples and recalculated the r values by classifying reads into G+, G-, and G+/- according to their phylum-level taxonomic assignments (GTDB) (SI2 Table\u0026nbsp;3) and scaling the relative abundances with respect to ALH and IMH. After correction, r values were supposed to be higher if the two spike-ins could reduce the biases resulted from different cell lysis conditions (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e1\u003c/span\u003ec).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e4.6 Applications of MAQ method in environmental samples\u003c/h2\u003e \u003cp\u003eSamples were collected from 8 different types of environments in August 2023, including influent (IN), effluent (EF) of five WWTPs in Hong Kong (HYW, NP, YL, STK, and Stone), activated sludge (AS) of HYW, NP and YL, settle sludge of Stone, as well as samples of five river water (RW), five marine bathing beach water (BBW), five marine water (MW), and five marine fishery water (FW). For HYW, AS is from a membrane bioreactor (MBR). Stone uses chemically enhanced primary treatment process. NP is a tertiary wastewater treatment plant (WWTP). Both YL and STK are secondary WWTPs.\u003c/p\u003e \u003cp\u003eAdding an appropriate number of spike-ins is crucial for absolute quantification to avoid technical drawbacks, such as big variation if the added amount is too small compared to the total biomass of the sample, and low yield of data useful for microbial analysis of the sample if the opposite is true. Based on previous DNA extraction data of one ST influent sample in August 2021 collected in our lab (sample volume used for extraction and the extracted DNA mass), absolute abundance of which has been reported [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e], we empirically estimated the total number of cells for each sample using their DNA extraction data, and calculated the required number of spike-ins in CFU to achieve an expected relative abundance of 4% for both ALH and IMH. After adding spike-ins, microbial cells of samples were collected on cellulose acetate membrane (pore size 0.45 \u0026micro;m) for IN, EF, RW, BBW, MW and FW, and those were collected by centrifugation for AS and settle sludge. DNAs were extracted using PowerSoil Pro kit (QIAGEN, Germany), quality and concentration of extracted DNA were measured by NanoDrop\u0026trade; One (Thermo Scientific\u0026trade;, Thermo Fisher Science) and Qubit\u0026reg; 2.0 Fluorometer (Invitrogen Life Technologies, NY, USA), respectively. Purified DNAs (by 1X AMPure XP beads, Beckman Coulter, USA) were individually prepared for Nanopore sequencing using SQK-NBD114.24 following the manufacturer\u0026rsquo;s protocol, loaded onto R10.4.1 (FLO- PRO114M) flow cells on PromethION, and sequenced until the pores were completely utilized.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e4.7 Statistical analyses and absolute quantification calculation methodology\u003c/h2\u003e \u003cp\u003eRaw sequencing signals were decoded by the built-in Guppy (v1.2.0) basecaller of NanoPore MinKNOW. Adapters were removed using Porechop (v0.2.4). NanoFilt (v2.8.0) was used for quality control [\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e], with a filtering criterion of reads shorter than 500 bp and quality scores lower than Q10. Taxonomic classification of reads was performed using Kraken2 [\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e] (v2.1.2) with GTDB [\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e] (v207). To minimize the interference from the spike-ins, all data related to these two species were excluded [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. ARGs were annotated through DIAMOND (v2.1.6) blastx [\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e] using the SARG protein database [\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e, \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e] (v3.2). Genetic context of ARG-carrying reads were predicted using PlasFlow [\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e] (v1.1). Potential hosts of ARGs were determined based on Kraken2\u0026rsquo;s results, requiring an additional 1 kb of sequence beyond the ARG region [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]. Mobile genetic elements associated with ARG-carrying reads were identified by aligning them to the MobileGeneticElementDatabase [\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e] using minimap2 [\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e, \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e] (v2.24, cutoffs: identity\u0026thinsp;\u0026ge;\u0026thinsp;80%, subject coverage\u0026thinsp;\u0026ge;\u0026thinsp;50%). ARG_ranker [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e] (v 2.0) was employed to assess the risk of ARGs. The methodology for metagenomic absolute quantification calculation was detailed in SI1 S1.\u003c/p\u003e \u003c/div\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eIn summary, we have developed a Nanopore-based cellular spike-in absolute quantification method that had been proven to be accurate, repeatable, and broadly applicable. This method used two spike-ins (one G\u0026thinsp;+\u0026thinsp;and one G-) to absolutely quantify G\u0026thinsp;+\u0026thinsp;and G- bacteria respectively, and the spike-ins used in this study have been confirmed to accurately represent the respective bacterial communities and effectively correct the microorganism ratio distortions. Based on the absolute quantification method, we quantified microorganism, pathogens, FIBs, ARGs, and MGEs in 34 samples across eight distinct sample types. We also identified the pathogenic hosts of high-risk ARGs based on Nanopore long-read sequences. Moreover, we assessed the efficiency of five different wastewater treatment processes in removing pathogens and ARGs. Finally, we constructed a microbial risk assessment framework applicable to multiple environments, covering four major aspects: pathogens, FIBs, resistance, and mobilome. Using this framework, the complex concentrations of these items could be converted into simple indexes to derive risk indices for various environments and provide guidance for formulating timely and effective risk management policies.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eDeclaration of Competing Interest\u003c/h2\u003e \u003cp\u003eThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.\u003c/p\u003e \u003ch2\u003eAcknowledgment\u003c/h2\u003e \u003cp\u003eThe authors appreciate the substantial support from the Theme-based Research Scheme funded by the University Grants Committee of Hong Kong, China (Grant No. T21-705/20-N). Ms. Xianghui Shi, Dr. Yu Yang, Mr. Xi Chen, Ms. Jiahui Ding and Shuxian Li would like to thank the University of Hong Kong for the postgraduate studentship. Dr. Chunxiao Wang, Dr. Xiaoqing Xu and Dr. Xuemei Mao would like to thank the University of Hong Kong for the postdoctoral fellowship. The authors would also like to thank Hong Kong Agriculture, Fisheries and Conservation Department and Hong Kong Environmental Protection Department for sample collections. The computations were performed using research computing facilities offered by Information Technology Services at the University of Hong Kong. The authors would also like to thank the lab technician, Ms. Vicky Fung, for assisting with the experimental process.\u003c/p\u003e\u003ch2\u003eData availability\u003c/h2\u003e \u003cp\u003eAll raw sequencing data from the six AS viromes generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) database under BioProject ID: PRJNA1158533.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eEssack, S.Y., \u003cem\u003eEnvironment: the neglected component of the One Health triad.\u003c/em\u003e The Lancet Planetary Health, 2018. \u003cstrong\u003e2\u003c/strong\u003e(6): p. e238-e239.\u003c/li\u003e\n\u003cli\u003eFederigi, I., et al., \u003cem\u003eThe application of quantitative microbial risk assessment to natural recreational waters: A review.\u003c/em\u003e Marine Pollution Bulletin, 2019. \u003cstrong\u003e144\u003c/strong\u003e: p. 334-350.\u003c/li\u003e\n\u003cli\u003eSchoen, M.E., et al., \u003cem\u003eQuantitative Microbial Risk Assessment of Antimicrobial Resistant and Susceptible Staphylococcus aureus in Reclaimed Wastewaters.\u003c/em\u003e Environmental Science \u0026amp; Technology, 2021. \u003cstrong\u003e55\u003c/strong\u003e(22): p. 15246-15255.\u003c/li\u003e\n\u003cli\u003eAstles, K.L., et al., \u003cem\u003eAn ecological method for qualitative risk assessment and its use in the management of fisheries in New South Wales, Australia.\u003c/em\u003e Fisheries Research, 2006. \u003cstrong\u003e82\u003c/strong\u003e(1): p. 290-303.\u003c/li\u003e\n\u003cli\u003eHordyk, A.R. and T.R. Carruthers, \u003cem\u003eA quantitative evaluation of a qualitative risk assessment framework: Examining the assumptions and predictions of the Productivity Susceptibility Analysis (PSA).\u003c/em\u003e PloS one, 2018. \u003cstrong\u003e13\u003c/strong\u003e(6): p. e0198298.\u003c/li\u003e\n\u003cli\u003eUSEPA 2012 Recreational Water Quality Criteria. The Health and Ecological Criteria Division, O.o.S.a.T. (ed), United States (U.S.) Environmental Protection Agency (EPA). \u003c/li\u003e\n\u003cli\u003eLim, K.-Y., et al., \u003cem\u003eEvaluation of the dry and wet weather recreational health risks in a semi-enclosed marine embayment in Southern California.\u003c/em\u003e Water Research, 2017. \u003cstrong\u003e111\u003c/strong\u003e: p. 318-329.\u003c/li\u003e\n\u003cli\u003eTymensen, L.D., et al., \u003cem\u003eComparative accessory gene fingerprinting of surface water Escherichia coli reveals genetically diverse naturalized population.\u003c/em\u003e Journal of Applied Microbiology, 2015. \u003cstrong\u003e119\u003c/strong\u003e(1): p. 263-277.\u003c/li\u003e\n\u003cli\u003eMbanga, J., et al., \u003cem\u003eQuantitative microbial risk assessment for waterborne pathogens in a wastewater treatment plant and its receiving surface water body.\u003c/em\u003e BMC microbiology, 2020. \u003cstrong\u003e20\u003c/strong\u003e: p. 1-12.\u003c/li\u003e\n\u003cli\u003eOwens, C.E., et al., \u003cem\u003eImplementation of quantitative microbial risk assessment (QMRA) for public drinking water supplies: systematic review.\u003c/em\u003e Water Research, 2020. \u003cstrong\u003e174\u003c/strong\u003e: p. 115614.\u003c/li\u003e\n\u003cli\u003eMiliotis, M., et al., \u003cem\u003eRole of epidemiology in microbial risk assessment.\u003c/em\u003e Food additives and contaminants, 2008. \u003cstrong\u003e25\u003c/strong\u003e(9): p. 1052-1057.\u003c/li\u003e\n\u003cli\u003eGoh, S.G., et al., \u003cem\u003eA new modelling framework for assessing the relative burden of antimicrobial resistance in aquatic environments.\u003c/em\u003e Journal of Hazardous Materials, 2022. \u003cstrong\u003e424\u003c/strong\u003e: p. 127621.\u003c/li\u003e\n\u003cli\u003eShao, Y., et al., \u003cem\u003eA systematic review on antibiotics misuse in livestock and aquaculture and regulation implications in China.\u003c/em\u003e Science of The Total Environment, 2021. \u003cstrong\u003e798\u003c/strong\u003e: p. 149205.\u003c/li\u003e\n\u003cli\u003eShao, S., et al., \u003cem\u003eResearch progress on distribution, migration, transformation of antibiotics and antibiotic resistance genes (ARGs) in aquatic environment.\u003c/em\u003e Critical reviews in biotechnology, 2018. \u003cstrong\u003e38\u003c/strong\u003e(8): p. 1195-1208.\u003c/li\u003e\n\u003cli\u003eHuemer, M., et al., \u003cem\u003eAntibiotic resistance and persistence\u0026mdash;Implications for human health and treatment perspectives.\u003c/em\u003e EMBO reports, 2020. \u003cstrong\u003e21\u003c/strong\u003e(12): p. e51034.\u003c/li\u003e\n\u003cli\u003eMart\u0026iacute;nez, J.L., \u003cem\u003eEcology and evolution of chromosomal gene transfer between environmental microorganisms and pathogens.\u003c/em\u003e Microbiology spectrum, 2018. \u003cstrong\u003e6\u003c/strong\u003e(1): p. 10.1128/microbiolspec. mtbp-0006-2016.\u003c/li\u003e\n\u003cli\u003eChe, Y., et al., \u003cem\u003eMobile antibiotic resistome in wastewater treatment plants revealed by Nanopore metagenomic sequencing.\u003c/em\u003e Microbiome, 2019. \u003cstrong\u003e7\u003c/strong\u003e: p. 1-13.\u003c/li\u003e\n\u003cli\u003eForster, S.C., et al., \u003cem\u003eStrain-level characterization of broad host range mobile genetic elements transferring antibiotic resistance from the human microbiome.\u003c/em\u003e Nature Communications, 2022. \u003cstrong\u003e13\u003c/strong\u003e(1): p. 1445.\u003c/li\u003e\n\u003cli\u003eLarsson, D. and C.-F. Flach, \u003cem\u003eAntibiotic resistance in the environment.\u003c/em\u003e Nature Reviews Microbiology, 2022. \u003cstrong\u003e20\u003c/strong\u003e(5): p. 257-269.\u003c/li\u003e\n\u003cli\u003eZhang, A.-N., et al., \u003cem\u003eAn omics-based framework for assessing the health risk of antimicrobial resistance genes.\u003c/em\u003e Nature communications, 2021. \u003cstrong\u003e12\u003c/strong\u003e(1): p. 4765.\u003c/li\u003e\n\u003cli\u003eDe Oliveira, D.M., et al., \u003cem\u003eAntimicrobial resistance in ESKAPE pathogens.\u003c/em\u003e Clinical microbiology reviews, 2020. \u003cstrong\u003e33\u003c/strong\u003e(3): p. 10.1128/cmr. 00181-19.\u003c/li\u003e\n\u003cli\u003eSantajit, S. and N. Indrawattana, \u003cem\u003eMechanisms of antimicrobial resistance in ESKAPE pathogens.\u003c/em\u003e BioMed research international, 2016. \u003cstrong\u003e2016\u003c/strong\u003e.\u003c/li\u003e\n\u003cli\u003eZhen, X., et al., \u003cem\u003eEconomic burden of antibiotic resistance in ESKAPE organisms: a systematic review.\u003c/em\u003e Antimicrobial Resistance \u0026amp; Infection Control, 2019. \u003cstrong\u003e8\u003c/strong\u003e: p. 1-23.\u003c/li\u003e\n\u003cli\u003eReyneke, B., et al., \u003cem\u003eComparison of EMA-, PMA-and DNase qPCR for the determination of microbial cell viability.\u003c/em\u003e Applied microbiology and biotechnology, 2017. \u003cstrong\u003e101\u003c/strong\u003e: p. 7371-7383.\u003c/li\u003e\n\u003cli\u003eMcLain, J.E., et al., \u003cem\u003eCulture\u003c/em\u003e\u003cem\u003e‐\u003c/em\u003e\u003cem\u003ebased methods for detection of antibiotic resistance in agroecosystems: Advantages, challenges, and gaps in knowledge.\u003c/em\u003e Journal of environmental quality, 2016. \u003cstrong\u003e45\u003c/strong\u003e(2): p. 432-440.\u003c/li\u003e\n\u003cli\u003eKo, K.K., K.R. Chng, and N. Nagarajan, \u003cem\u003eMetagenomics-enabled microbial surveillance.\u003c/em\u003e Nature Microbiology, 2022. \u003cstrong\u003e7\u003c/strong\u003e(4): p. 486-496.\u003c/li\u003e\n\u003cli\u003eXu, H.-S., et al., \u003cem\u003eSurvival and viability of nonculturable Escherichia coli and Vibrio cholerae in the estuarine and marine environment.\u003c/em\u003e Microbial ecology, 1982. \u003cstrong\u003e8\u003c/strong\u003e: p. 313-323.\u003c/li\u003e\n\u003cli\u003eLi, L., et al., \u003cem\u003eThe importance of the viable but non-culturable state in human bacterial pathogens.\u003c/em\u003e Frontiers in microbiology, 2014. \u003cstrong\u003e5\u003c/strong\u003e: p. 88860.\u003c/li\u003e\n\u003cli\u003eFrossard, A., F. Hammes, and M.O. Gessner, \u003cem\u003eFlow cytometric assessment of bacterial abundance in soils, sediments and sludge.\u003c/em\u003e Frontiers in microbiology, 2016. \u003cstrong\u003e7\u003c/strong\u003e: p. 195298.\u003c/li\u003e\n\u003cli\u003eRuijter, J., et al., \u003cem\u003eAmplification efficiency: linking baseline and bias in the analysis of quantitative PCR data.\u003c/em\u003e Nucleic acids research, 2009. \u003cstrong\u003e37\u003c/strong\u003e(6): p. e45-e45.\u003c/li\u003e\n\u003cli\u003eKrehenwinkel, H., et al., \u003cem\u003eEstimating and mitigating amplification bias in qualitative and quantitative arthropod metabarcoding.\u003c/em\u003e Scientific reports, 2017. \u003cstrong\u003e7\u003c/strong\u003e(1): p. 17668.\u003c/li\u003e\n\u003cli\u003eVandeputte, D., et al., \u003cem\u003eQuantitative microbiome profiling links gut community variation to microbial load.\u003c/em\u003e Nature, 2017. \u003cstrong\u003e551\u003c/strong\u003e(7681): p. 507-511.\u003c/li\u003e\n\u003cli\u003eZhang, Z., et al., \u003cem\u003eSoil bacterial quantification approaches coupling with relative abundances reflecting the changes of taxa.\u003c/em\u003e Scientific reports, 2017. \u003cstrong\u003e7\u003c/strong\u003e(1): p. 4837.\u003c/li\u003e\n\u003cli\u003eJi, B.W., et al., \u003cem\u003eQuantifying spatiotemporal variability and noise in absolute microbiota abundances using replicate sampling.\u003c/em\u003e Nature Methods, 2019. \u003cstrong\u003e16\u003c/strong\u003e(8): p. 731-736.\u003c/li\u003e\n\u003cli\u003eWang, C., et al., \u003cem\u003eAbsolute quantification and genome-centric analyses elucidate the dynamics of microbial populations in anaerobic digesters.\u003c/em\u003e Water Research, 2022. \u003cstrong\u003e224\u003c/strong\u003e: p. 119049.\u003c/li\u003e\n\u003cli\u003eTkacz, A., M. Hortala, and P.S. Poole, \u003cem\u003eAbsolute quantitation of microbiota abundance in environmental samples.\u003c/em\u003e Microbiome, 2018. \u003cstrong\u003e6\u003c/strong\u003e: p. 1-13.\u003c/li\u003e\n\u003cli\u003eYang, Y., et al., \u003cem\u003eRapid absolute quantification of pathogens and ARGs by nanopore sequencing.\u003c/em\u003e Science of The Total Environment, 2022. \u003cstrong\u003e809\u003c/strong\u003e: p. 152190.\u003c/li\u003e\n\u003cli\u003eYang, Y., et al., \u003cem\u003eQMRA of beach water by Nanopore sequencing-based viability-metagenomics absolute quantification.\u003c/em\u003e Water Research, 2023. \u003cstrong\u003e235\u003c/strong\u003e: p. 119858.\u003c/li\u003e\n\u003cli\u003eSmets, W., et al., \u003cem\u003eA method for simultaneous measurement of soil bacterial abundances and community composition via 16S rRNA gene sequencing.\u003c/em\u003e Soil Biology and Biochemistry, 2016. \u003cstrong\u003e96\u003c/strong\u003e: p. 145-151.\u003c/li\u003e\n\u003cli\u003eDe Coster, W., et al., \u003cem\u003eNanoPack: visualizing and processing long-read sequencing data.\u003c/em\u003e Bioinformatics, 2018. \u003cstrong\u003e34\u003c/strong\u003e(15): p. 2666-2669.\u003c/li\u003e\n\u003cli\u003eWood, D.E., J. Lu, and B. Langmead, \u003cem\u003eImproved metagenomic analysis with Kraken 2.\u003c/em\u003e Genome biology, 2019. \u003cstrong\u003e20\u003c/strong\u003e: p. 1-13.\u003c/li\u003e\n\u003cli\u003eParks, D.H., et al., \u003cem\u003eGTDB: an ongoing census of bacterial and archaeal diversity through a phylogenetically consistent, rank normalized and complete genome-based taxonomy.\u003c/em\u003e Nucleic acids research, 2022. \u003cstrong\u003e50\u003c/strong\u003e(D1): p. D785-D794.\u003c/li\u003e\n\u003cli\u003eBuchfink, B., K. Reuter, and H.-G. Drost, \u003cem\u003eSensitive protein alignments at tree-of-life scale using DIAMOND.\u003c/em\u003e Nature methods, 2021. \u003cstrong\u003e18\u003c/strong\u003e(4): p. 366-368.\u003c/li\u003e\n\u003cli\u003eBuchfink, B., C. Xie, and D.H. Huson, \u003cem\u003eFast and sensitive protein alignment using DIAMOND.\u003c/em\u003e Nature methods, 2015. \u003cstrong\u003e12\u003c/strong\u003e(1): p. 59-60.\u003c/li\u003e\n\u003cli\u003eYin, X., et al., \u003cem\u003eARGs-OAP v2. 0 with an expanded SARG database and Hidden Markov Models for enhancement characterization and quantification of antibiotic resistance genes in environmental metagenomes.\u003c/em\u003e Bioinformatics, 2018. \u003cstrong\u003e34\u003c/strong\u003e(13): p. 2263-2270.\u003c/li\u003e\n\u003cli\u003eYin, X., et al., \u003cem\u003eARGs-OAP v3. 0: Antibiotic-resistance gene database curation and analysis pipeline optimization.\u003c/em\u003e Engineering, 2022.\u003c/li\u003e\n\u003cli\u003eKrawczyk, P.S., L. Lipinski, and A. Dziembowski, \u003cem\u003ePlasFlow: predicting plasmid sequences in metagenomic data using genome signatures.\u003c/em\u003e Nucleic acids research, 2018. \u003cstrong\u003e46\u003c/strong\u003e(6): p. e35-e35.\u003c/li\u003e\n\u003cli\u003eP\u0026auml;rn\u0026auml;nen, K., et al., \u003cem\u003eMaternal gut and breast milk microbiota affect infant gut antibiotic resistome and mobile genetic elements.\u003c/em\u003e Nature Communications, 2018. \u003cstrong\u003e9\u003c/strong\u003e(1): p. 3891.\u003c/li\u003e\n\u003cli\u003eLi, H., \u003cem\u003eMinimap2: pairwise alignment for nucleotide sequences.\u003c/em\u003e Bioinformatics, 2018. \u003cstrong\u003e34\u003c/strong\u003e(18): p. 3094-3100.\u003c/li\u003e\n\u003cli\u003eLi, H., \u003cem\u003eNew strategies to improve minimap2 alignment accuracy.\u003c/em\u003e Bioinformatics, 2021. \u003cstrong\u003e37\u003c/strong\u003e(23): p. 4572-4574.\u003c/li\u003e\n\u003cli\u003ePendleton, J.N., S.P. Gorman, and B.F. Gilmore, \u003cem\u003eClinical relevance of the ESKAPE pathogens.\u003c/em\u003e Expert review of anti-infective therapy, 2013. \u003cstrong\u003e11\u003c/strong\u003e(3): p. 297-308.\u003c/li\u003e\n\u003cli\u003eRice, L.B., \u003cem\u003eFederal funding for the study of antimicrobial resistance in nosocomial pathogens: no ESKAPE\u003c/em\u003e. 2008, The University of Chicago Press. p. 1079-1081.\u003c/li\u003e\n\u003cli\u003eZhang, S., et al., \u003cem\u003eDissemination of antibiotic resistance genes (ARGs) via integrons in Escherichia coli: a risk to human health.\u003c/em\u003e Environmental Pollution, 2020. \u003cstrong\u003e266\u003c/strong\u003e: p. 115260.\u003c/li\u003e\n\u003cli\u003eHirshfeld, B., et al., \u003cem\u003ePrevalence and antimicrobial resistance profiles of Vibrio spp. and Enterococcus spp. in retail shrimp in Northern California.\u003c/em\u003e Frontiers in Microbiology, 2023. \u003cstrong\u003e14\u003c/strong\u003e: p. 1192769.\u003c/li\u003e\n\u003cli\u003eMartinez, J.L., et al., \u003cem\u003eFunctional role of bacterial multidrug efflux pumps in microbial natural ecosystems.\u003c/em\u003e FEMS microbiology reviews, 2009. \u003cstrong\u003e33\u003c/strong\u003e(2): p. 430-449.\u003c/li\u003e\n\u003cli\u003eCrossette, E., et al., \u003cem\u003eMetagenomic quantification of genes with internal standards.\u003c/em\u003e MBio, 2021. \u003cstrong\u003e12\u003c/strong\u003e(1): p. 10.1128/mbio. 03173-20.\u003c/li\u003e\n\u003cli\u003eTourlousse, D.M., et al., \u003cem\u003eValidation and standardization of DNA extraction and library construction methods for metagenomics-based human fecal microbiome measurements.\u003c/em\u003e Microbiome, 2021. \u003cstrong\u003e9\u003c/strong\u003e: p. 1-19.\u003c/li\u003e\n\u003cli\u003eContijoch, E.J., et al., \u003cem\u003eGut microbiota density influences host physiology and is shaped by host and microbial factors.\u003c/em\u003e Elife, 2019. \u003cstrong\u003e8\u003c/strong\u003e: p. e40553.\u003c/li\u003e\n\u003cli\u003eYin, X., et al., \u003cem\u003eToward a universal unit for quantification of antibiotic resistance genes in environmental samples.\u003c/em\u003e Environmental Science \u0026amp; Technology, 2023. \u003cstrong\u003e57\u003c/strong\u003e(26): p. 9713-9721.\u003c/li\u003e\n\u003cli\u003eYin, X., et al., \u003cem\u003eAn assessment of resistome and mobilome in wastewater treatment plants through temporal and spatial metagenomic analysis.\u003c/em\u003e Water Research, 2022. \u003cstrong\u003e209\u003c/strong\u003e: p. 117885.\u003c/li\u003e\n\u003cli\u003eTymensen, L., et al., \u003cem\u003eComparative accessory gene fingerprinting of surface water Escherichia coli reveals genetically diverse naturalized population.\u003c/em\u003e Journal of Applied Microbiology, 2015. \u003cstrong\u003e119\u003c/strong\u003e(1): p. 263-277.\u003c/li\u003e\n\u003cli\u003eGilmore, M.S., et al., \u003cem\u003eEnterococci: From commensals to leading causes of drug resistant infection [Internet].\u003c/em\u003e 2014.\u003c/li\u003e\n\u003cli\u003eOffenbaume, K.L., E. Bertone, and R.A. Stewart, \u003cem\u003eMonitoring approaches for faecal indicator bacteria in water: Visioning a remote real-time sensor for e. coli and enterococci.\u003c/em\u003e Water, 2020. \u003cstrong\u003e12\u003c/strong\u003e(9): p. 2591.\u003c/li\u003e\n\u003cli\u003eSharpe, T.J., \u003cem\u003eAssessing a Fluorescence Spectroscopy Method for In-Situ Microbial Drinking Water Quality\u003c/em\u003e. 2017, Portland State University.\u003c/li\u003e\n\u003cli\u003eCasta\u0026ntilde;eda-Barba, S., E.M. Top, and T. Stalder, \u003cem\u003ePlasmids, a molecular cornerstone of antimicrobial resistance in the One Health era.\u003c/em\u003e Nature Reviews Microbiology, 2024. \u003cstrong\u003e22\u003c/strong\u003e(1): p. 18-32.\u003c/li\u003e\n\u003cli\u003eEllabaan, M.M., et al., \u003cem\u003eForecasting the dissemination of antibiotic resistance genes across bacterial genomes.\u003c/em\u003e Nature communications, 2021. \u003cstrong\u003e12\u003c/strong\u003e(1): p. 2435.\u003c/li\u003e\n\u003cli\u003ehttps://www.airnow.gov/aqi/\u003c/li\u003e\n\u003cli\u003ehttps://www.epd.gov.hk/epd/sites/default/files/epd/sc_chi/environmentinhk/water/hkwqrc/files/waterquality/annual-report/riverreport2020.pdf\u003cstrong\u003e\u003c/strong\u003e\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5150537/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5150537/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe risk posed by microorganisms in diverse environments has emerged as a significant concern. Despite this, existing microbial risk assessment frameworks often lack comprehensiveness and systematicness. To tackle this constraint, we developed a cellular spike-in (one Gram-positive and one Gram-negative bacteria) method that enables absolute quantification of microorganisms in various environmental compartments. This method was rigorously evaluated for reproducibility, accuracy, and applicability. Furthermore, we investigated biases that might arise from DNA extraction to sequencing under different cell lysis conditions for both types of bacteria, and importantly, demonstrated that this spike-in absolute quantification method could correct such biases. We then applied this method to a range of samples to determine the absolute abundance of various microorganisms, pathogens, and antibiotic resistance genes (ARGs) across eight different sample types, including influent, effluent, primary sludge, activated sludge, marine water, marine bathing beach water, marine fishery water, and river water. Based on the results, we evaluated and compared the treatment efficiencies in terms of pathogens and ARGs in five WWTPs of different operational modes. Finally, we integrated the absolute abundances of 1) total pathogens and key pathogens used for cumulative pathogenic possibility calculation in the framework of Quantitative Microbial Risk Assessment (QMRA); 2) Risk Rank1\u0026amp;2 ARGs and high-risk ARGs associated with ESKAPE (\u003cem\u003eEnterococcus faecium\u003c/em\u003e, \u003cem\u003eStaphylococcus aureus\u003c/em\u003e, \u003cem\u003eKlebsiella pneumoniae\u003c/em\u003e, \u003cem\u003eAcinetobacter baumannii\u003c/em\u003e, \u003cem\u003ePseudomonas aeruginosa\u003c/em\u003e, and \u003cem\u003eEnterobacter spp.\u003c/em\u003e)\u0026thinsp;+\u0026thinsp;EV (\u003cem\u003eE.coli\u003c/em\u003e and \u003cem\u003eVibrio spp.\u003c/em\u003e); 3) two most common fecal indicator bacteria (FIBs), namely \u003cem\u003eEscherichia coli\u003c/em\u003e and \u003cem\u003eEnterococci\u003c/em\u003e; and 4) plasmids and other mobile genetic elements (MGEs), into an index to facilitate comprehensive microbial risk assessment and comparison across different environments.\u003c/p\u003e","manuscriptTitle":"Microbial Risk Assessment Across Diverse Environments Based on Metagenomic Absolute Quantification with Cellular Internal Standard","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-10-16 08:19:21","doi":"10.21203/rs.3.rs-5150537/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-water","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"natwater","sideBox":"Learn more about [Nature Water](https://www.nature.com/natwater/)","snPcode":"44221","submissionUrl":"https://mts-natwater.nature.com/cgi-bin/main.plex","title":"Nature Water","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"432ca124-a9a9-4e17-a79a-3055a73af004","owner":[],"postedDate":"October 16th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":38986504,"name":"Earth and environmental sciences/Environmental sciences/Environmental impact"},{"id":38986505,"name":"Biological sciences/Biotechnology/Environmental biotechnology"},{"id":38986506,"name":"Biological sciences/Microbiology/Bacteria/Metagenomics"},{"id":38986507,"name":"Biological sciences/Genetics/Sequencing/DNA sequencing"}],"tags":[],"updatedAt":"2025-04-23T07:11:44+00:00","versionOfRecord":{"articleIdentity":"rs-5150537","link":"https://doi.org/10.1038/s44221-025-00421-y","journal":{"identity":"nature-water","isVorOnly":false,"title":"Nature Water"},"publishedOn":"2025-04-22 04:00:00","publishedOnDateReadable":"April 22nd, 2025"},"versionCreatedAt":"2024-10-16 08:19:21","video":"","vorDoi":"10.1038/s44221-025-00421-y","vorDoiUrl":"https://doi.org/10.1038/s44221-025-00421-y","workflowStages":[]},"version":"v1","identity":"rs-5150537","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5150537","identity":"rs-5150537","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}