Comprehensive Pathogen Profiling in Adult Patients with Severe Acute Respiratory Infections Using Metagenomic Next-Generation Sequencing of Sputum Samples

Comprehensive Pathogen Profiling in Adult Patients with Severe Acute Respiratory Infections Using Metagenomic Next-Generation Sequencing of Sputum Samples | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 17 December 2025 V1 Latest version Share on Comprehensive Pathogen Profiling in Adult Patients with Severe Acute Respiratory Infections Using Metagenomic Next-Generation Sequencing of Sputum Samples Authors : Nang Kham-Kjing 0000-0001-8053-9468 , Rathakarn Kawila , Patcharaporn Tariyo , Kittiyaporn Puapun , Nicole Ngo-Giang-Huong , Sayamon Hongjaisee , and Woottichai Khamduang 0000-0003-0196-5691 [email protected] Authors Info & Affiliations https://doi.org/10.22541/au.176599276.68130321/v1 197 views 126 downloads Contents Abstract Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Severe Acute Respiratory Infection (SARI) remains a major global health burden, yet conventional diagnostics frequently fail to identify the causative pathogens. This study aimed to comprehensively characterize the respiratory microbial landscape in adult SARI patients and to evaluate the clinical utility of an optimized long-read metagenomic next-generation sequencing (mNGS) workflow. A total of 101 remnant specimens from hospitalized adults in northern Thailand were analyzed using SMART-9N-based cDNA synthesis, Oxford Nanopore sequencing, and a streamlined bioinformatics pipeline. Integrated clinical data was used to assess associations with disease severity. Long-read mNGS demonstrated superior diagnostic yield compared to multiplex PCR and culture, detecting pathogens in 78% of cases and uniquely identifying etiologic agents in 18% of specimens negative by routine diagnostics, including Klebsiella pneumoniae , Haemophilus parainfluenzae , and SARS-CoV-2. High concordance was observed for respiratory viruses, with genome-wide coverage supporting accurate detection. Phylogenetic analysis revealed the co-circulation of HRV-A, HRV-B, and HRV-C, with HRV-C predominating. Multivariable analysis identified male sex and reduced oxygen saturation as independent predictors of severe disease. These findings underscore the diagnostic power of long-read mNGS for uncovering atypical, unculturable, and polymicrobial infections in adult SARI and support its integration into respiratory pathogen surveillance. Introduction Acute respiratory infections (ARIs) remain a leading cause of hospitalization and mortality globally, with pneumonia accounting for a substantial proportion of respiratory-related deaths [1]. Severe Acute Respiratory Infection (SARI), as defined by the World Health Organization (WHO), represents the most severe end of the ARI spectrum and imposes a considerable burden on healthcare systems [2]. Despite its clinical significance, the causative pathogen remains unidentified in a substantial proportion of SARI cases when conventional diagnostic methods are used [3]. Although most SARI cases are attributed to respiratory viruses, bacterial and fungal pathogens can also play etiological roles. Multiplex PCR assays have enhanced sensitivity for selected pathogens but are inherently limited by their targeted nature, failing to identify unexpected, atypical, or emerging organisms. In contrast, metagenomic next-generation sequencing (mNGS) provides an unbiased approach capable of detecting a broad range of pathogens without prior assumptions [4, 5]. A recent study demonstrates that mNGS can identify unculturable or unanticipated organisms frequently missed by conventional diagnostics [6], offering new insights into pathogen ecology. SARI is increasingly recognized as a polymicrobial condition, however, the relationship between microbial diversity and disease severity remains unclear. Some evidence suggests that higher microbial diversity may exacerbate disease through synergistic interactions [7, 8], while other study associates lower diversity with dominance of a single pathogenic organism [9]. In addition, host-related factors such as age, comorbidities, and immune status are also critical determinants of disease severity and pathogen distribution. Older adults tend to have distinct pathogen profiles compared with younger populations, and face greater risks due to underlying conditions [10, 11]. Understanding how these host factors influence pathogen distribution may elucidate the heterogeneity in clinical outcomes among SARI patients. Most existing SARI etiology studies rely on targeted PCR-based diagnostics [12-14], leaving important gaps in our understanding of the full microbial spectrum and its relationship to clinical severity. In this retrospective study, we integrate mNGS data with clinical and outcome variables to provide a high-resolution characterization of the SARI etiological landscape. We aim to evaluate the diagnostic utility of mNGS for detecting pathogens missed by routine diagnostics and to determine whether specific microbial profiles or host factors correlate with disease severity. Our findings may inform clinical decision-making, guide antimicrobial stewardship, and enhance pathogen surveillance strategies for adult SARI populations. Material and methods Study design and patient enrollment This retrospective study included adult patients hospitalized with SARI at a regional hospital in Chiang Mai, northern Thailand, between November 2023 and April 2024. Inclusion criteria were based on the WHO case definition, which includes acute respiratory illness with onset of fever (≥38°C) and cough within 10 days, requiring hospital admission [15]. A total of 101 remnant respiratory specimens collected during routine diagnostic care were used for analysis. Clinical and laboratory data were retrieved from hospital records. Sample processing, extraction, and host genome depletion Sputum samples were treated with an equal volume of 0.1% dithiothreitol, DTT, (Thermo Fisher Scientific, MA, USA) and incubated at room temperature until complete liquefaction. The liquefied samples were centrifuged at 10,000×g for 3 minutes to pellet debris, and the supernatant was collected for nucleic acid extraction. For the nasopharyngeal swab sample, no pretreatment step was applied. Total nucleic acids were extracted from 200 µL of each liquefied sputum or nasopharyngeal sample using the QIAamp® Viral RNA Mini Kit (Qiagen GmbH, Hilden, Germany) following the manufacturer’s protocol. To deplete host genome, extracted nucleic acids were treated with RNase-Free DNase Set (Qiagen GmbH, Hilden, Germany) at room temperature, following the manufacturer’s instructions. The treated nucleic acids were then purified using the RNA Clean & Concentrator™-25 Kit (Zymo Research, CA, USA). Library preparation and Oxford Nanopore sequencing Complementary DNA (cDNA) was synthesized using SMART-9N primers ( S witching M echanism a t the 5′ end of R NA T emplate with random 9-n ucleotide primers), following our previously optimized protocol [16]. Briefly, a 12 µL reaction contained 0.16 µM 9N primer, 0.83 mM dNTPs, and purified nucleic acids, incubated at 65°C for 5 minutes. First-strand synthesis was performed by combining the annealed RNA with 8 µL of a mixture containing 2.5× SuperScript™ IV buffer, 12.5 mM DTT, 5 U RNase inhibitor, 25 U SuperScript™ IV reverse transcriptase (Thermo Fisher Scientific, MA, USA), and 0.25 µM strand-switching primer. The reaction was then incubated at 42°C for 90 minutes, followed by 70°C for 10 minutes. cDNA amplification was carried out in 25 µL of reaction containing 1× Q5® reaction buffer, 0.2 µM dNTPs, 0.8 µM PCR primer (AAGCAGTGGTATCAACGCAGAGT), 0.02 U of Q5® High-Fidelity DNA Polymerase (New England Biolabs, MA, USA), and 2.5 µL of cDNA. PCR cycling included an initial denaturation at 98°C for 45 seconds; 30 cycles of 98°C for 15 seconds, 62°C for 15 seconds, and 65°C for 5 minutes; with a final extension at 65°C for 10 minutes. The amplified products were purified using magnetic bead cleanup, VAHTS DNA Clean Beads (Vazyme, Nanjing, China) in equal ratio, and quantified using the DeNovix dsDNA Broad Range Assay Kit (DeNovix Inc., DE, USA) to ensure sufficient yield for sequencing. Sequencing libraries were prepared with the Oxford Nanopore Native Barcoding Kit (ONT-SQK-NBD114.24, Oxford Nanopore Technologies, Oxford, UK) following the manufacturer’s protocol. Sequencing was conducted on the MinION Mk1C platform (Oxford Nanopore Technologies, Oxford, UK) for up to 72 hours. Bioinformatics analysis Bioinformatic analysis followed protocols established in our prior study [16]. Briefly, raw sequencing data were basecalled using the Dorado basecaller with a high-accuracy model via the MinKNOW software (v24.11.10). NanoComp v1.23.1 and NanoPlot v1.42.0 were used for quality control [17]. Residual overhanging primer/adapter sequences introduced by the SMART-9N protocol were trimmed using Cutadapt v5.0 [18]. Shorter reads (< 200 base paris, bp)were filtered with fastp v0.24 [19]. Host genome were removed by mapping the reads to the human reference genome (GRCh38) using Minimap2 v2.28-r1209 with parameters optimized for nanopore reads (-x map-ont) [20]. The host-depleted reads were subjected to de novo assembly using MetaFlye v2.9.5-b1801 [21] and polished using Medaka tool v2.0.1 [22]. Taxonomy classification was done using Kraken2 v2.1.3 with the Kraken “PlusPF” database [23] and visualized using KronaTools v2.8.1 [24]. To access the genome coverage of the potential pathogens, we mapped the host-filtered reads to the respective reference genomes of key pathogens using Minimap2 and calculated per-base coverage depth using SAMtools v1.21 [25]. Phylogenetic analysis For phylogenetic tree reconstruction, HRV-positive samples from this study were compared with reference prototypes downloaded from the GenBank database. All consensus sequences were combined and aligned using MAFFT (v7.526) with the –auto strategy [26]. Poorly aligned positions and high-gap columns were removed using trimAl (v1.5.rev0) with the automated1 algorithm to retain reliably aligned regions [27]. Maximum-likelihood phylogenetic trees were reconstructed using IQ-TREE2 (v2.4.0), applying the ModelFinder Plus (MFP) algorithm for best-fit substitution model selection, and branch support was assessed using 1,000 ultrafast bootstrap replicates (UFBoot2) together with 1,000 SH-aLRT tests [28-30]. Enterovirus D68 (EV-D68) reference sequences were included as an outgroup to root the HRV phylogeny. The phylogenetic tree was visualized in RStudio (version 4.4.2; RStudio, MA, USA) using ggtree [31]. Statistical analysis Descriptive statistics were used to summarize clinical characteristics, laboratory parameters, sequencing performance, and pathogen prevalence, reported as frequencies and percentages for categorical variables and medians with interquartile ranges (IQR) for continuous variables. The distribution of continuous variables was first assessed using the Shapiro–Wilk test to determine normality. Continuous variables were compared using the Wilcoxon rank-sum test. Relative abundances were normalized using total-sum scaling (TSS). Alpha-diversity indices, including Shannon and Chao1, were calculated to assess within-sample microbial richness and diversity, while beta-diversity was evaluated using Bray–Curtis dissimilarity followed by principal coordinates analysis (PCoA). Differences in microbial community structure between clinical severity groups were tested using PERMANOVA with 999 permutations. To identify risk factors associated with severe disease, logistic regression modeling was performed using the finalfit package in R [32]. Univariable logistic regression analyses were initially performed, and variables with p < 0.20 were considered eligible for multivariable models. Statistical significance was defined as p < 0.05. All statistical analyses and visualizations were conducted using RStudio (version 4.4.2; RStudio, MA, USA). Ethics Approval This study was conducted under ethical approvals obtained from the Research Ethics Committee of the Faculty of Associated Medical Sciences, Chiang Mai University (Approval No. AMSEC-66EM-011) and the Ethics Committee of the participating hospital (Approval No. 112/66). The research utilized de-identified clinical data and remnant respiratory specimens that were originally collected for routine diagnostic purposes. All samples were coded and anonymized prior to analysis to protect patient confidentiality. As only leftover diagnostic specimens were used and no interventions were conducted on patients, a waiver of informed consent was granted in accordance with Thai ethical guidelines for research use of archived specimens. Sequencing performance and pathogen profiles To evaluate the performance of the optimized long-read mNGS workflow developed in our prior study, 101 respiratory samples from adult SARI patients were sequenced. Successful sequencing was achieved in 91% (92/101) of samples. Taxonomic classification at the family, genus, and species levels was obtained in 78% (79/101), 73% (74/101), and 72% (73/101) of samples, respectively (Figure 1A). Assembly failure occurred in 9% (9/101) of samples, while 13% (13/101) yielded host-derived reads only. To assess the added diagnostic value of mNGS, pathogen identification outcomes were compared with multiplex PCR results reported in our earlier investigation, as well as with conventional culture results from hospital records. The mNGS approach exhibited a notably higher detection rate (79/101; 78%) compared to multiplex PCR (47/101; 47%) and culture (40/101; 40%). Concordance analysis revealed that 30 samples (30%) were positive by both mNGS and multiplex PCR. Notably, mNGS uniquely identified pathogens in 18 samples (18%) that were negative by both multiplex PCR and culture (Figure 1B). To further explore the diagnostic utility of mNGS beyond targeted assays, we stratified the results based on multiplex PCR status. Among multiplex PCR-negative cases (n=34), mNGS detected a diverse array of clinically relevant pathogens, including Klebsiella pneumoniae (11.8%), Haemophilus parainfluenzae (8.8%), Pseudomonas aeruginosa (5.9%), severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2, 2.9%), and several oral anaerobes potentially implicated in lower respiratory tract infections in patients with comorbidities (Figure 1C). In multiplex PCR-positive samples (n=39), mNGS reliably confirmed the targeted pathogens while also uncovering additional co-infecting bacteria not included in the multiplex PCR panels (Figure 1D). Figure 1. Summary of the optimized metagenomic sequencing performance, method concordance, and species-level taxonomic findings. (A) Sequencing outcomes for 101 samples at the family, genus, and species levels. Bar plot showing successful taxonomic assignment (blue), partial host-genome interference (green), and sequencing failure (red). (B) Upset plot showing pathogen detection across mNGS, multiplex PCR, and culture. The upper bars represent intersection sample sizes, while connected orange nodes indicate method combinations. Horizontal bars show the total number of positive samples per method. (C) Top 20 species identified by mNGS among multiplex PCR–negative samples (n=34). (D) Top 20 species identified by mNGS among multiplex PCR–positive samples (n=39). Microbial Landscape of SARI Patients Metagenomic sequencing revealed a diverse and clinically significant microbial landscape among the 79 adult SARI patients (Figure 2). The most prevalent bacterial families included Enterobacteriaceae and Flavobacteriaceae , alongside several anaerobic and oral respiratory associated families such as Porphyromonadaceae, Prevotellaceae, Neisseriaceae, and Pasteurellaceae . Viral families were also represented, with Picornaviridae and Coronaviridae being the most commonly detected. At the genus level, Capnocytophaga , Porphyromonas , Escherichia , and Prevotella were predominant, collectively comprising a substantial portion of the detected microbial community. These genera reflect a blend of potential respiratory pathogens, oral anaerobes, and aspiration-associated taxa. Other notable genera included Enterovirus , Rothia, Haemophilus, Klebsiella, Pseudomonas, and Leptotrichia , all of which are recognized contributors to lower respiratory tract infections in adults. Species-level analysis further emphasized organisms of potential pathogenic importance. Capnocytophaga gingivalis was the most frequently detected species, followed by Porphyromonas sp. oral taxon 275, Escherichia coli , Capnocytophaga sputigena , Rothia mucilaginosa , and Haemophilus parainfluenzae . Several viral pathogens, including rhinovirus A, rhinovirus C, and SARS-CoV-2, were also identified. In addition, recognized bacterial respiratory pathogens such as Klebsiella pneumoniae , Prevotella melaninogenica , and Metamycoplasma hominis were detected, highlighting the complex and polymicrobial nature of SARI infections in this cohort. Figure 2. Prevalence of identified taxa among adult SARI patients based on metagenomic sequencing. Horizontal bar plots illustrate the distribution of detected pathogens at three taxonomic levels: family, genus, and species. Only the top 20 taxa at each level are displayed. Prevalence of each taxon was calculated as the proportion of samples with successful mNGS classification (n=79). Samples with host-only reads (n=13) or assembly failure (n=9) were excluded from denominator calculations. Comparison of pathogen detection across mNGS, multiplex PCR, and culture methods To assess the concordance between standard diagnostic methods (multiplex PCR and culture) and the optimized metagenomic sequencing approach, we compared species-level pathogen identifications across the three platforms (Figure 3). The mNGS approach demonstrated the broadest detection range, identifying a substantially more diverse array of viral and bacterial taxa compared to culture or multiplex PCR. Multiplex PCR showed strong performance for viral agents included in its detection panel but identified fewer bacterial species. Culture-based methods predominantly recovered fast-growing respiratory bacteria. Overlapping detections were mainly observed between mNGS and multiplex PCR for high-abundance viral species such as human rhinovirus (HRV), human metapneumovirus (HMPV), influenza virus, human coronaviruses (HCoV), and human parainfluenza virus (HPIV), demonstrating strong agreement for high-abundance respiratory viruses. In contrast, concordance between mNGS and culture was confined to culturable organisms, particularly Klebsiella pneumoniae and Acinetobacter baumannii , though these were detected in relatively few samples. Importantly, mNGS identified atypical pathogens such as Mycoplasma pneumoniae , which are typically challenging to culture and require molecular assays for confirmation. Figure 3. Comparison of pathogen detection across three diagnostic methods. Tile plot comparing organisms detected using mNGS, multiplex PCR, and culture among SARI patient samples. Rows represent patient samples, and columns represent detected species. Colors indicate species detected by mNGS + multiplex PCR (red), mNGS + culture (yellow), mNGS only (green), multiplex only (blue), culture only (purple), or no detection (grey). mNGS detected the highest diversity of pathogens, including atypical bacteria and unculturable pathogens, while multiplex PCR predominantly identified targeted viral species within the panel, and culture recovered only a limited subset of classical bacterial pathogens. Genome-wide coverage profiles confirm accurate pathogen detection by mNGS To evaluate the reliability and accuracy of pathogen detection achieved by the optimized mNGS workflow, genome-wide coverage profiles were generated for representative pathogens with concordant detection by both multiplex PCR and mNGS. These analyses demonstrated robust and contiguous read coverage across the genomes of various respiratory pathogens, including Mycoplasma pneumoniae , HCoV-HKU1, HCoV-OC43, HMPV, HPIV-3, HRV, influenza A virus (IAV), and SARS-CoV-2 (Figure 4). High-density, continuous coverage, particularly for HRV and IAV, confirmed effective enrichment and recovery of pathogen reads, even in the context of sputum samples characterized by high human genomic background. For lower-abundance pathogens such as M. pneumoniae, HCoV-HKU1 and HCoV-OC43, the read coverage patterns remained specific and aligned precisely to their respective reference genomes, supporting confident and accurate identification. Figure 4. Metagenomic coverage depth across representative respiratory pathogen genomes. Coverage plots show read depth across reference genomes for pathogens identified in representative SARI patient samples. Reads were mapped to respective reference genomes using minimap2, and per-base depth was visualized after host reads removal. Distinct coverage patterns are observed for each pathogen, with high and continuous depth for viruses such as HRV and IAV, and characteristic localized patterns for M. pneumoniae and coronaviruses. Age-stratified pathogen profile To investigate age-specific distributions of respiratory pathogens among adult SARI patients and assess whether certain pathogens disproportionately affect specific age groups, we conducted an age-stratified analysis of metagenomic findings (Supplementary Figure 1). The results revealed notable variations in pathogen prevalence across the three defined age groups: 18–30 years, 31–64 years, and ≥65 years. Overall, elderly patients (≥65 years) exhibited the greatest diversity of detected pathogens, suggesting a more complex microbial landscape in this population. Several organisms, including Haemophilus parainfluenzae , Klebsiella pneumoniae , and HRV, were consistently detected across all age groups, indicating their potential role as common etiologic agents in adult SARI. In younger adults (18–30 years), pathogens such as HCoV-HKU1, HMPV, and Mycoplasma pneumoniae were more frequently identified. Conversely, Acinetobacter baumannii was predominantly found in the elderly cohort, likely reflecting higher rates of healthcare-associated exposures in this age group. In the intermediate age group (31–64 years), increased frequencies of HCoV-OC43 and Klebsiella aerogenes were observed. Despite these trends, no single pathogen was uniformly dominant across all age groups. Microbial diversity analysis To determine whether microbial community diversity correlates with clinical severity and may serve as an indicator of disease progression, we analyzed alpha and beta diversity among mild and severe SARI cases. Alpha diversity metrics, including species richness and evenness, were assessed using the Chao1 and Shannon indices. For the mild group, the median Chao1 index was 3 (IQR: 2–7), and the Shannon index was 0.45 (IQR: 0.14–0.88). In the severe group, the corresponding values were 3 (IQR: 2–6) and 0.45 (IQR: 0.22–0.90), respectively. No statistically significant differences were observed between the groups for either index (Chao1: p = 0.85; Shannon: p = 0.49) (Supplementary Figure 2A). Beta diversity showed no distinct clustering by severity category, indicating similar microbial community structures across groups (R² = 0.013, p = 0.142) (Supplementary Figure 2B). These results suggest that disease severity in this adult SARI cohort is not driven by broad shifts in microbial diversity. Instead, severity may be more closely associated with the presence of specific pathogenic taxa or host factors. Pathogen-Specific Associations with Outcomes To explore whether specific microbial signatures correlate with disease severity, we conducted a heatmap-based analysis of species-level relative abundance across fungal, bacterial, and viral taxa (Figure 5). The results revealed substantial inter-patient variability, with diverse viral and bacterial profiles observed throughout the cohort. Viral pathogens, including HMPV, HPIV, HCoV, IAV, and HRV, were commonly detected in patients with mild diseases. In contrast, individuals with pneumonia or those requiring special care frequently exhibited enrichment of nosocomial or opportunistic bacterial pathogens such as Klebsiella pneumoniae , Haemophilus parainfluenzae , Klebsiella aerogenes , Acinetobacter baumannii , and Staphylococcus haemolyticus , suggesting possible bacterial mono- or super-infections contributing to increased disease severity. SARS-CoV-2 was identified in two fatal cases, and Klebsiella pneumoniae was detected in one of these, underscoring their potential role in adverse outcomes. Figure 5. Distribution and relative abundance of respiratory pathogens across clinical outcomes in adult SARI cases. Heatmap showing the relative abundance of fungal, bacterial, and viral taxa across SARI patient samples. Columns represent individual patient samples, annotated by clinical outcomes (mild, need special care, bronchiolitis, pneumonia, and death). Each row represents the identified species. Microbial taxa are grouped into fungal, bacterial, and viral domains and arranged based on hierarchical clustering. Color gradients indicate relative abundance per sample after host-read removal and normalization. Risk factors associated with disease severity in SARI patients To identify risk factors associated with severe outcomes among SARI patients, we conducted both univariable and multivariable logistic regression analysis incorporating demographic features, comorbidities, vaccination status, and clinical parameters (Supplementary Table 1). Male sex was significantly associated with increased risk of severe disease compared to female sex (odds ratio (OR) = 3.56, 95% confidence interval (CI): 1.09–13.21, p = 0.043), suggesting possible underlying immunological or behavioral factors influencing susceptibility. In contrast, higher oxygen saturation levels were strongly protective, with each unit increase associated with reduced odds of severe illness (OR = 0.54 per unit increase, 95% CI: 0.39–0.71, p < 0.001) (Figure 6). Notably, influenza vaccination status was not significantly associated with reduced severity in this cohort. Other clinical factors, including pre-existing pulmonary disease, hypertension, and overall comorbidity burden, also did not demonstrate statistically significant associations with severe disease. Figure 6. Multivariable logistic regression analysis for factors associated with severe outcomes among SARI patients in Northern Thailand. The forest plot illustrates odds ratios (OR) with 95% confidence intervals on a logarithmic scale. Variables include patient demographics, vaccination status, comorbidities, and clinical parameters. A vertical dashed line represents an OR of 1.00, indicating no association. Increased risk of disease severity was observed among male patients, whereas higher oxygen saturation was strongly associated with reduced disease severity. Phylogenetic characterization of rhinovirus To characterize the circulating human rhinovirus strains associated with adult SARI during the 2023–2024 period, we conducted phylogenetic analysis on all HRV-positive samples (n=8). The resulting maximum-likelihood tree revealed distinct clustering of HRV strains into three species: HRV-A (n=3), HRV-B (n=1), and HRV-C (n=4), as illustrated in Figure 7. Among the HRV-C strains, 4 clinical samples were grouped into genotypes C22 and C42. The 3 HRV-A sequences clustered within the HRV-A1 subgroup, aligning with the closest reference sequences. One sample mapped to the HRV-B lineage and showed close phylogenetic proximity to HRV-B48. Figure 7. Phylogenetic analysis of HRV-positive samples among SARI patients. A maximum-likelihood tree demonstrates distinct clustering of HRV-A (blue), HRV-B (green), and HRV-C (red) clades. Samples from this study (marked with red triangles) were grouped with established prototype strains, including HRV-C42, HRV-C22, HRV-A1, and HRV-B48. Reference genomes are shown in bold letters. Enterovirus D68 (EV-D68) was included as an outgroup. Bootstrap values are displayed along branches. Discussion This study highlights the substantial clinical and diagnostic utility of long-read mNGS for comprehensive pathogen detection in adult patients hospitalized with SARI in northern Thailand. An optimized nanopore-based sequencing workflow applied to sputum specimens achieved a diagnostic yield of 78%, significantly outperforming multiplex PCR (47%) and culture (40%). These results underscore the advantages of unbiased sequencing approaches for identifying both expected and atypical pathogens, especially in settings where conventional diagnostics are limited by targeted panels or low sensitivity. Notably, mNGS identified additional pathogens in 18% of cases that were negative by both PCR and culture, including clinically relevant organisms such as Klebsiella pneumoniae, Haemophilus parainfluenzae, Pseudomonas aeruginosa , and even SARS-CoV-2. This emphasizes the limitations of targeted assays and supports the role of mNGS in uncovering unanticipated etiologies. The detection of SARS-CoV-2 in PCR-negative samples may reflect genomic variation affecting primer-binding regions, highlighting the need for continuous molecular surveillance [33, 34]. Conversely, certain pathogens detected by culture or PCR were missed by mNGS, likely due to low microbial biomass, high host background, or degraded nucleic acids [35, 36]. Concordance between mNGS and multiplex PCR was particularly strong for common respiratory viruses, and genome-wide coverage profiles confirmed high-confidence detection, including near-complete coverage for HRV, HMPV, and coronaviruses. Even low-abundance targets such as Mycoplasma pneumoniae exhibited specific, although discontinuous, coverage patterns consistent with true positives. Mycoplasma pneumoniae is difficult to culture and often requires molecular tests such as PCR assays for diagnosis [37]. The ability of mNGS to detect this atypical pathogen without prior targeting is useful, particularly for community-acquired atypical pneumonia [38]. Age-stratified analysis revealed distinct pathogen patterns. Younger adults exhibited elevated detection of HMPV, HCoV-HKU1, and Mycoplasma pneumoniae , pathogens commonly associated with community-acquired respiratory infections [39, 40]. Middle-aged adults showed moderate frequency of HCoV-OC43 and Klebsiella aerogenes , while the elderly demonstrated broader pathogen diversity, with increased detection of Acinetobacter baumannii , a healthcare-associated infection [41]. The identification of Haemophilus parainfluenzae and Klebsiella pneumoniae in all age groups indicates their role as ubiquitous opportunistic pathogens in SARI. These patterns indicate varying exposures, immune responses, and comorbidity burdens among different age groups, which may influence respiratory microbial colonization and infection risk. Among viral etiologies, HRV was a predominant pathogen. Phylogenetic analysis indicated the circulation of HRV-A, HRV-B, and HRV-C, with HRV-C being the most common, consistent with report from Thailand showing a high prevalence of HRV-C in SARI patients [42]. HRV-C has been implicated in severe lower respiratory tract infections, particularly in individuals with chronic lung disease or immunosuppression [43, 44]. These findings highlight the potential of mNGS for molecular surveillance of circulating HRV strains. Despite detailed microbial profiling, alpha and beta diversity analyses showed no significant differences between mild and severe cases, suggesting that overall microbial diversity is not a major determinant of disease severity in adult SARI. Instead, severity appeared to be associated with specific pathogenic taxa and host factors. In severe and fatal cases, specific pathogenic taxa, such as Klebsiella pneumoniae, Acinetobacter baumannii, Staphylococcus haemolyticus, and Klebsiella aerogenes , were commonly found. These patterns align with viral–bacterial synergy, wherein viral infection facilitates secondary bacterial invasion [45, 46]. Notably, SARS-CoV-2 was detected in two fatal cases and Klebsiella pneumoniae in one, emphasizing their pathogenic significance in SARI mortality. Multivariable analysis identified male sex and decreased oxygen saturation as independent predictors of severe disease. The association between being male and having worse outcomes may be due to differences in the immune system, hormones, or other behavioral factors [47, 48]. Hypoxemia reflects impaired gas exchange and thus serves as a clinically relevant biomarker for risk stratification and determinations related to ICU admission or mechanical ventilation [49]. Overall, these findings support the integration of long-read mNGS into respiratory diagnostics and public health surveillance, enabling detection of unculturable or unexpected pathogens and facilitating real-time monitoring of pathogen evolution. However, limitations include the single-center design, limited ability of mNGS to distinguish colonization from infection without clinical context, and the absence of healthy controls to define baseline respiratory microbiota. Future multicenter studies incorporating healthy controls and integrative multi-omics approaches will be essential to refine pathogen interpretation and advance understanding of host-pathogen interactions in SARI. Conclusions This study demonstrates that long-read metagenomic sequencing is a robust and comprehensive diagnostic platform for the etiological investigation of severe acute respiratory infections (SARI) in adults. By enabling the detection of a broad spectrum of viral, bacterial, and atypical pathogens, including those frequently missed by conventional methods, mNGS significantly enhances diagnostic yield and clinical insight. Its capacity to identify co-infections and provide strain-level resolution further underscores its value in both clinical diagnostics and infectious disease surveillance. Importantly, the severity of SARI was more closely associated with specific pathogens and host-related factors than with overall microbial diversity, emphasizing the need for pathogen-specific and host-aware clinical management strategies. These findings support the integration of long-read mNGS into respiratory infection diagnostics and real-time pathogen surveillance frameworks. Future research incorporating healthy controls, multi-omics approaches, and multicenter cohorts will be critical to refine the clinical interpretation of mNGS data and to advance its application in outbreak preparedness and precision infectious disease management. Acknowledgements We gratefully acknowledge Nakornping hospital for providing clinical specimens, and the LUCENT international collaboration, Faculty of Associated Medical Sciences, Chiang Mai University, for providing laboratory facilities and technical support. N.K.K. was a candidate in PhD Program in Biomedical Sciences, Faculty of Associated Medical Sciences, Chiang Mai University, under the CMU Presidential Scholarship. Author contributions W.K. secured funding and conceptualized the study; R.K., P.T., and K.P. coordinated sample collection; N.K.K. conducted the experiments and analyzed the data; W.K., S.H., and N.N. managed the resources; W.K. and N.N. investigated and validated the data; N.K.K., prepared the original draft of the manuscript; W.K., N.N., and S.H. supervised the study; all authors have reviewed and approved the final version of the manuscript. Data availability The datasets generated during the current study are not publicly available due to potentially identifiable human genetic material and cannot be deposited in a public repository according to institutional and ethical restrictions but de-identified processed data are available from the corresponding author on reasonable request. Funding This study was supported by Chiang Mai University, Thailand. Conflict of interest The authors declare no conflict of interest. References 1. Chen C, You Y, Du Y, Zhou W, Jiang D, Cao K, et al. Global epidemiological trends in the incidence and deaths of acute respiratory infections from 1990 to 2021. Heliyon. 2024;10(16):e35841. doi:10.1016/j.heliyon.2024.e35841.2. Williams BG, Gouws E, Boschi-Pinto C, Bryce J, Dye C. Estimates of world-wide distribution of child deaths from acute respiratory infections. Lancet Infect Dis. 2002;2(1):25-32. doi:10.1016/s1473-3099(01)00170-0.3. Steffens A, Friedlander H, Como-Sabetti K, Boxrud D, Bistodeau S, Strain A, et al. Comparison of Respiratory Pathogen Detections from Routine Hospital Testing and Expanded Systematic Testing from the Minnesota Severe Acute Respiratory Illness Surveillance Program, 2015–2016. (2328-8957 (Electronic)). 4. Zhao Y, Zhang W, Zhang X. Application of metagenomic next-generation sequencing in the diagnosis of infectious diseases. Front Cell Infect Microbiol. 2024;14:1458316. doi:10.3389/fcimb.2024.1458316.5. van Boheemen S, van Rijn AL, Pappas N, Carbo EC, Vorderman RHP, Sidorov I, et al. Retrospective Validation of a Metagenomic Sequencing Protocol for Combined Detection of RNA and DNA Viruses Using Respiratory Samples from Pediatric Patients. J Mol Diagn. 2020;22(2):196-207. doi:10.1016/j.jmoldx.2019.10.007.6. Phan MVT, Agoti CN, Munywoki PK, Otieno GP, Ngama M, Kellam P, et al. Identification of missed viruses by metagenomic sequencing of clinical respiratory samples from Kenya. Sci Rep. 2022;12(1):202. doi:10.1038/s41598-021-03987-1.7. Yang J, Li J, Zhang L, Shen Z, Xiao Y, Zhang G, et al. Highly diverse sputum microbiota correlates with the disease severity in patients with community-acquired pneumonia: a longitudinal cohort study. Respir Res. 2024;25(1):223. doi:10.1186/s12931-024-02821-2.8. Langevin S, Pichon M, Smith E, Morrison J, Bent Z, Green R, et al. Early nasopharyngeal microbial signature associated with severe influenza in children: a retrospective pilot study. J Gen Virol. 2017;98(10):2425-37. doi:10.1099/jgv.0.000920.9. Vidaur L, Guridi A, Leizaola O, Marin J, Rello J, Sarasqueta C, et al. Respiratory dysbiosis as prognostic biomarker of disease severity for adults with community-acquired pneumonia requiring mechanical ventilation. Pneumonia (Nathan). 2025;17(1):10. doi:10.1186/s41479-025-00163-1.10. Branche A, Ramesh M, Francis B. A Narrative Review of Key Risk Factors for Severe Illness Following SARS-CoV-2, Influenza Virus, and Respiratory Syncytial Virus Infection. Infect Dis Ther. 2025;14(Suppl 1):39-61. doi:10.1007/s40121-024-01081-3.11. Kang SH, Cheong HJ, Song JY, Noh JY, Jeon JH, Choi MJ, et al. Analysis of Risk Factors for Severe Acute Respiratory Infection and Pneumonia and among Adult Patients with Acute Respiratory Illness during 2011-2014 Influenza Seasons in Korea. Infect Chemother. 2016;48(4):294-301. doi:10.3947/ic.2016.48.4.294.12. Chittaganpitch M, Waicharoen S, Yingyong T, Praphasiri P, Sangkitporn S, Olsen SJ, et al. Viral etiologies of influenza-like illness and severe acute respiratory infections in Thailand. Influenza Other Respir Viruses. 2018;12(4):482-9. doi:10.1111/irv.12554.13. Kandeel A, Fahim M, Deghedy O, W HR, M KK, El Shesheny R, et al. Multicenter study to describe viral etiologies, clinical profiles, and outcomes of hospitalized children with severe acute respiratory infections, Egypt 2022. Sci Rep. 2023;13(1):21860. doi:10.1038/s41598-023-48814-x.14. Milucky J, Pondo T, Gregory CJ, Iuliano D, Chaves SS, McCracken J, et al. The epidemiology and estimated etiology of pathogens detected from the upper respiratory tract of adults with severe acute respiratory infections in multiple countries, 2014–2015. PLoS One. 2020;15(10):e0240309. doi:10.1371/journal.pone.0240309.15. Fitzner J, Qasmieh S, Mounts AW, Alexander B, Besselaar T, Briand S, et al. Revision of clinical case definitions: influenza-like illness and severe acute respiratory infection. Bull World Health Organ. 2018;96(2):122-8. doi:10.2471/blt.17.194514.16. Kham-Kjing N, Ngo-Giang-Huong N, Khamduang W. Sputum Respiratory Pathogen Genomic Surveillance: A Practical Approach for Long-Read Metagenomic Sequencing. bioRxiv. 2025:2025.12.08.692896. doi:10.64898/2025.12.08.692896.17. De Coster W, Rademakers R. NanoPack2: population-scale evaluation of long-read sequencing data. Bioinformatics. 2023;39(5). doi:10.1093/bioinformatics/btad311.18. Martin M. Cutadapt removes adapter sequences from high-throughput sequencing reads. 2011. 2011;17(1):3. doi:10.14806/ej.17.1.200.19. Chen S. Ultrafast one-pass FASTQ data preprocessing, quality control, and deduplication using fastp. iMeta. 2023;2(2):e107. doi:10.1002/imt2.107.20. Li H. New strategies to improve minimap2 alignment accuracy. Bioinformatics. 2021;37(23):4572-4. doi:10.1093/bioinformatics/btab705.21. Kolmogorov M, Bickhart DM, Behsaz B, Gurevich A, Rayko M, Shin SB, et al. metaFlye: scalable long-read metagenome assembly using repeat graphs. Nat Methods. 2020;17(11):1103-10. doi:10.1038/s41592-020-00971-x.22. Oxford Nanopore Technologies Ltd. medaka: Sequence correction provided by ONT Research [Available from: https://github.com/nanoporetech/medaka].23. Lu J, Rincon N, Wood DE, Breitwieser FP, Pockrandt C, Langmead B, et al. Metagenome analysis using the Kraken software suite. Nat Protoc. 2022;17(12):2815-39. doi:10.1038/s41596-022-00738-y.24. Ondov BD, Bergman NH, Phillippy AM. Interactive metagenomic visualization in a Web browser. BMC Bioinformatics. 2011;12:385. doi:10.1186/1471-2105-12-385.25. Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, et al. Twelve years of SAMtools and BCFtools. GigaScience. 2021;10(2). doi:10.1093/gigascience/giab008.26. Katoh K, Standley DM. MAFFT Multiple Sequence Alignment Software Version 7: Improvements in Performance and Usability. Molecular Biology and Evolution. 2013;30(4):772-80. doi:10.1093/molbev/mst010.27. Capella-Gutiérrez S, Silla-Martínez JM, Gabaldón T. trimAl: a tool for automated alignment trimming in large-scale phylogenetic analyses. Bioinformatics. 2009;25(15):1972-3. doi:10.1093/bioinformatics/btp348.28. Minh BQ, Schmidt HA, Chernomor O, Schrempf D, Woodhams MD, von Haeseler A, et al. IQ-TREE 2: New Models and Efficient Methods for Phylogenetic Inference in the Genomic Era. Molecular Biology and Evolution. 2020;37(5):1530-4. doi:10.1093/molbev/msaa015.29. Hoang DT, Chernomor O, von Haeseler A, Minh BQ, Vinh LS. UFBoot2: Improving the Ultrafast Bootstrap Approximation. Molecular Biology and Evolution. 2017;35(2):518-22. doi:10.1093/molbev/msx281.30. Kalyaanamoorthy S, Minh BQ, Wong TKF, von Haeseler A, Jermiin LS. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat Methods. 2017;14(6):587-9. doi:10.1038/nmeth.4285.31. Yu G, Smith DK, Zhu H, Guan Y, Lam TT-Y. ggtree: an r package for visualization and annotation of phylogenetic trees with their covariates and other associated data. Methods Ecol Evol. 2017;8(1):28-36. doi:10.1111/2041-210X.12628.32. Harrison E, Drake T, Pius R. finalfit: Quickly Create Elegant Regression Results Tables and Plots when Modelling. R package version 1.1.0, [cited 2025 8 December,]. Available from: https://github.com/ewenharrison/finalfit].33. Mentes A, Papp K, Visontai D, Stéger J, Csabai I, Papp K, et al. Identification of mutations in SARS-CoV-2 PCR primer regions. Sci Rep. 2022;12(1):18651. doi:10.1038/s41598-022-21953-3.34. Khan KA, Cheung P. Presence of mismatches between diagnostic PCR assays and coronavirus SARS-CoV-2 genome. Royal Society Open Science. 2020;7(6):200636. doi:doi:10.1098/rsos.200636.35. Zhang H, Zhou F, Liu X, Huang J. Clinical application of metagenomic next-generation sequencing in patients with different organ system infection: A retrospective observational study. Medicine (Baltimore). 2024;103(4):e36745. doi:10.1097/md.0000000000036745.36. Xu J, Zhou P, Liu J, Zhao L, Fu H, Han Q, et al. Utilizing Metagenomic Next-Generation Sequencing (mNGS) for Rapid Pathogen Identification and to Inform Clinical Decision-Making: Results from a Large Real-World Cohort. Infectious Diseases and Therapy. 2023;12(4):1175-87. doi:10.1007/s40121-023-00790-5.37. Viz-Lasheras S, Gómez-Carballa A, Bello X, Rivero-Calle I, Dacosta AI, Kaforou M, et al. A diagnostic host-specific transcriptome response for Mycoplasma pneumoniae pneumonia to guide pediatric patient treatment. Nat Commun. 2025;16(1):673. doi:10.1038/s41467-025-55932-9.38. Miyashita N. Atypical pneumonia: Pathophysiology, diagnosis, and treatment. Respiratory Investigation. 2022;60(1):56-67. doi:10.1016/j.resinv.2021.09.009.39. Howard LM, Edwards KM, Zhu Y, Grijalva CG, Self WH, Jain S, et al. Clinical Features of Human Metapneumovirus-Associated Community-acquired Pneumonia Hospitalizations. Clin Infect Dis. 2021;72(1):108-17. doi:10.1093/cid/ciaa088.40. Kumar S. Mycoplasma pneumoniae : A significant but underrated pathogen in paediatric community-acquired lower respiratory tract infections. Indian J Med Res. 2018;147(1):23-31. doi:10.4103/ijmr.IJMR_1582_16.41. Javadi K. Acinetobacter baumannii in the Healthcare Facility Setting. Avicenna J Clin Microbiol Infect. 2024;11(3):141-6. doi:10.34172/ajcmi.3500.42. Wanlapakorn N, Thongpan I, Sarawanangkoor N, Vichaiwattana P, Auphimai C, Srimuan D, et al. Epidemiology and clinical characteristics of severe acute respiratory infections among hospitalized children under 5 years of age in a tertiary care center in Bangkok, Thailand, 2019-2020. Heliyon. 2023;9(11):e22300. doi:10.1016/j.heliyon.2023.e22300.43. Galindo-Fraga A, Guerra-de-Blas PDC, Ortega-Villa AM, Mateja A, Ruiz Quiñones JA, Ramos Cervantes P, et al. Different Clinical Presentations of Human Rhinovirus Species Infection in Children and Adults in Mexico. Open Forum Infect Dis. 2022;9(7):ofac303. doi:10.1093/ofid/ofac303.44. Kan OK, Washio Y, Fujimoto T, Shiroyama N, Nakano T, Wakamatsu K, et al. Differences in the spectrum of respiratory viruses and detection of human rhinovirus C in exacerbations of adult asthma and chronic obstructive pulmonary disease. Respir Investig. 2022;60(1):129-36. doi:10.1016/j.resinv.2021.08.009.45. Manna S, Baindara P, Mandal SM. Molecular pathogenesis of secondary bacterial infection associated to viral infections including SARS-CoV-2. Journal of Infection and Public Health. 2020;13:1397 - 404. 46. Shenoy B, Reddy HT. Mechanisms of Secondary Bacterial Infections in Viral Infections. Pediatr Infect Dis. 2023. 47. Takahashi T, Iwasaki A. Sex differences in immune responses. Science. 2021;371(6527):347-8. doi:doi:10.1126/science.abe7199.48. Taslem Mourosi J, Anwar S, Hosen MJ. The sex and gender dimensions of COVID-19: A narrative review of the potential underlying factors. Infection, Genetics and Evolution. 2022;103:105338. doi:10.1016/j.meegid.2022.105338.49. Midez R, Fehlmann CA, Marti C, Larribau R, Rouyer F, Boroli F, et al. Association between Prehospital Hypoxemia and Admission to Intensive Care Unit during the COVID-19 Pandemic: A Retrospective Cohort Study. Medicina. 2021;57(12):1362. Information & Authors Information Version history V1 Version 1 17 December 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords infection pathogenesis respiratory tract virus classification Authors Affiliations Nang Kham-Kjing 0000-0001-8053-9468 Chiang Mai University View all articles by this author Rathakarn Kawila Nakornping Hospital View all articles by this author Patcharaporn Tariyo Nakornping Hospital View all articles by this author Kittiyaporn Puapun Nakornping Hospital View all articles by this author Nicole Ngo-Giang-Huong Chiang Mai University View all articles by this author Sayamon Hongjaisee Chiang Mai University View all articles by this author Woottichai Khamduang 0000-0003-0196-5691 [email protected] Chiang Mai University View all articles by this author Metrics & Citations Metrics Article Usage 197 views 126 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Nang Kham-Kjing, Rathakarn Kawila, Patcharaporn Tariyo, et al. Comprehensive Pathogen Profiling in Adult Patients with Severe Acute Respiratory Infections Using Metagenomic Next-Generation Sequencing of Sputum Samples. Authorea . 17 December 2025. DOI: https://doi.org/10.22541/au.176599276.68130321/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.176599276.68130321/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9ffcb56fbb45e2c5',t:'MTc3OTQ2MjIxOA=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();