Comparative performance of three next-generation sequencing techniques in real clinical lower respiratory tract infections

doi:10.21203/rs.3.rs-4838495/v1

Download PDF

Research Article

Comparative performance of three next-generation sequencing techniques in real clinical lower respiratory tract infections

https://doi.org/10.21203/rs.3.rs-4838495/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background

Lower respiratory tract infections, notorious for high mortality, are inadequately addressed by traditional diagnostics, highlighting the need for more effective methods. The advent of next-generation sequencing (NGS) offers a promising solution. This study evaluates the performance of three NGS methodologies—metagenomic NGS (mNGS), amplification-based targeted NGS (tNGS), and capture-based tNGS—in identifying pathogens in bronchoalveolar lavage fluid.

Methods

We compared these methods against conventional microbiological tests and comprehensive clinical diagnosis in 205 patients, focusing on sensitivity, specificity, and pathogen detection capabilities.

Results

Capture-based tNGS demonstrated the highest sensitivity (99.43%) and positivity (90.73%), significantly outperforming the others in samples negative by conventional tests. While mNGS showed broader pathogen coverage, it underperformed in detecting RNA viruses. Amplification-based tNGS, constrained by primer and panel design, missed certain bacteria and DNA viruses. Both tNGS methods effectively identified SARS-CoV-2 genotypes, with capture-based tNGS providing more detailed distinctions. The study also detected several antimicrobial resistance genes and virulence factors, indicating a broader spectrum of pathogen identification by capture-based tNGS.

Conclusion

These findings suggest that the choice of NGS method should be tailored to specific clinical needs and objectives, with capture-based tNGS showing superior diagnostic utility.

metagenomic next-generation sequencing

amplification-based targeted next-generation sequencing

capture-based targeted next-generation sequencing

performance comparison

lower respiratory tract infection

diagnostic efficacy

Lower respiratory tract infections are a leading cause of mortality from infectious diseases worldwide (1). The inability to precisely identify pathogens in respiratory infections often hinders the effective administration of targeted drug therapies. This delay can aggravate the patient's condition, leading to worsening symptoms or, in severe cases, death. Diagnosing and managing certain persistent infections of the lower respiratory tract poses significant challenges (2). In such cases, accurate identification of the pathogen and subsequent targeted treatment are of critical importance.

However, in China, nearly half of all pulmonary infection cases remain etiologically undiagnosed (3). This is because the clinical conventional tests such as traditional microbial culture, antigen and/or antibody immunological methods, and polymerase chain reaction (PCR) detection can be time-consuming and have a low detection rate. They require extensive resources and may not identify fastidious pathogens, rare and atypical pathogens, or pathogens that are no longer viable due to antimicrobial therapy (4).

Next-generation sequencing (NGS) is a high-throughput method capable of directly detecting nucleic acids in clinical samples and generating millions to billions of reads per instrument run. Metagenomic next-generation sequencing (mNGS), one of the most widely recognized applications of NGS, has been progressively employed in clinical scenarios for identifying unknown infections, rare, and atypical pathogens (5–7). Additionally, a new variant of NGS, targeted next-generation sequencing (tNGS), has recently emerged. mNGS workflows aim to sequence as much DNA and/or RNA as possible from a sample, whereas tNGS workflows focus on enriching specific genetic targets for sequencing (4). Currently, tNGS enriches target pathogens primarily through primer amplification and probe capture techniques. All three NGS assays have been reported in studies on clinical applications, with most demonstrating significant value in aiding clinical diagnosis (8–12). However, no study has directly compared the efficacy of pathogen identification between mNGS, amplification-based tNGS, and capture-based tNGS. The distinct advantages and characteristics of each NGS method remain unclear, posing challenges for clinicians in selecting the most appropriate detection approach.

This study investigates the real-world performance of mNGS, amplification-based tNGS, and capture-based tNGS, comparing their respective advantages and disadvantages. It aims to guide clinicians in making more informed decisions for detection, thereby enhancing clinical diagnosis, disease management, and public health.

Study design and population

All patients with suspected pneumonia admitted to the Department of Respiratory and Critical Care Medicine at the Second Xiangya Hospital of Central South University from January 2023 to October 2023 were retrospectively investigated. Patients were eligible for enrolment if they (1) were at least 18 years of age; (2) suspected with pneumonia; (3) had BALF samples available for conventional microbiological tests (CMTs) and mNGS within 48h after admission. A comprehensive clinical diagnosis for each patient judged by more than two clinicians was collected. Patients without a clear comprehensive clinical diagnosis were excluded from this study. Patients with suspected pneumonia fulfilled the following criteria: (1) at least one compatible symptom, such as new-onset fever, cough, or dyspnoea; (2) new-onset radiological findings on chest images.

Sample collection

Bronchoalveolar lavage fluid (BALF) samples were collected from eligible patients and stored in sterile screw-capped cryovials. A volume of 5–10 mL was collected from each patient and divided equally into three for mNGS, amplification-based tNGS, and capture-based tNGS. These samples were kept at ≤ − 20 ℃ during transportation, carefully handled and preserved to maintain the quality of the genetic material.

Metagenomic next-generation sequencing

DNA was extracted from all samples using a QIAamp® UCP Pathogen DNA Kit (Qiagen) according to the manufacturer's instructions. Human DNA was removed using Benzonase (Qiagen) and Tween20 (Sigma). The QIAamp UCP pathogen mini kit (Qiagen, Valencia, CA, USA) was applied to extract total RNA. RNA was reversely transcribed, and amplified by Ovation RNA-Seq system (NuGEN, CA, USA). Following fragmentation, the library was constructed using Ovation Ultralow System V2 (NuGEN, CA, USA) and the concentration of the library was measured using Qubit. Sequencing was executed on Illumina Nextseq 550Dx with 75-bp single-end reads. For negative controls, peripheral blood mononuclear cell samples with 10⁵ cells/mL from healthy donors in parallel were prepared with each batch, using the same protocol, and sterile deionized water was extracted alongside the specimens to serve as non-template controls (NTC).

Raw sequencing data was processed using fastp to remove reads containing adapters or ambiguous "N" nucleotides and low-quality reads. Low-complexity reads were removed by Kcomplexity with default parameters. Human sequence data were identified and excluded by mapping to a human reference genome (hg38) using Burrows-Wheeler Aligner software. Microbial reads were then aligned to database with SNAP v1.0 beta.18. Approximately 20 million reads were generated for each sample. For pathogen with background reads in negative control, a positive detection was reported for a given species or genus if the reads per million (RPM) ratio was ≥ 10, where the RPM ratio was defined as the RPMsample/ RPMNTC. For pathogen without background reads in negative control, RPM was set ≥ 0.05. A penalty of 5% and 10% was used for species and genus, respectively.

Amplification-based targeted next-generation sequencing

A volume of 650 µL of the sample was liquefied by combining it with an equal volume of 80mmol/L dithiothreitol in a 1.5mL centrifuge tube. Meanwhile, a positive control and a negative control were set up to monitor the whole experiment process of tNGS. Subsequently, 500 µL of the homogenate was utilized for total nucleic acid extraction and purification via the MagPure Pathogen DNA/RNA Kit (R6672-01B, Magen, Guangzhou, China), following the manufacturer’s protocol. The library was constructed using the Respiratory Pathogen Detection Kit (KS608-100HXD96, KingCreate, Guangzhou, China), and a no template control was set up to monitor the library construction and sequencing process. This process encompassed two rounds of PCR amplification. The sample nucleic acid and cDNA were employed as templates, and a set of 198 microorganism-specific primers were selected for ultra-multiplex PCR amplification to enrich the target pathogen sequences, spanning bacteria, viruses, fungi, mycoplasma, and chlamydia. After the amplification, PCR products underwent purification with beads, followed by amplification using primers containing sequencing adapters and distinct barcodes. The quality and quantity of the constructed library were evaluated using the Qsep100 Bio-Fragment Analyzer (Bioptic, Taiwan, China) and Qubit 4.0 fluorometer (Thermo Scientific, Massachusetts, United States), respectively. The library was subsequently subjected to sequencing on an Illumina MiniSeq platform with each library yielded approximately 0.1 million reads, with a sequencing read length of single-end 100bp.

Sequencing data were analyzed using our in-house developed analysis pipeline. The raw data underwent initial identification via the adapter. Reads with single-end lengths exceeding 50bp were retained, followed by low-quality filtering to retain reads with Q30 > 75%, ensuring high-quality data. The single-ended aligned reads were then compared using the Self-Building clinical pathogen database to determine the read count of specific amplification targets in each sample. The reference sequences used for read mapping was a database curated from different sources, including Genbank database, Refseq database, and Nucleotide database from NCBI.2.5.

Capture-based targeted next-generation sequencing

BALF specimens were collected, to which a concoction of lysis buffer, protease K, and binding buffer was added within a grinding tube. This mixture was subjected to mechanical disruption via a shock breaker for a duration of 30 seconds. Following this, a simultaneous extraction of DNA and RNA was performed utilizing the VAMNE Magnetic Pathogen DNA/RNA Extraction Kit (Vazyme, Nanjing, China). Nucleic acid quantification was conducted employing a Qubit 3.0 fluorometer, utilizing the double-stranded DNA (dsDNA) and RNA high sensitivity assay kits. For the synthesis of cDNA and the preparation of the sequencing library, the HieffNGS®C37P4 OnePot cDNA & gDNA Library Prep Kit (Yeasen, Shanghai, China) was utilized, adhering to the manufacturer's instructions. Target sequence enrichment was achieved by incubating the samples with GenePlus probes for an approximate duration of 4 hours, followed by an 18-cycle PCR amplification of the captured products. The amplified products were then processed to form DNA nanoballs (DNBs). Sequencing was executed on the DNBSEQ-G99 platform with 100-bp single-end reads, targeting a sequencing depth of 5 million reads to ensure thorough coverage of the targeted regions.

For the analysis of sequencing data, initial preprocessing of raw data involved the removal of low-quality sequences, residual adapters, and reads of insufficient length. Sequences corresponding to microbial rRNA and human genomic material were also excluded. The refined data set was then aligned and annotated against a comprehensive database of pathogenic microorganisms, leveraging BLAST software for sequence comparison. Reads aligning with the target capture regions were classified as target reads and normalized to reads per million to facilitate quantitative comparisons. The resultant data provided a detailed profile of pathogens present in the sample, facilitating the identification of potential causative agents of infection.

Statistical analysis

Scoring structure for NGS comparisons to clinical testing is referenced from Karius (13). Categorical variables were compared using the Kappa test and Cochran's Q test. Paired nonparametric variables were compared using the Wilcoxon test and Friedman test. Unpaired nonparametric variables were compared using the Mann-Whitney test. All tests were two-tailed and significance was set at 5%. All figures were drawn using R version 4.3.1 and GraphPad Prism version 9.5.0 for Windows (GraphPad Software LLC., San Diego, CA, USA). All analyses were performed with SPSS version 26.0 for Windows (SPSS Inc., Chicago, Illinois, USA).

In this study, a total of 205 BALF samples were collected from patients diagnosed with lower respiratory tract infections (Supplementary Fig. 1). Each sample underwent testing using mNGS, amplification-based tNGS, capture-based tNGS, and conventional microbiological tests (CMTs). Furthermore, more than two clinicians independently conducted a comprehensive diagnosis for each case. Notably, among these samples, only 40 were tested using both DNA mNGS and RNA mNGS due to considerations such as clinical needs, cost, and sample size, leaving the remaining 165 samples untested for RNA mNGS.

The performance of NGS, benchmarked against CMTs and the comprehensive clinical diagnosis as standards, is comprehensively detailed in Table 1. Capture-based tNGS demonstrated the greatest consistency with the comprehensive clinical diagnosis, exhibiting the highest rates of positivity, sensitivity, negative predictive value (NPV), and F₁ score. On the other hand, mNGS excels in specificity and positive predictive value (PPV), while the performance of amplification-based tNGS occupies a middle ground. Nonetheless, a comparative analysis, summarized in Table 2, indicates a significant alignment in diagnostic efficacy across the three methods, with no substantial differences observed. This suggests that under certain conditions, the methods could be used interchangeably. Table 3 elaborates on the performance of NGS in CMT-negative cases, evaluated against the comprehensive clinical diagnosis. Here, capture-based tNGS continues to lead in positivity rates, sensitivity, NPV, and F₁ score, albeit with the lowest specificity and PPV. Notably, amplification-based tNGS and mNGS exhibit similar performance, except for a notably higher specificity in the former.

Table 1

Performance of NGS in clinical pathogen detecting.
Standard	Approach	Consistency	Sensitivity	Specificity	PPV	NPV	Positivity	F₁ Score
	mNGS	58.05%	66.15%	44.00%	67.19%	42.86%	62.44%	0.6667
CMT	A_tNGS	60.98%	73.85%	38.67%	67.61%	46.03%	69.27%	0.7059
	C_tNGS	51.22%	71.54%	16.00%	59.62%	24.49%	76.10%	0.6504
	mNGS	90.73%	91.95%	83.87%	96.97%	65.00%	80.49%	0.9439
Diagnosis	A_tNGS	90.24%	92.53%	77.42%	95.83%	64.86%	81.95%	0.9415
	C_tNGS	93.17%	99.43%	58.06%	93.01%	94.74%	90.73%	0.9611

PPV, positive predictive value; NPV, negative predictive value.

Table 2

Consistency and differences comparison of NGS performance.
Test	Approach	Value	P-value
	mNGS vs. A_tNGS	0.504	< 0.001
Kappa	mNGS vs. C_tNGS	0.477	< 0.001
	A_tNGS vs. C_tNGS	0.471	< 0.001
Cochran's Q	mNGS vs. A_tNGS vs. C_tNGS	1.476	0.478

Table 3

Performance of NGS against diagnosis in CMT-negative cases.
Approach	Consistency	Sensitivity	Specificity	PPV	NPV	Positivity	F₁ Score
mNGS	84.00%	84.91%	81.82%	91.84%	69.23%	65.33%	0.8824
A_tNGS	85.33%	84.91%	86.36%	93.75%	70.37%	64.00%	0.8911
C_tNGS	84.00%	100.00%	45.45%	81.54%	100.00%	86.67%	0.8983

Overall, when compared to the comprehensive clinical diagnosis, capture-based tNGS demonstrated the highest sensitivity rate but the lowest PPV. In contrast, amplification-based tNGS exhibited the highest miss rate. Regarding pathogen types, capture-based tNGS showed a high sensitivity rate across all categories, with its lowest PPV primarily observed in DNA viruses, followed by Gram-positive and Gram-negative bacteria. Amplification-based tNGS displayed its highest miss rates with Gram-positive bacteria, followed by Gram-negative bacteria and DNA viruses. Meanwhile, mNGS significantly underperformed in detecting RNA viruses (Fig. 1). Among the 505 pathogens identified through the comprehensive clinical diagnosis, 264 (52.3%) were detected by all three NGS techniques. These primarily included Acinetobacter baumannii, Klebsiella pneumoniae, Candida albicans, Pneumocystis jirovecii, Aspergillus fumigatus, Human betaherpes virus 5 (HHV-5), Human gammaherpes virus 4 (HHV-4), Human alphaherpes virus 1 (HHV-1), among others. The remaining 239 (47.3%) pathogens were detected by only two of the NGS techniques (Fig. 2A). In these cases, mNGS demonstrated high sensitivity for Enterococcus faecium, Haemophilus parainfluenzae, Corynebacterium striatum, Streptococcus mitis, C. albicans, Stenotrophomonas maltophilia, and P. jirovecii, but exhibited a high miss rate for SARS-CoV-2, HHV-1, Haemophilus influenzae, and A. fumigatus. Amplification-based tNGS showed high sensitivity in SARS-CoV-2, C. albicans, H. influenzae, and A. fumigatus, but had a high miss rate for E. faecium, HHV-4, H. parainfluenzae, C. striatum, HHV-5, S. mitis, HHV-1, S. maltophilia, and P. jirovecii. Capture-based tNGS demonstrated high sensitivity across all pathogens except C. albicans. However, in terms of PPV, capture-based tNGS showed relatively low PPV in HHV-4, C. striatum, HHV-5, and HHV-1 (Fig. 2B).

In this study, a comparative analysis was conducted between mNGS and tNGS at the nucleic acid level, due to the observed low sensitivity of mNGS. Overall, the RPM for pathogens reported by mNGS was significantly lower compared to those reported by capture-based tNGS, across all pathogen types with the exception of fungi (Fig. 3A). This disparity not only highlights the reduced sensitivity of mNGS but also underscores the enhanced nucleic acid enrichment capability of capture-based tNGS. Moreover, while amplification-based tNGS demonstrated the highest RPM for pathogen nucleic acids, this method's reliance on PCR amplification complicates direct comparisons (Supplementary Fig. 2). Specifically, PCR amplification can obfuscate the distinction between original and amplified DNA strands. Moreover, the observed shortfall of mNGS in detecting RNA viruses was initially thought to result from the absence of RNA sequencing in the bulk of our samples. However, even after these samples were excluded and a subsequent reevaluation was conducted, the shortfall remained. This suggests that the limitations are inherent to the mNGS technique itself (Supplementary Fig. 3A).

In-depth analysis was conducted on the pathogens missed by amplification-based tNGS. Firstly, a subset of pathogens fell outside the scope of the amplification-based tNGS panel utilized in this study, constituting 16.34% of the pathogens identified in the comprehensive clinical diagnosis (Supplementary Fig. 3B) (Supplementary Fig. 3C showed the most commonly missed pathogens within the amplification-based tNGS panel). After excluding the pathogens not covered by the panel, significant detection shortfalls persisted in amplification-based tNGS for both Gram-negative bacteria and DNA viruses (Supplementary Fig. 3A). This suggests a potential diagnostic gap in real-world clinical settings. Secondly, the detection of low-abundance pathogens poses a challenge for amplification-based tNGS. Among the pathogens identified by the comprehensive clinical diagnosis, those reported by both amplification-based tNGS and mNGS exhibited significantly higher RPM than those reported by mNGS alone. This discrepancy was particularly pronounced for Gram-negative bacteria and DNA viruses (Fig. 3B).

The lower PPV observed with capture-based tNGS can be attributed to its reports of redundant pathogens not identified in the comprehensive clinical diagnosis. Analysis of common redundant pathogen species was conducted at the nucleic acid level. Pathogens reported exclusively by capture-based tNGS generally displayed significantly lower RPM compared to those identified by both capture-based tNGS and the comprehensive clinical diagnosis (Fig. 3C). Consequently, receiver operating characteristic (ROC) curves were employed to establish improved reporting thresholds, aiming to mitigate the issue of low PPV (Fig. 3D). Subsequent optimization of reporting thresholds for capture-based tNGS markedly enhanced the PPV with the comprehensive clinical diagnosis (Fig. 3E).

In addition, this study assessed the capability to identify pathogen genotypes. For SARS-CoV-2, mNGS was limited to species-level identification, prompting a comparison between two tNGS approaches. Interestingly, among all the SARS-CoV-2 genotypes consistently identified by both tNGS methods, it was always the capture-based tNGS that provided a more refined genotype distinction (Fig. 4A).

Remarkably, examination of the outputs from these two tNGS approaches unveiled the presence of numerous antimicrobial resistance (AMR) genes and virulence factors (VFs). Both tNGS methodologies consistently identified several genes, including KPC, mecA, iucA, peg344, among others (Fig. 4B) (Fig. 4C). However, the capture-based tNGS approach demonstrated superior detection capabilities, identifying a broader spectrum of AMR genes and VFs.

In the current landscape, an array of sophisticated tests leveraging NGS technology has been progressively integrated into clinical settings. Numerous studies have highlighted their diagnostic utility, yet selecting the most suitable NGS assay for practical clinical application remains a formidable challenge for healthcare practitioners. This study conducted a comprehensive comparison among mNGS, amplification-based tNGS, and capture-based tNGS, benchmarked against CMTs, and diagnosis. While no significant differences in diagnostic efficacy were observed among the three NGS approaches, notable distinctions in their detection characteristics and pathogen detection preferences were evident.

Previous studies have already noted the diagnostic capabilities of the three NGS methodologies as largely equivalent. An investigation evaluated the performance of a commercially manufactured tNGS workflow and a complementary mNGS workflow, using a composite clinical standard consisting of provider-ordered microbiology testing, clinical data, and orthogonal testing as the comparator (4). Findings from this study suggests that the performance of tNGS and mNGS is remarkably similar, which is consistent with our findings. Notably, the mNGS approach incurs a higher cost, representing a significant barrier to its widespread adoption in clinical settings. Given their comparable diagnostic efficacy, tNGS could potentially serve as a more cost-effective alternative to mNGS in clinical practice (7). Nonetheless, our research went beyond these observations, delving deeper into the distinctions between the three NGS strategies to uncover nuanced differences.

In comparing the distribution of pathogens among the three NGS methodologies, distinct differences were observed. Overall, mNGS demonstrated lower sensitivity compared to tNGS; however, tNGS exhibited a lower PPV. Specifically, mNGS showed poor performance in detecting SARS-CoV-2, HHV-1, H. influenzae, and A. fumigatus. In contrast, capture-based tNGS only performed poorly in detecting C. albicans. Notably, amplification-based tNGS exhibited poor performance for HHV-4, HHV-5, and HHV-1. Moreover, amplification-based tNGS faced significant challenges with pathogen omission, including E. faecium, H. parainfluenzae, C. striatum, S. mitis, and S. maltophilia, highlighting a substantial gap in its detection scope.

The primary pathogens not detected by the amplification-based tNGS methods in our study included Gram-positive bacteria, Gam-negtive bacteria, and DNA viruses, notably E. faecium, C. striatum, H. parainfluenzae, HHV-4, S. mitis, and S. maltophilia, among others. This limitation largely stemmed from the specific pathogens being outside the scope of the targeted panel employed by the amplification-based tNGS approach used, coupled with its challenge in identifying pathogens present in low abundance. The constraints inherent in amplification strategy significantly limit the number of primers that can be effectively used, with an initial cap of approximately 20000 primers to prevent the occurrence of primer dimerization. However, when additional factors such as pathogen specificity, the melting temperatures of the primers, and their secondary structures are taken into account, the practical number of usable primers is substantially reduced. Consequently, this leads to a bias in pathogen identification by the amplification-based tNGS method, hindering the detection of low-abundance pathogens and those beyond the targeted panel. Nevertheless, amplification-based tNGS can perform exceptionally well when aligned with the specific requirements of researchers. For example, in a study of COVID-19 detection, although amplification-based and capture-based tNGS methods generally achieve comparable enrichment effects for target gene fragments, amplification-based approaches exhibit superior enrichment for viruses at lower concentrations (below 10⁴ copies/mL or Ct > 28.7) (14). Thus, it is crucial to carefully select the appropriate amplification-based tNGS method tailored to the specific needs of clinical practice.

The observed low PPV in capture-based tNGS can be attributed to the reports of pathogens not corroborated by the comprehensive clinical diagnosis. Predominantly, these pathogens included C. striatum, HHV-5, HHV-1, and HHV-4, which are mainly classified as bacteria and DNA viruses. Notably, a greatly enhancement in the PPV was achieved through the optimization of the RPM threshold for these pathogens, indicating a substantial untapped potential in tNGS technology. This suggests that tNGS possesses significant capabilities for pathogen detection in lower respiratory tract infections, although further refinement and validation are required to fully exploit its potential.

In comparison with mNGS, tNGS methods demonstrated superior nucleic acid enrichment capabilities, leading to their heightened sensitivity. This observation is supported by prior research indicating that tNGS can significantly enhance the concentration of microbial nucleic acids by orders of magnitude more than mNGS (15). Additionally, tNGS offers the advantage of identifying pathogen genotypes, a capability absents in mNGS approaches. Notably, during the COVID-19 pandemic, tNGS played a pivotal role in infection surveillance and genomic characterization (16, 17). Our findings further reveal that, among tNGS methodologies, capture-based tNGS outperforms amplification-based tNGS in gene typing accuracy, despite both being capable of discerning SARS-CoV-2 genotypes. Moreover, tNGS methods also demonstrated AMR genes and VFs identification capabilities in this study. This discovery highlights the critical clinical importance of targeted sequencing strategies in precisely detecting key genetic indicators of antimicrobial resistance and pathogenicity. Regrettably, our study lacked a clinical gold standard against which the accuracy of these tNGS methods could be evaluated. However, the findings revealed a degree of consistency between the two tNGS approaches, with the capture-based tNGS method exhibiting a broader detection range. This discrepancy may stem from variations in panel design, indicating that the selection of the most suitable method should be tailored to the specific goals of the clinical scenario at hand.

The limited efficacy of mNGS in detecting RNA viruses, particularly SARS-CoV-2, can mainly be attributed to the general absence of RNA mNGS testing in this study. This limitation is largely attributable to conventional mNGS platforms detect DNA and RNA separately, coupled with the higher costs associated with RNA mNGS assays, which leads to a clinical preference for DNA-only mNGS testing, marking a significant limitation in the clinical deployment of mNGS. Crucially, our study reveals that tNGS not only parallels mNGS in performance but also provides potential advantages under certain conditions. This suggests that tNGS, with its lower cost, could serve as a feasible pre-test to mNGS in the clinical detection of lower respiratory tract infections.

In conclusion, each of the three NGS methods exhibits unique characteristics and advantages, though no significant differences in diagnostic efficacy were observed among them. In theory, mNGS offers the most comprehensive coverage of pathogens. Nevertheless, both tNGS techniques also demonstrated great accuracy for the identification of pathogens in BALF. In particular, capture-based tNGS, with its heightened sensitivity, is more appropriate for clinical scenarios where conventional testing fails to identify the causative microorganism, or when empirical treatments prove ineffective. Additionally, for precise pathogen genotyping, AMR genes and VFs identification, capture-based tNGS stands out as the most effective methodology.

PCR, polymerase chain reaction

NGS, next-generation sequencing

mNGS, metagenomic next-generation sequencing

tNGS, targeted next-generation sequencing

CMT, conventional microbiological test

BALF, bronchoalveolar lavage fluid

NTC, non-template control

RPM, reads per million

dsDNA, double-stranded DNA

DNB, DNA nanoball

NPV, negative predictive value

PPV, positive predictive value

HHV-5, human betaherpes virus 5

HHV-4, human gammaherpes virus 4

HHV-1, human alphaherpes virus 1

ROC, receiver operating characteristic

AMR, antimicrobial resistance

VF, virulence factor

Ethics approval and consent to participate

The study design received approval from the Ethics Committee of the Second Xiangya Hospital of Central South University in accordance with the Declaration of Helsinki (LYF2022229). Given the retrospective nature of the research, the Ethics Committee determined that an informed consent was not required.

Consent for publication

Not applicable.

Availability of data and materials

All sequencing data were deposited in the Genome Warehouse in the National Genomics Data Center (National Genomics Data Center Members and Partners, 2022) under project PRJCA025979, which are publicly accessible at https://bigd.big.ac.cn/gsa.

Competing interests

The authors declare that they have no competing interests.

Funding

This study was supported by National Natural Science Foundation of China (No. 82102499; ZL), Hunan Natural Science Foundation (No. 2021JJ40840; ZL), and the Scientific Research Launch Project for new employees of the Second Xiangya Hospital of Central South University.

Authors' contributions

ZL and MH were instrumental in guiding the study and spearheading the drafting of the manuscript. LT and QL assessed the diagnostic utility of three NGS methodologies. ZG and XL analyzed the concordance among the three NGS methods at detecting specific pathogens. WW and ZC investigated the variances in nucleic acid detection capabilities across the NGS techniques. DG, HL, HG and YY focused on evaluating the effectiveness of NGS in pathogen genotyping and identifying resistance and virulence genes. MH and DG offered essential revisions during the manuscript review stage. All authors read and approved the final manuscript.

Acknowledgments

We owe thanks to the patients in our study and their family members. We acknowledge the staffs of all departments for their assistance to this study.

Langelier C, Kalantar KL, Moazed F, Wilson MR, Crawford ED, Deiss T, et al. Integrating host response and unbiased microbe detection for lower respiratory tract infection diagnosis in critically ill adults. Proc Natl Acad Sci U S A. 2018;115(52):E12353-e62.
Zheng Y, Qiu X, Wang T, Zhang J. The Diagnostic Value of Metagenomic Next-Generation Sequencing in Lower Respiratory Tract Infection. Front Cell Infect Microbiol. 2021;11:694756.
Zhu YG, Tang XD, Lu YT, Zhang J, Qu JM. Contemporary Situation of Community-acquired Pneumonia in China: A Systematic Review. J Transl Int Med. 2018;6(1):26-31.
Gaston DC, Miller HB, Fissel JA, Jacobs E, Gough E, Wu J, et al. Evaluation of Metagenomic and Targeted Next-Generation Sequencing Workflows for Detection of Respiratory Pathogens from Bronchoalveolar Lavage Fluid Specimens. J Clin Microbiol. 2022;60(7):e0052622.
Simner PJ, Miller S, Carroll KC. Understanding the Promises and Hurdles of Metagenomic Next-Generation Sequencing as a Diagnostic Tool for Infectious Diseases. Clin Infect Dis. 2018;66(5):778-88.
Filkins LM, Bryson AL, Miller SA, Mitchell SL. Navigating Clinical Utilization of Direct-from-Specimen Metagenomic Pathogen Detection: Clinical Applications, Limitations, and Testing Recommendations. Clin Chem. 2020;66(11):1381-95.
Miao Q, Ma Y, Wang Q, Pan J, Zhang Y, Jin W, et al. Microbiological Diagnostic Performance of Metagenomic Next-generation Sequencing When Applied to Clinical Practice. Clin Infect Dis. 2018;67(suppl_2):S231-s40.
Chen X, Ding S, Lei C, Qin J, Guo T, Yang D, et al. Blood and Bronchoalveolar Lavage Fluid Metagenomic Next-Generation Sequencing in Pneumonia. Can J Infect Dis Med Microbiol. 2020;2020:6839103.
Chen Y, Feng W, Ye K, Guo L, Xia H, Guan Y, et al. Application of Metagenomic Next-Generation Sequencing in the Diagnosis of Pulmonary Infectious Pathogens From Bronchoalveolar Lavage Samples. Front Cell Infect Microbiol. 2021;11:541092.
Duan H, Li X, Mei A, Li P, Liu Y, Li X, et al. The diagnostic value of metagenomic next⁃generation sequencing in infectious diseases. BMC Infect Dis. 2021;21(1):62.
Li S, Tong J, Liu Y, Shen W, Hu P. Targeted next generation sequencing is comparable with metagenomic next generation sequencing in adults with pneumonia for pathogenic microorganism detection. J Infect. 2022;85(5):e127-e9.
Lin R, Xing Z, Liu X, Chai Q, Xin Z, Huang M, et al. Performance of targeted next-generation sequencing in the detection of respiratory pathogens and antimicrobial resistance genes for children. J Med Microbiol. 2023;72(11).
Blauwkamp TA, Thair S, Rosen MJ, Blair L, Lindner MS, Vilfan ID, et al. Analytical and clinical validation of a microbial cell-free DNA sequencing test for infectious disease. Nat Microbiol. 2019;4(4):663-74.
Xiao M, Liu X, Ji J, Li M, Li J, Yang L, et al. Multiple approaches for massively parallel sequencing of SARS-CoV-2 genomes directly from clinical samples. Genome Med. 2020;12(1):57.
Singh RR. Target Enrichment Approaches for Next-Generation Sequencing Applications in Oncology. Diagnostics (Basel). 2022;12(7).
Cheng LL, Li SY, Zhong NS. [New characteristics of COVID-19 caused by the Omicron variant in Guangzhou]. Zhonghua Jie He He Hu Xi Za Zhi. 2023;46(5):441-3.
Ramos N, Panzera Y, Frabasile S, Tomás G, Calleros L, Marandino A, et al. A multiplex-NGS approach to identifying respiratory RNA viruses during the COVID-19 pandemic. Arch Virol. 2023;168(3):87.

No competing interests reported.

SupplementMaterials.pdf

Download PDF

Editor assigned by journal
02 Aug, 2024
Submission checks completed at journal
02 Aug, 2024
First submitted to journal
31 Jul, 2024

You are reading this latest preprint version

Comparative performance of three next-generation sequencing techniques in real clinical lower respiratory tract infections

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Background

Methods

Statistical analysis

Results

Discussion

Conclusion

Abbreviations

Declarations

Ethics approval and consent to participate

Consent for publication

Availability of data and materials

Competing interests

Funding

Authors' contributions

Acknowledgments

References

Additional Declarations

Supplementary Files

Status:

Version 1