teamNGS Balances Sensitivity for Viruses with Comprehensive Microbial Detection in Clinical Specimens

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Title

“teamNGS balances sensitivity for viruses with comprehensive microbial detection”

The title is somewhat vague: it does not clearly indicate whether teamNGS is a workflow, a method, a platform, or a combined sequencing strategy

Abstract

“teNGS achieved 100–10,000X increases in depth and >50% genome coverage for pathogens with titers ≥1000 cp/ml using only 3–4% of the number of reads usually required for mNGS.”

The sentence is overly long and contains too much information, making it difficult to follow. Consider splitting it into two shorter sentences and briefly clarifying the improvement or added value of this approach compared with standard mNGS.

 

“Coupling methods maintains the sensitivity and coverage for viruses achieved by enrichment alone while also ensuring comprehensive recovery of non-viral microbes.”

There is a grammatical error in this sentence: “Coupling methods maintains …” should be corrected to “maintain.”

1. Introduction

No sentence clearly states that “existing methods cannot simultaneously achieve high sensitivity for viruses and comprehensive non-viral detection.” The introduction should better highlight the novelty and rationale of the current study.

 

The manuscript extensively lists previous work (e.g., Galileo One, Karius, VirCapSeq) but lacks a comparative summary. The introduction should synthesize these studies and clearly emphasize the advantages and novelty of the current work.

 

The introduction is disproportionately long compared to standard IMRaD structure, containing extensive descriptions of commercial products (e.g., Galileo, Karius, Illumina), which detracts from highlighting the research question. Merge repetitive information, such as multiple mentions of mNGS host background and probe enrichment, and limit commercial product descriptions to only those directly relevant to the study.

 

2. Materials and Methods

L127: “Patient specimens were pre-treated with Benzonase” – the final concentration, temperature, and incubation time are not provided, making replication across batches impossible.

 

L185: “using an in-house metagenomics and virus discovery pipeline (DiVir 3.0)” – no GitHub/DOI/commit number is provided, limiting reproducibility.

 

L300: “EMCV was spiked into the lysis buffer” – the amount/dose and whether fresh preparations were made for each tube are not specified.

 

 

Software versions missing: all software mentioned (CLC, R, MAFFT) lack version numbers, reducing reproducibility.

3. Results

L284: “as more libraries were multiplexed, the on-target percentage steadily increased” – slope or correlation (r) values are not reported; readers cannot judge whether this is statistically significant or merely a trend.

 

4. Discussion

L607: The statement “combining them and illustrating the benefits of this approach directly on clinical specimens has not been published.” lacks literature comparison and should be substantiated or rephrased.

 

Clinical and public health: The manuscript mentions potential clinical application but could be more specific. How can the rich data generated by teamNGS be converted into actionable reports for clinicians? Should the reports distinguish between confirmed infection, incidental colonization, or environmental background organisms? Additionally, how could this approach integrate into existing public health surveillance systems to enable earlier outbreak detection?

Author Response

Reviewer 1: Comments and Suggestions for Authors

Title

“teamNGS balances sensitivity for viruses with comprehensive microbial detection”

The title is somewhat vague: it does not clearly indicate whether teamNGS is a workflow, a method, a platform, or a combined sequencing strategy.

We believe that teamNGS is explained adequately in the abstract, and that the NGS in ‘teamNGS’ implies sequencing, while mention of sensitivity and detection also implies a diagnostic method. As Reviewer #2 also had a comment about the title, suggesting we emphasize ‘scalability or clinical applicability’, and noted that sequencing >2000 clinical specimens ‘strengthens the translational significance’, we have modified the title as follows:

“teamNGS balances sensitivity for viruses with comprehensive microbial detection in clinical specimens”

Abstract

“teNGS achieved 100–10,000X increases in depth and >50% genome coverage for pathogens with titers ≥1000 cp/ml using only 3–4% of the number of reads usually required for mNGS.”

The sentence is overly long and contains too much information, making it difficult to follow. Consider splitting it into two shorter sentences and briefly clarifying the improvement or added value of this approach compared with standard mNGS.

Due to the 200-word limit for the abstract, we have switched the clauses to make the sentence easier to follow. Mention of ‘standard’ mNGS and ‘increased sensitivity’ now clarifies the added value of the teNGS approach.

“Using only 3–4% of reads required for standard mNGS, teNGS achieved increased sensitivity, 100–10,000X increases in depth, and obtained >50% genome coverage for pathogens with titers ≥1000 cp/ml.”

“Coupling methods maintains the sensitivity and coverage for viruses achieved by enrichment alone while also ensuring comprehensive recovery of non-viral microbes.”

There is a grammatical error in this sentence: “Coupling methods maintains …” should be corrected to “maintain.”

We believe this sentence is grammatically correct as written.  ‘Methods’, while plural, is the direct object and not the subject of the sentence, which is actually ‘coupling’.  Remove ‘methods’ and the verb is singular: e.g. coupling maintains the sensitivity…

1. Introduction

No sentence clearly states that “existing methods cannot simultaneously achieve high sensitivity for viruses and comprehensive non-viral detection.” The introduction should better highlight the novelty and rationale of the current study.

Thank you for this suggestion. In the final paragraph of the Introduction, we have paraphrased your above statement to state the rationale for why teamNGS is needed and to highlight its novelty.

The manuscript extensively lists previous work (e.g., Galileo One, Karius, VirCapSeq) but lacks a comparative summary. The introduction should synthesize these studies and clearly emphasize the advantages and novelty of the current work.

Considering the length of the Introduction, as mentioned below, expanding it further to provide a comparative summary is not preferable. Rather, we have trimmed text in several areas describing the commercial offerings and then synthesize the thought at the end to state that these offerings and published methods do not combine the sensitivity of teNGS with the comprehensiveness of mNGS.

The introduction is disproportionately long compared to standard IMRaD structure, containing extensive descriptions of commercial products (e.g., Galileo, Karius, Illumina), which detracts from highlighting the research question. Merge repetitive information, such as multiple mentions of mNGS host background and probe enrichment, and limit commercial product descriptions to only those directly relevant to the study.

As mentioned above, we cut down the length for the commercial product descriptions and limited it to those directly relevant.

1. Materials and Methods

L127: “Patient specimens were pre-treated with Benzonase” – the final concentration, temperature, and incubation time are not provided, making replication across batches impossible.

This information has now been added: 200 U/ml final, 37^oC for 3 hrs. 

L185: “using an in-house metagenomics and virus discovery pipeline (DiVir 3.0)” – no GitHub/DOI/commit number is provided, limiting reproducibility.

DiVir 3.0 is a proprietary pipeline.  The essential, publicly available software used therein is described in broad terms in our recent publication (Weiss SL et al 2025 Frontiers in Microbiology) which we now cite.

L300: “EMCV was spiked into the lysis buffer” – the amount/dose and whether fresh preparations were made for each tube are not specified.

The titer (1 TCID₅₀) was mentioned both in the legend of Figure 2D and in the text on line 300.  Large stocks were grown in culture and then aliquoted and frozen back for individual use. Note this final titer is determined by both the initial volume of the lysis buffer placed on the instrument and the specific volume added to each sample.

Software versions missing: all software mentioned (CLC, R, MAFFT) lack version numbers, reducing reproducibility.

The following software versions have been added to the M & M section.

R 4.3.1 programming language

tidyverse v.2.0

rnaturalearth v.1.0

MAFFT v.7.487

Biostrings v.2.6.81

CLC Genomics Workbench v.23

 

1. Results

L284: “as more libraries were multiplexed, the on-target percentage steadily increased” – slope or correlation (r) values are not reported; readers cannot judge whether this is statistically significant or merely a trend.

Thank you for suggesting this analysis: we performed a Pearson’s product-moment correlation test and have briefly mentioned the results in the text describing Figure 2 and the procedure in the Materials and Methods.

Percent on-target values were standardized in two steps: first within each Virus (scaling to the within‑virus maximum), and then within each Virus × Level subgroup. “High” and “Low” levels corresponded to input concentrations of 5000 and 1000 copies/mL, respectively.

In high‑concentration samples (5000 copies/mL), standardized on‑target percentage showed a moderate positive correlation with increased multiplexing (r = 0.51, 95% CI: 0.02–0.80, p = 0.044). However, when excluding one outlier (HIV, 24-plex pool), the correlation was much stronger (r = 0.82, 95% CI: 0.52–0.94, p = 0.0002).

Low‑concentration samples (1000 copies/mL) also demonstrated a strong positive correlation (r = 0.77, 95% CI: 0.45–0.92, p < 0.001). Even when combining across all samples (high & low), the association remained significant (r = 0.60, 95% CI: 0.32–0.79, p < 0.001).

Overall, these results indicate that higher plexity values are associated with higher relative on‑target performance, with the relationship being most pronounced at lower viral input concentrations.

 

1. Discussion

L607: The statement “combining them and illustrating the benefits of this approach directly on clinical specimens has not been published.” lacks literature comparison and should be substantiated or rephrased.

As Reviewer #2 points out, “The claim that this combination has not been previously described or benchmarked with real clinical specimens appears valid within the scope of current literature.”  It is impossible to prove a negative, however, we performed a formal query as follows:

After exploring PubMed category and term variants in MS Copilot and Google Gemini genAI tools, we performed broad independent PubMed search for metagenomic and target enrichment NGS strategies for identifying microorganisms in clinical specimens. Combining these search outputs, we retrieved 189 potentially relevant records where the results overlapped. We manually reviewed these and found no evidence of a published report describing the combined usage of these two technologies (mNGS and teNGS) as we have done here with teamNGS.

(((("Metagenomics"[MeSH Terms] OR "High-Throughput Nucleotide Sequencing"[MeSH Terms] OR "metagenomi*"[Title/Abstract] OR "generation sequencing"[Title/Abstract] OR "shotgun"[Title/Abstract])

AND ("hexame*"[Title/Abstract] OR "random"[Title/Abstract] OR "unbiased"[Title/Abstract] OR "agnostic"[Title/Abstract])

AND ("clinical"[Title/Abstract] OR "patien*"[Title/Abstract] OR "sampl*"[Title/Abstract])

AND ("Microbiota"[MeSH Terms] OR "microb*"[Title/Abstract] OR "pathoge*"[Title/Abstract] OR "infect*"[Title/Abstract])))

AND

((("enrich*"[Title/Abstract]

OR "capture"[Title/Abstract] OR "hybridiz*"[Title/Abstract] OR "probe"[Title/Abstract] OR "Twist"

OR "Comprehensive Viral Research Panel"[Title/Abstract] OR "VirCapSe*"[All Fields]

OR "Virus Capture Vertebrate"[All Fields] OR "Vir-Cap-Vert"[All Fields] OR "virome capture sequencing"[All Fields] OR "targeted virome sequencing"[All Fields])

AND ("clinical diagnostics"[All Fields] OR "microb*"[Title/Abstract] OR "pathogen"[Title/Abstract] OR "infect*"[Title/Abstract]))))

 

Clinical and public health: The manuscript mentions potential clinical application but could be more specific. How can the rich data generated by teamNGS be converted into actionable reports for clinicians? Should the reports distinguish between confirmed infection, incidental colonization, or environmental background organisms? Additionally, how could this approach integrate into existing public health surveillance systems to enable earlier outbreak detection?

Thank you for these helpful comments. Indeed, in the last paragraph of the Discussion we addressed the potential clinical applications and mentioned the many hurdles that would need to be overcome to realize this vision. Throughout the manuscript we have been careful to stress this is for research and that additional work is required to fully validate the method. Regardless, efforts are underway to integrate teamNGS data with patient metadata to generate actionable reports for clinicians. And modifications to the DiVir pipeline are being implemented that will better distinguish between confirmed infection, incidental colonization, or environmental background organisms.

In this same final paragraph, we also stressed the importance of epidemiologic based ‘case finding’ to spot the interesting samples and the role of networks for shared resources and increased communication. 

On line 747 (marked-up version), we previously stated, “Finding these individuals will give the scientific community a head-start on developing diagnostics and learning how to treat or prevent spread.” To substantiate this statement with examples of how these NGS approaches have been incorporated into the existing public health system to enable early detection, we now mention our work using in Uganda standing up mortuary surveillance and at APDC fever clinics in Colombia.

We have added the following sentence:

Indeed, integrating case finding with sequencing into the existing public health systems (Rodgers MA, 2025 Int J Inf Dis) has enabled early detection, with mortuary surveillance in Uganda spotting outbreaks of Echovirus 7 (Weiss SL 2025 Front Micro) and anthrax (Bbosa N 2024 ASTMH), and outpatient fever clinics in Colombia detecting Oropouche (Ciuoderis K, 2022 Emerg Micro Infect) and Mayaro (Perez-Restrepo L 2024 Front Micro) viruses.   

Reviewer 2 Report

Comments and Suggestions for Authors

1. General Comments

The manuscript presents an applied and technically valuable contribution describing the development, optimization, and evaluation of a combined metagenomic and target-enrichment sequencing workflow termed teamNGS. The work is ambitious in scale, using thousands of clinical samples sourced globally, and integrates both methodological validation and real-world diagnostic implications. The manuscript is generally well-written, methodologically rich, and has relevance for infectious disease research, public health surveillance, and translational diagnostics. The narrative is clear, although somewhat lengthy and dense in its methodological parts, and in several sections would benefit from clearer signposting to reduce the perceived complexity of the workflow. The inclusion of multiple real-world cohorts, plus the comparative evaluation against mNGS alone, strengthens the scientific value of the paper.

2. Originality and Relevance of the Topic

The topic is highly relevant. The current global diagnostic landscape demands affordable and scalable technologies capable of detecting both known and unexpected viral pathogens. The novelty lies not only in the demonstration of CVRP-based target enrichment performance at scale, but more importantly, in the introduction of teamNGS as a practical operational solution that merges sensitivity with broad microbial profiling while reducing workflow and cost. The claim that this combination has not been previously described or benchmarked with real clinical specimens appears valid within the scope of current literature. The application across >2000 clinical samples adds substantial originality and strengthens translational significance. 

3. Methodological Rigor and Suggested Improvements

The methodological framework is thorough and well-controlled. Use of internal controls, various titers, multiple pooling strategies, and cross-platform sequencing strengthens reliability. The comparative analysis between mNGS, teNGS, and teamNGS is carefully interpreted, with adequate use of quantitative metrics such as RPM, genome coverage, and fold-changes. Control of confounding factors such as index hopping and species detection bias is acknowledged.

Suggested methodological improvements:

1. Clarify inclusion/exclusion criteria for clinical samples.This is particularly relevant because samples came from heterogeneous cohorts with variable pre-test knowledge and may affect interpretation.
2. Add more explicit definitions for “complete genome,” “partial genome,” and sensitivity thresholds.Although RPM and % genome coverage are provided, precise operational classification criteria would enhance clarity.
3. Consider including statistical teststo support statements regarding performance differences, especially in Figures comparing methods across multiple samples.
4. Expand on bioinformatics reproducibility and benchmarking.While DiVir 3.0 is briefly mentioned, the reproducibility and accessibility of the pipeline (open availability, parameters, version control) could be clarified.
5. Discuss potential biases introduced by mNGS library preparation kits, especially considering comparisons between Nextera XT and sparQ.

4. Consistency of Conclusions with Evidence Presented

The conclusions align with the experimental evidence presented. The authors appropriately avoid over-claiming regarding the use of teamNGS as a stand-alone diagnostic tool and acknowledge limitations, particularly regarding discovery of highly divergent viruses. The economic and workflow advantages are convincingly supported by the data and cost breakdown table. The scope of clinical applicability and relevance to real-world diagnostic pipelines is reasonable and well contextualized.

5. Specific Comments By Line or Section

Title and Abstract

• The title accurately reflects the main contribution but consider emphasizing scalability or clinical applicability.
• In the abstract, consider specifying that >2000 samples were evaluated, as this enhances the perceived strength of evidence.

Introduction (Lines ~43–116)

• The introduction is comprehensive, but could be slightly condensed; currently it reads more like a narrative literature review than a problem-statement-driven section.
• An explicit final paragraph outliningresearch questions or hypotheses would be beneficial. 

Materials and Methods (Lines ~117–209)

• Line 121: What do the authors mean by numerous samples were purchased from commercial vendors? Authors should specify ethics committee approval number.
• The internal control description is thorough, though a diagram or supplementary table specifying concentrations and passage number for all controls would help reproducibility.
• Consider specifying how frequently batches were re-validated and whether batch effects were tested statistically.

Results

• In Figure 3 and related text, the explanation of sample subgroups (circles, squares, diamonds, triangles) is useful but could benefit from a compact summary table.
• In the virus discovery section, it would be valuable to clarify thepractical implications of the 65% identity threshold and whether this can inform probe design algorithms prospectively. 

Discussion

• The economic and time-efficiency claims are compelling, but including a comparison withcurrent clinical molecular diagnostic standards or reimbursement considerations would make the translational impact more concrete.
• Consider expanding a paragraph on ethical and biosurveillance considerations, given the implied global public health relevance.

Author Response

Reviewer 2: Comments and Suggestions for Authors

1. General Comments

The manuscript presents an applied and technically valuable contribution describing the development, optimization, and evaluation of a combined metagenomic and target-enrichment sequencing workflow termed teamNGS. The work is ambitious in scale, using thousands of clinical samples sourced globally, and integrates both methodological validation and real-world diagnostic implications. The manuscript is generally well-written, methodologically rich, and has relevance for infectious disease research, public health surveillance, and translational diagnostics. The narrative is clear, although somewhat lengthy and dense in its methodological parts, and in several sections would benefit from clearer signposting to reduce the perceived complexity of the workflow. The inclusion of multiple real-world cohorts, plus the comparative evaluation against mNGS alone, strengthens the scientific value of the paper.

Thank you for this positive evaluation.

2. Originality and Relevance of the Topic

The topic is highly relevant. The current global diagnostic landscape demands affordable and scalable technologies capable of detecting both known and unexpected viral pathogens. The novelty lies not only in the demonstration of CVRP-based target enrichment performance at scale, but more importantly, in the introduction of teamNGS as a practical operational solution that merges sensitivity with broad microbial profiling while reducing workflow and cost. The claim that this combination has not been previously described or benchmarked with real clinical specimens appears valid within the scope of current literature. The application across >2000 clinical samples adds substantial originality and strengthens translational significance. 

Again, thank you for summarizing well what teamNGS offers and affirming that the approach is indeed novel. Its application to numerous clinical samples certainly suggests it has promise beyond that of a research tool.

3. Methodological Rigor and Suggested Improvements

The methodological framework is thorough and well-controlled. Use of internal controls, various titers, multiple pooling strategies, and cross-platform sequencing strengthens reliability. The comparative analysis between mNGS, teNGS, and teamNGS is carefully interpreted, with adequate use of quantitative metrics such as RPM, genome coverage, and fold-changes. Control of confounding factors such as index hopping and species detection bias is acknowledged.

Suggested methodological improvements:

1. Clarify inclusion/exclusion criteria for clinical samples. This is particularly relevant because samples came from heterogeneous cohorts with variable pre-test knowledge and may affect interpretation.

In this paper, while the extent of information on each sample (e.g. medical history, other diagnostic results) certainly varied considerably, it did not form the basis for including or excluding it. Rather, the focus was on detection/no detection to compare methods (e.g. mNGS vs teNGS vs teamNGS).  In future work where we define sensitivity and specificity, having orthogonal results like PCR or serology will then dictate which samples to restrict analysis to.

1. Add more explicit definitions for “complete genome,” “partial genome,” and sensitivity thresholds. Although RPM and % genome coverage are provided, precise operational classification criteria would enhance clarity.

Per later comments from Reviewer 2, we have expanded the legend of Figure 3 in a table format to define ‘complete’ and ‘partial’ genomes.

1. Consider including statistical tests to support statements regarding performance differences, especially in Figures comparing methods across multiple samples.

As requested by Reviewer 1, we have performed statistical analysis to determine the significance of the trend resulting from increased multiplexing (Figure 2C).  In addition, we asserted that teamNGS yields greater genome coverage compared to mNGS or teNGS alone.  An Aligned Rank Transform test of the samples in Figures 6B & 6C was performed to compare across sequencing methods since the dependent variable, genome coverage, violated the assumptions required for parametric ANOVA.

Across all pathogens, teamNGS (10%, P2) consistently produced higher aligned‑rank genome coverage than both pure mNGS (p = 0.0004) and pure teNGS (p = 0.0150), whereas the individual methods did not differ significantly from each other (p = 0.2025). As expected, when grouped by only viral infections, teamNGS was superior to teNGS (p = 0.0201) and mNGS (p = 0.0001), while no significant differences in coverage were observed for non-viral pathogens.

Dataset	Comparison	Estimate	p‑value
All pathogens	Pure mNGS – Pure teNGS	–10.9	0.2025
All pathogens	Pure mNGS – teamNGS (10%, P2)	–30.2	0.0004
All pathogens	Pure teNGS – teamNGS (10%, P2)	–19.4	0.0150
Viral only	Pure mNGS – Pure teNGS	–13.5	0.0166
Viral only	Pure mNGS – teamNGS (10%, P2)	–26.6	0.0001
Viral only	Pure teNGS – teamNGS (10%, P2)	–13.1	0.0201
Non‑viral only	Pure mNGS – Pure teNGS	+7.0	0.1110
Non‑viral only	Pure mNGS – teamNGS (10%, P2)	+2.33	0.6225
Non‑viral only	Pure teNGS – teamNGS (10%, P2)	–4.67	0.2557

This demonstrates that the teamNGS protocol systematically yields higher genome coverage for viruses compared to the individual ‘pure’ methods, while still also detecting non-viral pathogens.

With both analyses reinforcing the conclusions stated, we have now added a brief statistical analysis section to the Materials and Methods. The results in the table above for Figure 6 are now included in the Supplement as Table S6.

1. Expand on bioinformatics reproducibility and benchmarking. While DiVir 3.0 is briefly mentioned, the reproducibility and accessibility of the pipeline (open availability, parameters, version control) could be clarified.

Reviewer #1 had a similar comment. DiVir is proprietary and only available to Abbott personnel and our APDC partners.  In the Methods section we now cite Weiss SL, et al 2025 where we briefly described the essential elements of the pipeline.

1. Discuss potential biases introduced by mNGS library preparation kits, especially considering comparisons between Nextera XT and sparQ.

Aside from QuantaBio sparQ libraries producing comparatively larger average library sizes, we did not observe meaningful differences or biases to report on. Our intent was to show that either input library was amenable to downstream CVRP enrichment, not determine whether one was better than the other.

4. Consistency of Conclusions with Evidence Presented

The conclusions align with the experimental evidence presented. The authors appropriately avoid over-claiming regarding the use of teamNGS as a stand-alone diagnostic tool and acknowledge limitations, particularly regarding discovery of highly divergent viruses. The economic and workflow advantages are convincingly supported by the data and cost breakdown table. The scope of clinical applicability and relevance to real-world diagnostic pipelines is reasonable and well contextualized.

5. Specific Comments By Line or Section

Title and Abstract

• The title accurately reflects the main contribution but consider emphasizing scalability or clinical applicability.

In light of Reviewer 1 comments and this suggestion here, we have changed the title to emphasize its clinical applicability.

• In the abstract, consider specifying that >2000 samples were evaluated, as this enhances the perceived strength of evidence.

We actually did mention this in the original version, but we are glad to receive this comment.

Introduction (Lines ~43–116)

• The introduction is comprehensive, but could be slightly condensed; currently it reads more like a narrative literature review than a problem-statement-driven section.
• An explicit final paragraph outlining research questions or hypotheses would be beneficial. 

Thank you: Reviewer 1 had similar comments and so we have shortened the Introduction and revised the final paragraph to state the research question and highlight the novelty of the work we are about to describe.

Materials and Methods (Lines ~117–209)

• Line 121: What do the authors mean by numerous samples were purchased from commercial vendors? Authors should specify ethics committee approval number.

We have added this information for purchased samples in the “Institutional Review Board Statement” section at the end of the article.

• The internal control description is thorough, though a diagram or supplementary table specifying concentrations and passage number for all controls would help reproducibility.

In the Methods subsection: Virus and bacteria stocks and NGS positive control preparation, we indicate the strain and the source from which it was purchased (ATCC, Exact Diagnostics). Data pages hosted by these vendors contain the initial titer and passage number information.  While this could certainly change if inventory is depleted, the relevant piece here regarding reproducibility is that we dilute them each individually to log 4.0 or log 3.0 copies/ml depending on whether target enrichment is performed or not.

• Consider specifying how frequently batches were re-validated and whether batch effects were tested statistically.

Given the previous question, it is unclear whether the reviewer is referring to the positive control virus cocktail or to patient samples generally. Assuming the former, we indeed show the consistency of the two controls in Supplemental Figure S2 for 11 experiments over 3 years, in terms of RPM and % genome coverage. The boxplots indicate the median and the 25%/75% confidence intervals and there does not appear to be a noticeable deterioration over time.  Importantly, just as with the EMCV internal control, these stocks represent one batch, produced in large quantities and aliquoted and frozen for later use. Therefore, not needing to remake them, we have not revalidated or tested for batch effects.

Results

• In Figure 3 and related text, the explanation of sample subgroups (circles, squares, diamonds, triangles) is useful but could benefit from a compact summary table.

Thank you for this suggestion. We have made a compact summary table to replace the figure legend in Fig 3E which now defines complete and partial genomes and enumerates mean read numbers for mNGS and teNGS for each class.

• In the virus discovery section, it would be valuable to clarify the practical implications of the 65% identity threshold and whether this can inform probe design algorithms prospectively. 

Since this is a topic for discussion, we have inserted a sentence on Line 717 (marked-up version) and cite our previous work designing probes for HIV where we did just this (e.g. design probes based off a consensus).

“The practical implication of this 65% threshold for probe design is that using an alignment of related viruses to generate a consensus sequence is preferable to tiling individual genomes.”

Discussion

• The economic and time-efficiency claims are compelling, but including a comparison with current clinical molecular diagnostic standards or reimbursement considerations would make the translational impact more concrete.

As mentioned for Reviewer #1, the data provided in this paper is restricted to research and ‘proof of principle’ on clinical samples.  Future work will define teamNGS method parameters (e.g. sensitivity, specificity) comparing orthogonal testing results (e.g. PCR) to judge against current diagnostic standards.  At that stage, one can begin to discuss reimbursement and other considerations for moving this into the clinic.

• Consider expanding a paragraph on ethical and biosurveillance considerations, given the implied global public health relevance.

One of this reviewer’s opening comments was that, “The narrative is clear, although somewhat lengthy…”.  We tend to agree, and adding more to the Discussion on these topics goes beyond the scope of the paper considering the stage of development that teamNGS is currently at.

Reviewer 3 Report

Comments and Suggestions for Authors

This study introduces teamNGS, a new method that combines metagenomic next-generation sequencing (mNGS) and probe-based target enrichment (teNGS). The study provides convincing evidence of teamNGS's effectiveness by analyzing thousands of clinical specimens. Notably, teamNGS's ability to markedly improve viral detection sensitivity and comprehensively detect non-viral pathogens in a single sequencing run is a clear advantage over conventional approaches and substantially contributes to the fields of clinical infectious disease diagnostics and outbreak surveillance.

However, several limitations exist. For instance, systematic comparisons of sensitivity and specificity against established diagnostic methods, such as PCR and serological assays, are limited. The constraints of sequence-similarity thresholds (>65%) are evident when searching for previously unknown viruses. Issues specific to target enrichment, such as cross-hybridization and index hopping, remain areas for further consideration. Nevertheless, these limitations do not undermine the technique's value; rather, they indicate areas for future refinement.

Overall, this manuscript merits a very high evaluation in terms of novelty, technical soundness, practical application to large-scale clinical specimens, and public health relevance. The concept of teamNGS, which reduces costs and simplifies workflows, will likely impact the implementation of infectious disease diagnostics in the future.

For these reasons, I consider this manuscript a high-quality study worthy of publication and recommend acceptance. I encourage the authors to consider the following minor revisions:

1. In several figures, especially Figures 3–6, the color legends are small and difficult to interpret at a glance. Adding direct labels within the figures may improve readability.
2. The relationship between the 15,000 CVRP probe strains and the 3,100 reference sequences is unclear. Adding a brief clarification in the legend or main text would help readers understand this relationship.
3. The term "pancaking" (over-amplification bias) may seem informal to non-specialist readers. It is recommended that a brief explanation be provided when it first appears.
4. Figures 1 and 6 contain English labels such as "dry down" and "aliquot." While this is not problematic, adding a brief textual note for consistency would improve clarity.
5. In Figure 4, it would be helpful to indicate the lengths of the probe-covered regions (e.g., Tacheng tick virus 1, Burana virus) visually on the plots.
6. If serotype or genotype information is available for the detected viruses, such as the chikungunya or dengue viruses, including this information in the supplementary data would enhance the study's epidemiological value.

 

Author Response

This study introduces teamNGS, a new method that combines metagenomic next-generation sequencing (mNGS) and probe-based target enrichment (teNGS). The study provides convincing evidence of teamNGS's effectiveness by analyzing thousands of clinical specimens. Notably, teamNGS's ability to markedly improve viral detection sensitivity and comprehensively detect non-viral pathogens in a single sequencing run is a clear advantage over conventional approaches and substantially contributes to the fields of clinical infectious disease diagnostics and outbreak surveillance.

However, several limitations exist. For instance, systematic comparisons of sensitivity and specificity against established diagnostic methods, such as PCR and serological assays, are limited. The constraints of sequence-similarity thresholds (>65%) are evident when searching for previously unknown viruses. Issues specific to target enrichment, such as cross-hybridization and index hopping, remain areas for further consideration. Nevertheless, these limitations do not undermine the technique's value; rather, they indicate areas for future refinement.

Overall, this manuscript merits a very high evaluation in terms of novelty, technical soundness, practical application to large-scale clinical specimens, and public health relevance. The concept of teamNGS, which reduces costs and simplifies workflows, will likely impact the implementation of infectious disease diagnostics in the future.

Thank you for the endorsement of our work and the insights provided. Among the limitations listed, these are indeed areas of ongoing work and future refinement of the method.

For these reasons, I consider this manuscript a high-quality study worthy of publication and recommend acceptance. I encourage the authors to consider the following minor revisions:

1. In several figures, especially Figures 3–6, the color legends are small and difficult to interpret at a glance. Adding direct labels within the figures may improve readability.

Each of these figures is packed with information.  Where possible, we have increased font sizes for labels and for Figure 3E we have enhanced the legend in a table format to aid interpretation.

1. The relationship between the 15,000 CVRP probe strains and the 3,100 reference sequences is unclear. Adding a brief clarification in the legend or main text would help readers understand this relationship.

Thank you for pointing out this is unclear.  We have modified the sentence in the main text to indicate that there are probes for more than 15,000 strains representing >3100 viral species.

1. The term "pancaking" (over-amplification bias) may seem informal to non-specialist readers. It is recommended that a brief explanation be provided when it first appears.

We believe what is written, “…since teNGS read numbers could be inflated by overamplification and numerous reads mapping to identical regions (i.e., ‘pancaking’)”, adequately explains the issue.

1. Figures 1 and 6 contain English labels such as "dry down" and "aliquot." While this is not problematic, adding a brief textual note for consistency would improve clarity.

For Figure 1 we added, “…and dried down in a heated vacufuge…to the legend.  For Figure 6, we added clarification in the Methods subsection: Preparation of teNGS and teamNGS libraries, to indicate that mNGS equal volume aliquots are synonymous with equal volume portions.

1. In Figure 4, it would be helpful to indicate the lengths of the probe-covered regions (e.g., Tacheng tick virus 1, Burana virus) visually on the plots.

The location of the Burana virus probe is now indicated in the revised figure.  Tacheng tick virus probes tile the entire segment, as was previously indicated.

1. If serotype or genotype information is available for the detected viruses, such as the chikungunya or dengue viruses, including this information in the supplementary data would enhance the study's epidemiological value.

When viral sequences exceeded 90% coverage, we submitted the genomes to GenBank along with this information. Supplemental Table S7 indicates there are 17 Dengue sequences representing serotypes 1-4 and there are 3 Chikungunya sequences from Honduras, Thailand, and India, likely representing ECSA and IOL genotypes.