Impact of Preweaning Vaccination on Host Gene Expression Patterns Linked to Future Bovine Respiratory Disease Development in Beef Calves

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

General Comments to the Authors:

• Materials and Methods: This Section completely lacks sub-numbering, which again significantly hindered the review process by making it difficult to indicate corrections at specific locations.
• Section 3. Results and Discussion: I assume that the authors have submitted this manuscript as a full-length research article. However, the Results and Discussion have been presented as a single combined section, a format that is generally more appropriate for a short communication. If the manuscript is intended for publication as a full-length research article, the authors are requested to present the Results and Discussion as separate sections.
• The authors are requested to follow uniform language throughout the manuscript like “GLMMseq” or “glmmSeq”.
• The authors are requested to avoid the numerous typographical or grammatical errors throughout the manuscript like “analyis,” “obsevation,” “adminstered,” “vaccaination,” “comparsion,” “downregualted,” “intial,” etc. Hence, the manuscript requires thorough English language editing prior to publication. There are numerous grammatical errors, awkward constructions, and non-idiomatic usages throughout the manuscript.

Major Reviewer’s Comments to the Authors:

Title

• The title is appropriate.

Abstract

• The authors are requested to provide the full form of “TIME”, when this abbreviation is first introduced in the Abstract. Thereafter, only the abbreviated forms should be used consistently throughout the remainder of the Abstract.
• The authors mentioned as “identifying 5,364 differentially expressed genes (DEGs) for TIME, 84 DEGs for vaccination (VAX), and 129 for BRD status” in the abstract. But, in the Results portion, the authors mentioned as “Analysis of genes from glmmSeq resulted in 11,068, 358, and 9,241 for TIME, VAX, and BRD, respectively. Between glmmSeq and QLF testing, a total of 5,364, 84, and 129 DEGs were identified for TIME, VAX, and BRD.” Further, the authors evaluated differential expression in two complementary ways, a generalized linear mixed model (GLMMseq); and pairwise comparisons for DEGs between each vaccination group at each timepoint were performed using edgeR and QLF. Hence, the authors are requested to include following things:
  • Include the evaluation of differential expression by two complementary ways as methods in the abstract.
  • Include both the results of “glmmSeq resulted in 11,068, 358, and 9,241 for TIME, VAX, and BRD, respectively. Between glmmSeq and QLF testing, a total of 5,364, 84, and 129 DEGs were identified for TIME, VAX, and BRD” in the abstract.
• The authors failed to mention the key limitation in the abstract that BRD × VAX interaction was with minimal DEGs (vaccination status did not modify BRD-associated gene expression), which is one of the more clinically relevant findings of this manuscript.

Keywords

• The authors are requested to use minimum of 6 Keywords.

Introduction

• The introduction is adequate.

Materials and Methods

• This Section completely lacks sub-numbering, which again significantly hindered the review process by making it difficult to indicate corrections at specific locations.
• The authors are requested to write the sub-heading for the first paragraph.
• The second paragraph is excessively long. The authors are requested to divide it into smaller sections using appropriate subheadings to improve readability and clarity.
• The authors are requested to avoid the sample number inconsistencies in the manuscript. The authors mentioned as 81 bull calves were enrolled (VAX n=39, NOVAX n=42). But, in the “Next-generation RNA sequencing and bioinformatic data processing,” authors mentioned as “A total of 285 samples were sequenced and included in the analysis representing 73 animals across 4 timepoints (NOVAX; n=40; VAX; n=32)”. Further, the authors are requested to note that 40 + 32 = 72, not 73. This calculation failure also must be corrected. The authors are requested to mention about what happened to remaining animals (81 – 72/73 = ~8–9), whether these animals are died? or removal from experiment?
• The authors are failed to mention about the data on number of animals classified as bovine respiratory disease (BRD) in either Materials and Methods or Results portions. This is important for statistical analysis.
• The authors are requested to validate the RNA-seq-derived candidate biomarkers by qPCR, which would strengthen the data accuracy.

Results

• The authors are requested to present the Results and Discussion as separate sections.
• Section 3. Results and Discussion: I assume that the authors have submitted this manuscript as a full-length research article. However, the Results and Discussion have been presented as a single combined section, a format that is generally more appropriate for a short communication. If the manuscript is intended for publication as a full-length research article, the authors are requested to present the Results and Discussion as separate sections.
• The authors are requested to correlate the antibody titer data with the transcriptomic findings, which would help in understand whether transcriptomic vaccine responses were correspond to humoral immunity?
• Figures 2 & 4 have significant readability issues. The labels are not clearly visible and overlap with each other. The figures require significant revision.

Discussion

• The authors are requested to discuss the rationale for the specific four sampling timepoints (T1-4; median age 107, 114, 183, and 283 respectively).
• The authors are requested to discuss about in the T3 group, revaccination and surgical castration (median age=183 days) happened simultaneously, whether castration-associated inflammatory signaling could influenced the vaccine-associated signals?
• Whole blood transcriptomics data and neutrophil degranulation pathways are variable according to the leukocyte subpopulation shift, which happens during stress, age, and vaccination. Hence, the authors are requested to discuss/justify how this variation was overcome?
• The authors are requested to discuss about in the T1 group, calves were vaccinated at a median age of 107 days, whether maternal antibodies interference the vaccine-induced transcriptional response?
• The authors mentioned as “absence of antiviral signalling” (absence of interferon signalling), but which not a biologically appropriate terminology. The authors are requested to use the proper language like “was not detected” rather than “absence of”.

Conclusion

• The authors are requested to mention the key limitation in the conclusion that BRD × VAX interaction was with minimal DEGs (vaccination status did not modify BRD-associated gene expression), which is one of the more clinically relevant findings of this manuscript.
• The authors are requested to write the targeted and actionable recommendations for future research, which would strengthen this manuscript.

References

The authors are requested to check and format once again all the references according to journal format especially while abbreviating the Journal names. In some References, volume, issue and page numbers are missing.

Comments on the Quality of English Language

The authors are requested to avoid the numerous typographical or grammatical errors throughout the manuscript like “analyis,” “obsevation,” “adminstered,” “vaccaination,” “comparsion,” “downregualted,” “intial,” etc. Hence, the manuscript requires thorough English language editing prior to publication. There are numerous grammatical errors, awkward constructions, and non-idiomatic usages throughout the manuscript.

Author Response

Reviewer 1 Open Review

( ) I would not like to sign my review report
(x) I would like to sign my review report

Quality of English Language

(x) The English could be improved to more clearly express the research.
( ) The English is fine and does not require any improvement.

 

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	( )	(x)	( )	( )
Are the results clearly presented?	( )	(x)	( )	( )
Are the conclusions supported by the results?	( )	(x)	( )	( )
Are all figures and tables clear and well-presented?	( )	(x)	( )	( )

Comments and Suggestions for Authors

General Comments to the Authors:

• Materials and Methods: This Section completely lacks sub-numbering, which again significantly hindered the review process by making it difficult to indicate corrections at specific locations.

Thank you. The material methods now have sub-numbering sections 2.1 – 2.7.

• Section 3. Results and Discussion: I assume that the authors have submitted this manuscript as a full-length research article. However, the Results and Discussion have been presented as a single combined section, a format that is generally more appropriate for a short communication. If the manuscript is intended for publication as a full-length research article, the authors are requested to present the Results and Discussion as separate sections.

 

• The authors are requested to follow uniform language throughout the manuscript like “GLMMseq” or “glmmSeq”.

Thank you. The package name glmmSeq is now used consistently throughout the manuscript including sections 2.5, 3.1, 3.2, and 3.3.

• The authors are requested to avoid the numerous typographical or grammatical errors throughout the manuscript like “analyis,” “obsevation,” “adminstered,” “vaccaination,” “comparsion,” “downregualted,” “intial,” etc. Hence, the manuscript requires thorough English language editing prior to publication. There are numerous grammatical errors, awkward constructions, and non-idiomatic usages throughout the manuscript.

All of the errors listed have been corrected and the manuscript has been reviewed for spelling, grammar, and awkward construction.

Major Reviewer’s Comments to the Authors:

Title

• The title is appropriate.

Abstract

• The authors are requested to provide the full form of “TIME”, when this abbreviation is first introduced in the Abstract. Thereafter, only the abbreviated forms should be used consistently throughout the remainder of the Abstract.

Thank you. TIME is now defined at first use in the Abstract.

• The authors mentioned as “identifying 5,364 differentially expressed genes (DEGs) for TIME, 84 DEGs for vaccination (VAX), and 129 for BRD status” in the abstract. But, in the Results portion, the authors mentioned as “Analysis of genes from glmmSeq resulted in 11,068, 358, and 9,241 for TIME, VAX, and BRD, respectively. Between glmmSeq and QLF testing, a total of 5,364, 84, and 129 DEGs were identified for TIME, VAX, and BRD.” Further, the authors evaluated differential expression in two complementary ways, a generalized linear mixed model (GLMMseq); and pairwise comparisons for DEGs between each vaccination group at each timepoint were performed using edgeR and QLF. Hence, the authors are requested to include following things:
• Include the evaluation of differential expression by two complementary ways as methods in the abstract.
• Include both the results of “glmmSeq resulted in 11,068, 358, and 9,241 for TIME, VAX, and BRD, respectively. Between glmmSeq and QLF testing, a total of 5,364, 84, and 129 DEGs were identified for TIME, VAX, and BRD” in the abstract.

The abstract now clarifies that differential expression was evaluated in two ways and reports both sets of results. Thank you.

• The authors failed to mention the key limitation in the abstract that BRD × VAX interaction was with minimal DEGs (vaccination status did not modify BRD-associated gene expression), which is one of the more clinically relevant findings of this manuscript.

Thank you. Additional information and discussion have been added in Discussion Section 4.5. Due to word limit concerns in the abstract, we identify this key limitation and expand upon it in the discussion section.

Keywords

• The authors are requested to use minimum of 6 Keywords.

Thank you for bringing this to our attention and additional keywords have been added. Additional Keywords added were RNA sequencing, gene expression, and immune development.

Introduction

• The introduction is adequate.

Materials and Methods

• This Section completely lacks sub-numbering, which again significantly hindered the review process by making it difficult to indicate corrections at specific locations.

Sub numbering has been added throughout Section 2.

• The authors are requested to write the sub-heading for the first paragraph.

This suggestion has been incorporated. Section 2.1 is now titled Animal Use and Previous Work.

• The second paragraph is excessively long. The authors are requested to divide it into smaller sections using appropriate subheadings to improve readability and clarity.

The second paragraph is now Section 2.3 Marketing Strategy, Sampling Timepoints, and BRD Case Definition and split into two paragraphs. Thank you for the suggestion.

• The authors are requested to avoid the sample number inconsistencies in the manuscript. The authors mentioned as 81 bull calves were enrolled (VAX n=39, NOVAX n=42). But, in the “Next-generation RNA sequencing and bioinformatic data processing,” authors mentioned as “A total of 285 samples were sequenced and included in the analysis representing 73 animals across 4 timepoints (NOVAX; n=40; VAX; n=32)”. Further, the authors are requested to note that 40 + 32 = 72, not 73. This calculation failure also must be corrected. The authors are requested to mention about what happened to remaining animals (81 – 72/73 = ~8–9), whether these animals are died? or removal from experiment?

A correction has been made in the first paragraph of Section 2.4 Next-generation Sequencing and Bioinformatic Data Processing. It is now explicitly stated how samples were chosen to get from the original enrolled 81 calves to the 73 animals that were sequenced and a total of 292 samples.

• The authors are failed to mention about the data on number of animals classified as bovine respiratory disease (BRD) in either Materials and Methods or Results portions. This is important for statistical analysis.

This suggestion has been incorporated in Section 2.3 Marketing Strategy, Sampling Timepoints, and BRD Case Definition where BRD is defined.

• The authors are requested to validate the RNA-seq-derived candidate biomarkers by qPCR, which would strengthen the data accuracy.

We respectfully address the reviewer's request for RT-qPCR validation of our bulk RNA-Seq findings. While we appreciate the suggestion and acknowledge RT-qPCR's historical role in gene expression confirmation, we contend that RT-qPCR validation would not meaningfully strengthen the conclusions drawn from our RNA-Seq analysis and may in fact introduce methodological artifacts that could obscure biological signals. Our RNA-Seq data were generated with strict biological replicates per condition to meet estimated statistical power and processed using a standardized bioinformatics pipeline employing a transcriptome-wide normalization strategy that is statistically robust to compositional differences between samples. In contrast, RT-qPCR normalization depends on the stability of one or several endogenous reference genes across experimental conditions. Given the nature of our experimental system, it has been established in previous literature that commonly used housekeeping genes such as GAPDH and ACTB are themselves differentially expressed under such conditions, meaning RT-qPCR results could systematically contradict our RNA-Seq findings not because the RNA-Seq is incorrect, but because the reference gene normalization is unstable. Our research team has recently addressed this in a now published article (Barber et al., 2026). Furthermore, both methods derive measurements from the same extracted RNA, meaning they share identical upstream sources of biological and technical variation and do not constitute truly orthogonal validation in the way that, for example, protein-level quantification potentially would. Furthermore, the extracted RNA for these samples was sequenced in 2023 and is no longer available.

Beyond normalization concerns, RT-qPCR validation of a small subset of DEGs cannot meaningfully confirm the transcriptome-wide findings, pathway enrichment analyses, nor gene-level network interpretations that form the basis of our conclusions. Our DEG lists is comprised of several hundreds of genes, and validating 10–20 targets by RT-qPCR would represent a small percentage of identified DEGs; this would be biased toward high-confidence, large fold-change targets to be statistically representative of the larger dataset. Critically, the biological interpretability of our findings rests on the aggregate enrichment of functionally related gene sets, which RT-qPCR is fundamentally incapable of confirming. Selecting a handful of genes that confirm our hypothesis while omitting borderline or contradictory cases would introduce confirmation bias into the validation process rather than resolve it.

DOIs: 10.1186/gb-2002-3-7-research0034; 10.1373/clinchem.2008.112797; 10.1158/0008-5472.CAN-04-0496; 10.1093/bioinformatics/btp616; 10.1371/journal.pone.0352137; 10.1186/1471-2105-11-94; 10.1021/ac202028g

Results

• The authors are requested to present the Results and Discussion as separate sections.
• Section 3. Results and Discussion: I assume that the authors have submitted this manuscript as a full-length research article. However, the Results and Discussion have been presented as a single combined section, a format that is generally more appropriate for a short communication. If the manuscript is intended for publication as a full-length research article, the authors are requested to present the Results and Discussion as separate sections.

This is a full-length research article and as such we have separated the two sections and the manuscript now reflects both the results (Section 3) and discussion (Section 4) sections.

• The authors are requested to correlate the antibody titer data with the transcriptomic findings, which would help in understand whether transcriptomic vaccine responses were correspond to humoral immunity?

Thank you for bringing this to our attention. The previous work that includes titer analysis is only in a subset of healthy calves and are not available for the full cohort analyzed here including the animals that subsequently were diagnosed with BRD.  A direct correlation between titers and this dataset is therefore not possible.

• Figures 2 & 4 have significant readability issues. The labels are not clearly visible and overlap with each other. The figures require significant revision.

Thank you for bringing this to our attention.

Discussion

• The authors are requested to discuss the rationale for the specific four sampling timepoints (T1-4; median age 107, 114, 183, and 283 respectively).

This suggestion has been incorporated in Section 2.3 Marketing Strategy, Sampling Timepoints, and BRD Case Definition paragraph 2 explaining our rationale.

• The authors are requested to discuss about in the T3 group, revaccination and surgical castration (median age=183 days) happened simultaneously, whether castration-associated inflammatory signaling could influenced the vaccine-associated signals?

Blood was drawn before any application of treatment or other management tactics were applied to the study animals. Further information has been incorporated in Section 2.3 Marketing Strategy, Sampling Timepoints, and BRD Case Definition paragraph 2 and further discussed in Sections 4.1 Time-Dependent Gene Expression Patterns paragraph 2 and Section 4.2 Vaccination-associated Gene Expression paragraph 3.

• Whole blood transcriptomics data and neutrophil degranulation pathways are variable according to the leukocyte subpopulation shift, which happens during stress, age, and vaccination. Hence, the authors are requested to discuss/justify how this variation was overcome?

Thank you and this is an important consideration. We have added additional information to the discussion (Section 4.3) and have updated the language to reflect that transcriptional patterns are upstream of what is happening in the cell and further work would be needed to understand that full relationship.

• The authors are requested to discuss about in the T1 group, calves were vaccinated at a median age of 107 days, whether maternal antibodies interference the vaccine-induced transcriptional response?

This suggestion has been incorporated into Section 4.6 Principal Findings and Limitations paragraph 2.

• The authors mentioned as “absence of antiviral signalling” (absence of interferon signalling), but which not a biologically appropriate terminology. The authors are requested to use the proper language like “was not detected” rather than “absence of”.

Thank you and this language has been corrected throughout Section 3 Results and Section 4 Discussion.

Conclusion

• The authors are requested to mention the key limitation in the conclusion that BRD × VAX interaction was with minimal DEGs (vaccination status did not modify BRD-associated gene expression), which is one of the more clinically relevant findings of this manuscript.

Thank you and additional information has been added in Section 4.6 Principal Findings and Limitations paragraph 1.

• The authors are requested to write the targeted and actionable recommendations for future research, which would strengthen this manuscript.

In Section 4.6 Principal Findings and Limitations paragraph 2 future work is directly defined in the last sentence. Thank you.

References

The authors are requested to check and format once again all the references according to journal format especially while abbreviating the Journal names. In some References, volume, issue and page numbers are missing.

References have been updated and reviewed to ensure correctness.

Reviewer 2 Report

Comments and Suggestions for Authors

Comments:
The manuscript by McAllister et al. reported an important topic in bovine health management: whether preweaning vaccination influences host gene expression patterns and whether peripheral blood transcriptomic signatures can be associated with later bovine respiratory disease development. The longitudinal sampling design, inclusion of vaccinated and unvaccinated calves, and use of RNA-seq-based transcriptomic analysis are potentially valuable. The study may provide useful information for understanding vaccine-associated immune signatures and future BRD susceptibility in beef cattle. However, the following concerns have to be properly addressed before the manuscript can be considered for publication in Vaccines.

Major concerns:

1. The study design is complex, involving vaccination and control groups, four sampling timepoints, revaccination, castration, weaning, and subsequent BRD monitoring during the backgrounding phase. A schematic figure should be added to clearly show the animal grouping, vaccination schedule, sample collection timepoints, and BRD outcome classification. This would greatly improve readability and help readers understand the logic of the study.
2. The manuscript reports many DEGs related to TIME, VAX, BRD, and interactions, but the biological meaning and consistency between statistical models are not always clear. The authors should clearly present representative DEGs, their direction of regulation, fold changes, adjusted p-values, and biological relevance. A focused table listing key DEGs for each major comparison would strengthen the manuscript.
3. Since the conclusions rely heavily on differential gene expression and pathway enrichment analyses, the lack of confirmatory PCR data weakens the reliability of the biological interpretation.
4. The manuscript suggests that blood gene expression at weaning may indicate later BRD susceptibility. This is potentially important, but the authors should avoid overstating the predictive value unless formal predictive modeling is performed. If prediction is claimed, the authors should provide model performance metrics such as sensitivity, specificity, AUC, or cross-validation results.

Minor concerns:

1. The current Materials and Methods section contains several important components. However, these parts are presented in long, continuous text, which makes the experimental workflow difficult to follow. The authors should divide Section 2 into clear subsections
2. Please double-check spelling and typographical errors, such as “intial,” “adminstered,” “vaccaination,” “comparsion,” “downregualted,” and similar mistakes.

Author Response

The manuscript by McAllister et al. reported an important topic in bovine health management: whether preweaning vaccination influences host gene expression patterns and whether peripheral blood transcriptomic signatures can be associated with later bovine respiratory disease development. The longitudinal sampling design, inclusion of vaccinated and unvaccinated calves, and use of RNA-seq-based transcriptomic analysis are potentially valuable. The study may provide useful information for understanding vaccine-associated immune signatures and future BRD susceptibility in beef cattle. However, the following concerns have to be properly addressed before the manuscript can be considered for publication in Vaccines.

Major concerns:

1. The study design is complex, involving vaccination and control groups, four sampling timepoints, revaccination, castration, weaning, and subsequent BRD monitoring during the backgrounding phase. A schematic figure should be added to clearly show the animal grouping, vaccination schedule, sample collection timepoints, and BRD outcome classification. This would greatly improve readability and help readers understand the logic of the study.

Additional schematics have been added in the supplemental files to improve clarity. Thank you for the suggestion.

2. The manuscript reports many DEGs related to TIME, VAX, BRD, and interactions, but the biological meaning and consistency between statistical models are not always clear. The authors should clearly present representative DEGs, their direction of regulation, fold changes, adjusted p-values, and biological relevance. A focused table listing key DEGs for each major comparison would strengthen the manuscript.

Thank you for the suggestion and an additional table (Table 1) highlighting DEGs related to pathways mentioned in the manuscript has been included.

3. Since the conclusions rely heavily on differential gene expression and pathway enrichment analyses, the lack of confirmatory PCR data weakens the reliability of the biological interpretation.

We respectfully address the reviewer's request for RT-qPCR validation of our bulk RNA-Seq findings. While we appreciate the suggestion and acknowledge RT-qPCR's historical role in gene expression confirmation, we contend that RT-qPCR validation would not meaningfully strengthen the conclusions drawn from our RNA-Seq analysis and may in fact introduce methodological artifacts that could obscure biological signals. Our RNA-Seq data were generated with strict biological replicates per condition to meet estimated statistical power and processed using a standardized bioinformatics pipeline employing a transcriptome-wide normalization strategy that is statistically robust to compositional differences between samples. In contrast, RT-qPCR normalization depends on the stability of one or several endogenous reference genes across experimental conditions. Given the nature of our experimental system, it has been established in previous literature that commonly used housekeeping genes such as GAPDH and ACTB are themselves differentially expressed under such conditions, meaning RT-qPCR results could systematically contradict our RNA-Seq findings not because the RNA-Seq is incorrect, but because the reference gene normalization is unstable. Our research team has recently addressed this in a now published article (Barber et al., 2026). Furthermore, both methods derive measurements from the same extracted RNA, meaning they share identical upstream sources of biological and technical variation and do not constitute truly orthogonal validation in the way that, for example, protein-level quantification potentially would. Furthermore, the extracted RNA for these samples was sequenced in 2023 and is no longer available.

Beyond normalization concerns, RT-qPCR validation of a small subset of DEGs cannot meaningfully confirm the transcriptome-wide findings, pathway enrichment analyses, nor gene-level network interpretations that form the basis of our conclusions. Our DEG lists is comprised of several hundreds of genes, and validating 10–20 targets by RT-qPCR would represent a small percentage of identified DEGs; this would be biased toward high-confidence, large fold-change targets to be statistically representative of the larger dataset. Critically, the biological interpretability of our findings rests on the aggregate enrichment of functionally related gene sets, which RT-qPCR is fundamentally incapable of confirming. Selecting a handful of genes that confirm our hypothesis while omitting borderline or contradictory cases would introduce confirmation bias into the validation process rather than resolve it.

DOIs: 10.1186/gb-2002-3-7-research0034; 10.1373/clinchem.2008.112797; 10.1158/0008-5472.CAN-04-0496; 10.1093/bioinformatics/btp616; 10.1371/journal.pone.0352137; 10.1186/1471-2105-11-94; 10.1021/ac202028g

4. The manuscript suggests that blood gene expression at weaning may indicate later BRD susceptibility. This is potentially important, but the authors should avoid overstating the predictive value unless formal predictive modeling is performed. If prediction is claimed, the authors should provide model performance metrics such as sensitivity, specificity, AUC, or cross-validation results.

The authors agree with this suggestion and language has been softened throughout the manuscript.

Minor concerns:

1. The current Materials and Methods section contains several important components. However, these parts are presented in long, continuous text, which makes the experimental workflow difficult to follow. The authors should divide Section 2 into clear subsections

Thank you. Materials and Methods have been split into subsections.

2. Please double-check spelling and typographical errors, such as “intial,” “adminstered,” “vaccaination,” “comparsion,” “downregualted,” and similar mistakes.

Spelling has been corrected throughout the manuscript. Thank you.

Reviewer 3 Report

Comments and Suggestions for Authors

I found the study interesting and potentially valuable because it combines a longitudinal sampling design with transcriptomic profiling to explore both vaccination-associated responses and future BRD susceptibility in beef calves. The dataset is substantial, the biological question is relevant to the cattle industry, and the attempt to link pre-weaning transcriptomic signatures with later BRD outcomes is novel.

However, in its current form, I have significant concerns regarding overinterpretation of the findings, statistical rigor, study design limitations, and the distinction between association and prediction. Throughout the manuscript, the authors frequently make strong mechanistic and translational claims that are only partially supported by the data presented. In several places, the Discussion moves beyond the evidence generated by the study and presents speculative explanations as if they were demonstrated findings. In addition, details regarding model structure, sample attrition, BRD case distribution, and validation of predictive claims are insufficient.

Major Comments

1. Claims regarding prediction of BRD susceptibility are not supported by the analyses presented

The authors repeatedly state that blood transcriptomics may be used to identify cattle at risk of future BRD (e.g., lines 83–85, 416–439, 503–505, 528–529). However, the analyses performed are differential expression analyses comparing animals retrospectively classified as BRD or NOBRD.

The manuscript does not include:

• predictive modelling,
• classification analysis,
• ROC/AUC assessment,
• cross-validation,
• independent validation cohorts,
• sensitivity or specificity estimates.

Therefore, the authors have demonstrated association, not prediction.

For example, lines 417–419 state that the 129 BRD-associated DEGs provide evidence that "transcriptomic susceptibility signatures precede clinical disease." This is reasonable. However, statements such as line 433–435 suggesting future precision intervention strategies and line 528–529 implying prediction of subsequent disease overstate the findings.

I recommend substantially softening these claims throughout the manuscript and clearly distinguishing between:

• biomarkers associated with later disease occurrence, and
• biomarkers that have been validated for prediction.

These are not equivalent.

 

2. Serious concerns regarding confounding of age and management events

The study interprets T3 as a biologically meaningful period of immune maturation and vaccine-associated response (lines 256–274, 321–328, 451–462).

However, T3 coincides with surgical castration (lines 124–125 and 130–131).

Castration is a major inflammatory and physiological stressor capable of altering:

• leukocyte dynamics,
• cytokine production,
• acute phase responses,
• transcriptomic profiles.

The manuscript repeatedly attributes T3-specific transcriptional patterns to vaccination and developmental immune maturation without adequately accounting for the potential effects of castration.

For example:

• lines 316–318 report enrichment of stress response and neutrophil degranulation pathways,
• lines 322–323 describe a prolonged vaccine response,
• lines 458–462 interpret the findings as evidence of immune memory development.

Because all calves were castrated at T3, it is impossible to disentangle vaccine effects from castration-related effects at this sampling point.

The statement on lines 271–274 that the T3 sample was collected before castration reduces this concern but does not fully eliminate it because:

1. animals were sampled immediately before a planned invasive procedure,
2. handling and restraint effects remain possible,
3. developmental age and management interventions remain completely confounded.

The authors should more formally acknowledge this issue and avoid attributing T3 phenomena primarily to vaccination.

 

3. The statistical strategy is insufficiently justified and potentially problematic

The combination of GLMMseq and edgeR analyses (lines 184–193) is difficult to follow and raises several concerns.

Specifically:

• Why was significance defined using overlap between GLMMseq and edgeR results?
• What biological or statistical rationale justified using FDR <0.05 for one analysis and FDR <0.10 for another (lines 186–190)?
• How were discrepancies between methods handled?
• Were interaction terms fitted simultaneously in a single model?
• Were pasture and animal treated as nested random effects?
• Was repeated-measures correlation assessed?

Furthermore, BRD status is determined after transcriptome collection, yet it is included as a factor in the differential expression model (lines 181–182). The rationale is briefly discussed on lines 194–197 but remains unclear.

More importantly, the manuscript reports:

• 11,068 TIME genes from GLMMseq,
• 9,241 BRD genes from GLMMseq,

yet only 129 BRD genes remained after the filtering strategy (lines 235–238).

Such drastic reductions raise questions about model stability and false positive control that are not adequately addressed.

The statistical framework requires far more transparency.

 

4. Insufficient information regarding BRD outcomes and statistical power

The manuscript does not clearly report:

• how many calves developed BRD,
• how many remained healthy,
• BRD incidence within VAX and NOVAX groups,
• BRD incidence within DIRECT and AUCTION groups,
• whether vaccination reduced BRD risk.

This omission is surprising given that BRD status represents a central outcome variable.

The reader cannot evaluate:

• balance between groups,
• statistical power,
• possible subgroup biases,
• biological significance of the BRD comparisons.

Moreover, the manuscript repeatedly discusses vaccination effects on future BRD susceptibility (lines 84–85, 473–488), but no clear epidemiologic analysis is presented to demonstrate whether vaccination altered BRD incidence.

A table summarizing group allocation, sample losses, BRD incidence, and final analytical sample sizes is essential.

 

5. Overinterpretation of mechanistic biology

Numerous sections infer biological mechanisms that are not directly measured.

Examples include:

• active inflammatory resolution (lines 263–264),
• impaired oxygen-dependent immune function (lines 424–426),
• early subclinical pulmonary abnormalities (line 427),
• prolonged antigen availability (line 467),
• durable memory formation (line 467),
• verification of vaccination history using transcriptomics (lines 360–362 and 527–528).

The study measured blood mRNA abundance only.

There were no data on:

• antibody titres,
• antigen persistence,
• T-cell responses,
• pulmonary pathology,
• oxygen transport physiology,
• vaccine efficacy endpoints.

Many mechanistic statements should therefore be reframed as hypotheses rather than conclusions.

 

6. The manuscript repeatedly uses causal language despite an observational outcome framework

Although vaccination was randomized, BRD status was not.

Examples include:

• lines 416–417 ("suggests that host gene expression at weaning could be indicative of subsequent BRD susceptibility"),
• lines 430–432 ("indicated immune gene expression patterns accumulate before pathogen exposure"),
• lines 487–488 ("host response factors rather than immune status modified by vaccine are quantifiable and targetable").

The analyses do not establish causality. The authors should consistently use language such as:

• associated with,
• correlated with,
• linked to,

rather than implying biological causation.

 

Minor Comments

1. Sample number inconsistencies require clarification

The study begins with 81 calves (line 23, lines 103–104), but sequencing analyses include only 73 animals (lines 150–151).

The pathway by which 81 became 73 is unclear.

Seven failed samples are listed (lines 152–154), but this alone does not explain the reduction. A CONSORT-style flow diagram would be helpful.

 

2. Research questions are written informally

Lines 83–85 present the research questions as: "can we predict future BRD outcome..."

These should be rewritten as formal hypotheses or objectives.

 

3. Figure references are not always sufficiently integrated

Figures 1–5 are repeatedly cited, but statistical outcomes shown in those figures are often not quantified in the text. For example, Figure 4 is used to support separation of T3 VAX animals (lines 323–324), but there are no statistical measures of cluster separation.

 

4. Use of the term "predict vaccination history"

Lines 360–362 and 500–501 suggest that transcriptomics can verify vaccination history.

This is a highly speculative statement that is not demonstrated. No classification framework was developed.

 

5. Missing description of sequencing depth

The manuscript reports sequencing platform information (lines 148–149) but does not provide:

• average reads per sample,
• mapping percentages,
• alignment quality metrics.

These are important RNA-seq quality indicators.

 

6. K-means clustering approach requires more explanation

Lines 206–207 state that 36 clusters were selected using the Elbow method.

The manuscript should show:

• Elbow plot results,
• rationale for 36 clusters,
• robustness assessment.

Thirty-six clusters appear unusually high and is currently insufficiently justified.

 

7. Typographical and grammatical issues

Several errors should be corrected:

• "analyis" (line 266)
• "backgounding" (line 267)
• "intial" (lines 327, 349)
• "identifed" (line 311)
• "adminstered" (line 371)
• "vaccaination" (line 373)
• "comparsion" (line 374)
• "obsevation" (line 378)
• "inteferon" (line 379)
• "downregualted" (line 380)
• "vaccination-associcated" (line 481)

There are numerous additional spacing and formatting inconsistencies throughout the manuscript.

 

8. Discussion and Results are overly intertwined

Section 3 is entitled "Results and Discussion", but the manuscript often presents long speculative interpretations immediately after reporting findings.

For example, lines 261–270, 378–386, and 424–435 contain extensive biological speculation.

I recommend separating results from interpretation more clearly or reducing speculation.

 

9. Conclusions exceed the evidence

The statements on lines 527–532 regarding vaccination history prediction, transcriptomic risk profiling, genomic selection, and optimized vaccination strategies are ambitious and extend considerably beyond what was directly tested.

The conclusions should focus on demonstrated findings rather than future applications.

Author Response

Comments and Suggestions for Authors

I found the study interesting and potentially valuable because it combines a longitudinal sampling design with transcriptomic profiling to explore both vaccination-associated responses and future BRD susceptibility in beef calves. The dataset is substantial, the biological question is relevant to the cattle industry, and the attempt to link pre-weaning transcriptomic signatures with later BRD outcomes is novel.

However, in its current form, I have significant concerns regarding overinterpretation of the findings, statistical rigor, study design limitations, and the distinction between association and prediction. Throughout the manuscript, the authors frequently make strong mechanistic and translational claims that are only partially supported by the data presented. In several places, the Discussion moves beyond the evidence generated by the study and presents speculative explanations as if they were demonstrated findings. In addition, details regarding model structure, sample attrition, BRD case distribution, and validation of predictive claims are insufficient.

Major Comments

 

1. Claims regarding prediction of BRD susceptibility are not supported by the analyses presented

The authors repeatedly state that blood transcriptomics may be used to identify cattle at risk of future BRD (e.g., lines 83–85, 416–439, 503–505, 528–529). However, the analyses performed are differential expression analyses comparing animals retrospectively classified as BRD or NOBRD.

The manuscript does not include:

• predictive modelling,
• classification analysis,
• ROC/AUC assessment,
• cross-validation,
• independent validation cohorts,
• sensitivity or specificity estimates.

Therefore, the authors have demonstrated association, not prediction.

For example, lines 417–419 state that the 129 BRD-associated DEGs provide evidence that "transcriptomic susceptibility signatures precede clinical disease." This is reasonable. However, statements such as line 433–435 suggesting future precision intervention strategies and line 528–529 implying prediction of subsequent disease overstate the findings.

I recommend substantially softening these claims throughout the manuscript and clearly distinguishing between:

• biomarkers associated with later disease occurrence, and
• biomarkers that have been validated for prediction.

These are not equivalent.

Thank you for your suggestions. The objectives have been reframed to reflect associations and an explicit note that no prediction models were constructed or tested. Additionally, clarification was added around BRD status being a later clinical outcome and BRD DEGs describe association with a later outcome and not prediction of it (Section 2.5 Differential Gene Expression Analysis paragraph 2). Previous claims about precision interventions and prediction of subsequent disease have been made conditional on future validation and softened throughout the manuscript.

 

2. Serious concerns regarding confounding of age and management events

The study interprets T3 as a biologically meaningful period of immune maturation and vaccine-associated response (lines 256–274, 321–328, 451–462).

However, T3 coincides with surgical castration (lines 124–125 and 130–131).

Castration is a major inflammatory and physiological stressor capable of altering:

• leukocyte dynamics,
• cytokine production,
• acute phase responses,
• transcriptomic profiles.

The manuscript repeatedly attributes T3-specific transcriptional patterns to vaccination and developmental immune maturation without adequately accounting for the potential effects of castration.

For example:

• lines 316–318 report enrichment of stress response and neutrophil degranulation pathways,
• lines 322–323 describe a prolonged vaccine response,
• lines 458–462 interpret the findings as evidence of immune memory development.

Because all calves were castrated at T3, it is impossible to disentangle vaccine effects from castration-related effects at this sampling point.

The statement on lines 271–274 that the T3 sample was collected before castration reduces this concern but does not fully eliminate it because:

1. animals were sampled immediately before a planned invasive procedure,
2. handling and restraint effects remain possible,
3. developmental age and management interventions remain completely confounded.

The authors should more formally acknowledge this issue and avoid attributing T3 phenomena primarily to vaccination.

Thank you. We have added clarification (Section 4.1 paragraph 2) and softened language around these points throughout the discussion (Section 4).

 

3. The statistical strategy is insufficiently justified and potentially problematic

The combination of GLMMseq and edgeR analyses (lines 184–193) is difficult to follow and raises several concerns.

Specifically:

• Why was significance defined using overlap between GLMMseq and edgeR results?
• What biological or statistical rationale justified using FDR <0.05 for one analysis and FDR <0.10 for another (lines 186–190)?
• How were discrepancies between methods handled?
• Were interaction terms fitted simultaneously in a single model?
• Were pasture and animal treated as nested random effects?
• Was repeated-measures correlation assessed?

Furthermore, BRD status is determined after transcriptome collection, yet it is included as a factor in the differential expression model (lines 181–182). The rationale is briefly discussed on lines 194–197 but remains unclear.

More importantly, the manuscript reports:

• 11,068 TIME genes from GLMMseq,
• 9,241 BRD genes from GLMMseq,

yet only 129 BRD genes remained after the filtering strategy (lines 235–238).

Such drastic reductions raise questions about model stability and false positive control that are not adequately addressed.

The statistical framework requires far more transparency.

Thank you for your suggestion and Section 2.5 has been substantially expanded and now should address our justifications and rationales for the statistical methods chosen. In Section 2.5 Differential Gene Expression Analysis paragraph one there is now a more thorough discussion on modeling. In Section 2.5 Paragraph 2 there is now a discussion on rationale for the FDRs chosen and further discussion on the statistical approach.

 

4.  Insufficient information regarding BRD outcomes and statistical power

The manuscript does not clearly report:

• how many calves developed BRD,
• how many remained healthy,
• BRD incidence within VAX and NOVAX groups,
• BRD incidence within DIRECT and AUCTION groups,
• whether vaccination reduced BRD risk.

This omission is surprising given that BRD status represents a central outcome variable.

The reader cannot evaluate:

• balance between groups,
• statistical power,
• possible subgroup biases,
• biological significance of the BRD comparisons.

Moreover, the manuscript repeatedly discusses vaccination effects on future BRD susceptibility (lines 84–85, 473–488), but no clear epidemiologic analysis is presented to demonstrate whether vaccination altered BRD incidence.

A table summarizing group allocation, sample losses, BRD incidence, and final analytical sample sizes is essential.

 Thank you for the suggestion, additional information about statistical power and BRD incidence have been added in Section 2.5 Paragraph 3. We appreciate the suggestion of a table however this information was included in the Methods where appropriate (Section 2.3 Paragraph 2, Section 2.4 Paragraph 1, Section 2.5 paragraph 1).

 

5. Overinterpretation of mechanistic biology

Numerous sections infer biological mechanisms that are not directly measured.

Examples include:

• active inflammatory resolution (lines 263–264),
• impaired oxygen-dependent immune function (lines 424–426),
• early subclinical pulmonary abnormalities (line 427),
• prolonged antigen availability (line 467),
• durable memory formation (line 467),
• verification of vaccination history using transcriptomics (lines 360–362 and 527–528).

The study measured blood mRNA abundance only.

There were no data on:

• antibody titres,
• antigen persistence,
• T-cell responses,
• pulmonary pathology,
• oxygen transport physiology,
• vaccine efficacy endpoints.

Many mechanistic statements should therefore be reframed as hypotheses rather than conclusions.

 Language has been softened throughout the manuscript. Specifically in Section 4.2 paragraph 2, Section 4.3 paragraph 1, and Section 4.6 paragraph 2.  

 

6. The manuscript repeatedly uses causal language despite an observational outcome framework

Although vaccination was randomized, BRD status was not.

Examples include:

• lines 416–417 ("suggests that host gene expression at weaning could be indicative of subsequent BRD susceptibility"),
• lines 430–432 ("indicated immune gene expression patterns accumulate before pathogen exposure"),
• lines 487–488 ("host response factors rather than immune status modified by vaccine are quantifiable and targetable").

The analyses do not establish causality. The authors should consistently use language such as:

• associated with,
• correlated with,
• linked to,

rather than implying biological causation.

 Language throughout the manuscript has been softened focusing in the discussion (Section 4). Thank you for your suggestions.

Minor Comments

1. Sample number inconsistencies require clarification

The study begins with 81 calves (line 23, lines 103–104), but sequencing analyses include only 73 animals (lines 150–151).

The pathway by which 81 became 73 is unclear.

Seven failed samples are listed (lines 152–154), but this alone does not explain the reduction. A CONSORT-style flow diagram would be helpful.

 A correction has been made in the first paragraph of Section 2.4 Next-generation Sequencing and Bioinformatic Data Processing. It is now explicitly stated how samples were chosen to get from the original enrolled 81 calves to the 73 animals that were sequenced and a total of 292 samples.

2. Research questions are written informally

Lines 83–85 present the research questions as: "can we predict future BRD outcome..."

These should be rewritten as formal hypotheses or objectives.

Formal objectives have been reworked into the final paragraph of the Introduction.

3. Figure references are not always sufficiently integrated

Figures 1–5 are repeatedly cited, but statistical outcomes shown in those figures are often not quantified in the text. For example, Figure 4 is used to support separation of T3 VAX animals (lines 323–324), but there are no statistical measures of cluster separation.

 Thank you. Section 3.2 Time Dependent Gene Expression Patterns paragraph 2 directly addresses this suggestion.

4. Use of the term "predict vaccination history"

Lines 360–362 and 500–501 suggest that transcriptomics can verify vaccination history.

This is a highly speculative statement that is not demonstrated. No classification framework was developed.

 Thank you. The statement has been softened and framed as a hypothesis.

5. Missing description of sequencing depth

The manuscript reports sequencing platform information (lines 148–149) but does not provide:

• average reads per sample,
• mapping percentages,
• alignment quality metrics.

These are important RNA-seq quality indicators.

 This suggestion has been incorporated in Section 2.4 paragraph 2.

6. K-means clustering approach requires more explanation

Lines 206–207 state that 36 clusters were selected using the Elbow method.

The manuscript should show:

• Elbow plot results,
• rationale for 36 clusters,
• robustness assessment.

Thirty-six clusters appear unusually high and is currently insufficiently justified.

 Thank you and Additional information addressing your suggestions will be included in the Supplemental Files.

7. Typographical and grammatical issues

Several errors should be corrected:

• "analyis" (line 266)
• "backgounding" (line 267)
• "intial" (lines 327, 349)
• "identifed" (line 311)
• "adminstered" (line 371)
• "vaccaination" (line 373)
• "comparsion" (line 374)
• "obsevation" (line 378)
• "inteferon" (line 379)
• "downregualted" (line 380)
• "vaccination-associcated" (line 481)

There are numerous additional spacing and formatting inconsistencies throughout the manuscript.

 Spelling has been corrected throughout the manuscript. Thank you.

8. Discussion and Results are overly intertwined

Section 3 is entitled "Results and Discussion", but the manuscript often presents long speculative interpretations immediately after reporting findings.

For example, lines 261–270, 378–386, and 424–435 contain extensive biological speculation.

I recommend separating results from interpretation more clearly or reducing speculation.

 Results and Discussions are now two separate sections in the manuscript.

9. Conclusions exceed the evidence

The statements on lines 527–532 regarding vaccination history prediction, transcriptomic risk profiling, genomic selection, and optimized vaccination strategies are ambitious and extend considerably beyond what was directly tested.

The conclusions should focus on demonstrated findings rather than future applications.

Conclusion and the language around the conclusions have been softened. Thank you.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I have carefully reviewed the revised manuscript and the authors’ responses. Most of the Reviewer comments have been addressed adequately and comprehensively, and the authors have made thoughtful and meticulous revisions to enhance the manuscript.

However, few queries remain unaddressed while revising the manuscript.

• In the Discussion Section, the authors are requested to write the discussion without any subheadings by correlating the findings of each other.
• Figures 2 & 4 have significant readability issues. The labels are not clearly visible and overlap with each other. The figures require significant revision.

Author Response

Open Review ( ) I would not like to sign my review report
(x) I would like to sign my review report Quality of English Language ( ) The English could be improved to more clearly express the research.
(x) The English is fine and does not require any improvement.            

	Yes	Can be improved	Must be improved	Not applicable
Does the introduction provide sufficient background and include all relevant references?	(x)	( )	( )	( )
Is the research design appropriate?	(x)	( )	( )	( )
Are the methods adequately described?	(x)	( )	( )	( )
Are the results clearly presented?	(x)	( )	( )	( )
Are the conclusions supported by the results?	(x)	( )	( )	( )
Are all figures and tables clear and well-presented?	(x)	( )	( )	( )

    Comments and Suggestions for Authors

I have carefully reviewed the revised manuscript and the authors’ responses. Most of the Reviewer comments have been addressed adequately and comprehensively, and the authors have made thoughtful and meticulous revisions to enhance the manuscript.

However, few queries remain unaddressed while revising the manuscript.

• In the Discussion Section, the authors are requested to write the discussion without any subheadings by correlating the findings of each other.
• Figures 2 & 4 have significant readability issues. The labels are not clearly visible and overlap with each other. The figures require significant revision

We would like to thank the reviewer for their time and helpful comments to our revised manuscript. Regarding the figures, we believe this was due ,in part, to file formatting once the PDF version of the manuscript was generated and uploaded. We have corrected the axis overlapping to Figure 2 and have ensured that all figures now uploaded are legible and of high resolution. Regarding the Discussion section subheadings, we would maintain how they are currently written for two reasons. First, we believe that, due to the number of analytical approaches and sub objectives of this project (i.e., VAX, BRD, TIME, TIME x VAX, etc.), the subheadings help direct readers toward the context and interpretations of all relevant findings; removing these subheadings would create a continuous flow of information that may be difficult to disentangle at times. Second, it was suggested by the editorial team and another reviewer that we generate the results and discussions into two separate sections and provide specific context to our interpretations. We believe that these subheadings provide such context and that the Abstract and Conclusion sections concatenate these data and results in a concise manner.

Reviewer 2 Report

Comments and Suggestions for Authors

No further comment.

Author Response

Thank you for your time and input. 

Reviewer 3 Report

Comments and Suggestions for Authors

I have no further comments.

Author Response

Thank you for your time and input.