Reference Gene Stability in Agrostemma githago Using Quantitative Real-Time PCR

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thanks for inviting me to review this manuscript. The work falls within the scope of the journal and provides useful data for future gene expression studies in Agrostemma githago and related species. However, the novelty, completeness, and presentation of the manuscript require substantial improvements before it meets the standards for acceptance.

Major comments:

1. The candidate reference genes selected are all classic, well-characterized housekeeping genes commonly used across plant research. Given that the genome assembly of A. githago has been published and RNA-Seq datasets are presumably available, it would significantly enhance the novelty of this study to incorporate novel reference gene candidates. Therefore, I recommend adding an initial analysis step where both classic and novel candidate genes are identified from public or in-house RNA-Seq datasets.
2. The authors should provide detailed sequence information for all candidate reference genes as supplementary file. This should include gene/transcript IDs (linked to the published genome assembly), genomic DNA sequences, cDNA sequences, protein sequences, and annotated features such as start/stop codons, exon/introns, and primer binding sites. Such information is critical for reproducibility, as researchers ofter struggle to cross-reference gene names with updated genome annotations.
3. The authors could supplement the melt curve analysis (Figure 1) with agarose gel electrophoresis results. Gel data would visually confirm the presence of a single amplicon of the expected size, providing more robust evidence of primer efficiency and specificity.
4. Several figures should be improved:

1. Figure 4 presents results from two independent algorithms (geNorm as an orange line and NormFinder as blue bars) in a single plot. This is misleading, as readers may incorrectly interpret the line and bars as complementary data from the same method.
2. Figure 5 includes a redundant table beneath the graph that replicates the pairwise variation values. This table is unnecessay and could be removed or moved to supplementary if needed.
3. Figure 7 uses figure legends to identify samples, forcing readers to cross-them with data points. To improve readability, I suggest placing sample identifiers directly on the x-aixs.
4. Figure 4 and 6 both use "Stability values" as the y-aixs label but with different scales (0-1 vs. 0-10), which is confusing. Please adjust the labels to relect the specific metric and ensure scale clarity.
5. Figure legends (annotations after figure title) need more detailed explanations to enable independent interpretation.

Minor comments:

1. There are several grammatical errors and typos that require correction. Key examples include:

1. Table 2, "Delta CT" -> ΔCt
2. Line 381, "TIFA1-2" -> TIF5A1-2
3. Line 456, "aplication" -> application
4. Line 522, "Wroclaw Medcal University" -> Medical
5. Line 549, "comercialy available soil" -> commercially
6. Line 553, "developemental stages" -> developmental
7. Line 554, "vegetative seasson" -> season

2. Although an abbreviation section is provided, abbreviations should be paired with their full names at first mention, such as RIPs (ribosome-inactivating proteins), SA (salicylic acid), MeJA (methyl jasmonate), etc.

Author Response

Dear Reviewer,

We sincerely thank you for taking the time to review our manuscript and for providing constructive and insightful comments. Your feedback has been invaluable in improving the quality and clarity of our work. We have carefully considered each suggestion, revised the manuscript accordingly, and highlighted all changes using the track-changes function. Detailed point-by-point responses to all issues raised are provided below, with our answers outlined in red corresponding directly to the reviewer’s remarks.

 

Major comments:

Comment 1: The candidate reference genes selected are all classic, well-characterized housekeeping genes commonly used across plant research. Given that the genome assembly of A. githago has been published and RNA-Seq datasets are presumably available, it would significantly enhance the novelty of this study to incorporate novel reference gene candidates. Therefore, I recommend adding an initial analysis step where both classic and novel candidate genes are identified from public or in-house RNA-Seq datasets.

 Response 1: Thank you for bringing this point up. We agree that incorporating additional analyses of novel candidate reference genes could further enhance the novelty of the study. Nevertheless, we deliberately chose to restrict our analysis to seven well-established housekeeping genes for the following reasons:

1. a) Although the genome of Agrostemma githago has been published recently (with raw sequencing reads available earlier but gene annotation completed only very recently), transcriptomic resources for this species remain extremely limited. Currently, only a single RNA-seq dataset is publicly available in NCBI/SRA (BioSample: SAMN41876039; sample name: P451_124; SRA: SRS21684434). This dataset represents a pooled transcriptome derived from fully mature organs of githago, including roots, stems, leaves, developing buds, flowers, and fruits. As a result, it does not permit organ-specific assessment of transcript abundance, which is essential for evaluating the expression stability of potential reference genes.

A second RNA-seq dataset, representing the transcriptome of a very young flower, was published by our collaborators (Weise et al., 2020; 10.1038/s41598-020-72282-2). Collectively, the available A. githago RNA-seq datasets provide reliable nucleotide sequence information but lack robust quantitative expression data (e.g., FPKM/TPM values) suitable for stability analysis. Consequently, to obtain sequences of well-characterized housekeeping genes, we relied on the annotated transcriptome provided by Weise et al. (2020).

1. b) Numerous studies have attempted to identify universally stable reference genes in plants by mining large transcriptomic datasets encompassing multiple organs, developmental stages, and environmental conditions. Such comprehensive approaches are typically feasible for model plant species or major crops with extensive transcriptomic resources. In contrast, our study was not designed as a broad-scale discovery project. Instead, to facilitate downstream gene expression analyses in githago, we leveraged existing knowledge from Caryophyllaceae species, selected previously validated candidate reference genes, and systematically evaluated their stability under our specific experimental conditions. We believe that the primary strength of our work lies in the rigorous assessment of these pre-selected, strong candidate genes across 40 independent A. githago cDNA libraries representing diverse developmental stages and a wide range of cultivation conditions. Our study adopts a technical and application-oriented approach, building upon prior research to provide a practical and timely framework for selecting suitable reference genes for gene expression analyses in A. githago, whether grown under traditional soil conditions or cultivated in vitro.

While we fully acknowledge the importance of identifying novel, highly stable reference genes in plants, extending the present analysis to additional, uncharacterized genes would substantially alter the primary objective of this study. We therefore chose to focus on a targeted evaluation of established candidate genes, in line with the rationale outlined above. We sincerely hope that this reasoning is acceptable and clearly justifies our methodological choices.

A paragraph describing our methodological reasoning has been added to the Introduction section of the manuscript.

 

Comment 2: The authors should provide detailed sequence information for all candidate reference genes as supplementary file. This should include gene/transcript IDs (linked to the published genome assembly), genomic DNA sequences, cDNA sequences, protein sequences, and annotated features such as start/stop codons, exon/introns, and primer binding sites. Such information is critical for reproducibility, as researchers ofter struggle to cross-reference gene names with updated genome annotations.

Response 2: We thank the Reviewer for this important comment regarding reproducibility and annotation clarity. We fully agree that transparent linkage between transcript identifiers, genome assemblies, and functional annotation is critical, particularly for non-model organisms with evolving genome annotations.

In response, we have revised Supplementary Table S8 (now the Supplementary Table S7) to improve cross-referencing and interpretability. Specifically, we added BLAST-derived annotation columns (Top BLAST hit, BLAST max score, % identity) retrieved from the reference genome assembly dcAgrGith1.1 using the same parameters described in the manuscript. These values provide an explicit and reproducible link between the transcript sequences reported in this study and public reference databases.

Regarding sequence information, we would like to clarify the following. The nucleotide sequences provided in Table S7 represent experimentally derived transcriptome contigs generated in Weise et al. 2020 and therefore constitute the primary data underlying our analyses. In contrast, genomic DNA sequences, exon–intron structures, predicted coding sequences, and protein sequences from the dcAgrGith1.1 assembly originate from a recently published genome resource and are not generated by the authors of this manuscript. For this reason, we do not reproduce full genomic, transcript, or protein sequences from dcAgrGith1.1 in the Supplementary Materials.

Instead, to ensure reproducibility while respecting data provenance, we explicitly link our transcriptome-derived sequences to the published genome annotation via accession identifiers and BLAST-based homology. This approach enables readers to directly retrieve genomic coordinates, exon–intron structures, coding sequences, and protein annotations from the original genome resource.

With respect to annotated features such as start/stop codons, introns, and primer binding sites, we note that our analyses are based on cDNA sequences, and intronic regions are therefore not present. Primer sequences used in this study are reported elsewhere in the manuscript, while exon–exon junctions and coding features can be reconstructed by linking the reported contigs to the reference genome annotation.

Finally, we have clarified in the manuscript that BLAST annotation was performed against public NCBI databases without taxonomy restriction (i.e., across all organisms). This ensures transparency regarding the scope of homology searches and avoids misinterpretation as species-specific orthology.

We believe that these revisions substantially improve the clarity, traceability, and reproducibility of the presented data while maintaining appropriate separation between newly generated data and previously published genome resources.

 

Comment 3: The authors could supplement the melt curve analysis (Figure 1) with agarose gel electrophoresis results. Gel data would visually confirm the presence of a single amplicon of the expected size, providing more robust evidence of primer efficiency and specificity.

Response 3: We have supplemented Figure 1 with the agarose gel result. Each PCR product is represented by a single band; no unspecific bands or primer-dimers were detected.  The appropriate Result section has been amended accordingly.

Comment 4: Several figures should be improved:

1. Figure 4 presents results from two independent algorithms (geNorm as an orange line and NormFinder as blue bars) in a single plot. This is misleading, as readers may incorrectly interpret the line and bars as complementary data from the same method.

Results from geNorm and NormFinder has been represented by two separate figures – Figure 4 and Figure 6, respectively.

2. Figure 5 includes a redundant table beneath the graph that replicates the pairwise variation values. This table is unnecessay and could be removed or moved to supplementary if needed.

The table with pairwise variation values has been moved to the Supplementary table 2.

3. Figure 7 uses figure legends to identify samples, forcing readers to cross-them with data points. To improve readability, I suggest placing sample identifiers directly on the x-aixs.

We agree with this opinion, however, this particular figure has been produced by the Maestro software (dedicated to CFX data analysis) and Maestro does not allow for such modifications. Therefore, we had to leave the Figure 7 (now Figure 8) as it was produced by Maestro.

4. Figure 4 and 6 both use "Stability values" as the y-aixs label but with different scales (0-1 vs. 0-10), which is confusing. Please adjust the labels to relect the specific metric and ensure scale clarity.

We have tried to follow this recommendation, but unfortunately adjusting the labels and using a uniform scale have reduced the clarity of the charts due to the large differences in values (between charts of Figure 4, 6 and 7). Therefore, each figure presenting stability values was plotted using an individual scale optimized for graphical clarity. To further enhance readability, data labels were included in all charts of Figure 4, 6 and 7.

5. Figure legends (annotations after figure title) need more detailed explanations to enable independent interpretation.

Detailed explanations have been added in each figure legend to enable independent interpretation.

 

Minor comments:

 Comment 1: There are several grammatical errors and typos that require correction. Key examples include:

Table 2, "Delta CT" -> ΔCt

It has been corrected

Line 381, "TIFA1-2" -> TIF5A1-2

Corrected

Line 456, "aplication" -> application

Corrected

Line 522, "Wroclaw Medcal University" -> Medical

Corrected

Line 549, "comercialy available soil" -> commercially

Corrected

Line 553, "developemental stages" -> developmental

Corrected

Line 554, "vegetative seasson" -> season

Corrected

Comment 2: Although an abbreviation section is provided, abbreviations should be paired with their full names at first mention, such as RIPs (ribosome-inactivating proteins), SA (salicylic acid), MeJA (methyl jasmonate), etc.

Abbreviations have been paired with their full names at first mention.

 

 

The responses are also included in the attached PDF file

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is a novel work and will be very useful for researchers in future. However, materials and method section needs some improvements in terms of writing. The introduction and result section are presented well. However, I feel that result obtained should have been supported with available literatures. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Needs more work. some of the things are commented on PDF file as well but it needs some detailed review. I suggest to revise it by native English speaker. 

Author Response

Dear Reviewer, We sincerely thank you for your careful evaluation of our manuscript and for your constructive and insightful comments. Your feedback has been invaluable in improving the clarity, readability, and overall quality of our work. We have carefully addressed all suggestions provided in the PDF file. All corrections are highlighted in the revised manuscript file using the track-changes function. A detailed response to your comment is provided below, with our answer shown in red and reference to the corresponding modifications in the revised manuscript.

Comment 1: Result obtained should have been supported with available literatures.

Response 1: Thank you for bringing this point up. Following this as well as one of the other reviewers recommendation, we have amended the Discussion with some additional citations to consider the ecological differences between corn cockle and other Caryophyllaceae species as well as reports on more distantly related plants from the Caryophyllales order representing Amaranthaceae, e.g. pseudocereals quinoa, amaranth and several halophytes – Halostachys, Salicornia, Salsola, Suaeda. and added a paragraph comparing our results with studies published on these plants. In the original submission, some other studies have already discussed from the methodological point of view. Nonetheless, all these papers also align with the conclusions that each time, various genes may perform best/worst as reference and that a specific set should be established before an qRT-PCR measurement. 

Additional comments included in the PDF file

Abstract – the blind link (although, it was copied from the original source website and used to work earlier) has been replaced by the active one - https://www.ciidirsinaloa.com.mx/RefFinder-master/

Page 2 – references have been added

Page 3 – the sentence has been corrected

Page 3 – we decided to use qRT-PCR

The redundant letter L. (authorship in the binomial name) has been deleted

Pages 21/22 – terminology changed to “open-pollinated”

“cultivated” changed to “grown”

Data on the soil vendor and some details on it have been provided.

“cultivation” replaced with “experiment”

The fertilizer was in the ready-to-use substrate, not added. This information has been included in the revised manuscript.

 

Finally, the manuscript has been proofread for English correctness by a hired editing service associated with our institution.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This study provides a rigorously validated and reliable set of reference genes for molecular biology research in Agrostemma githago. This foundational work not only resolves the normalization challenge in gene expression quantification for this species, but its rigorous multi-algorithm validation paradigm also offers a replicable template for similar studies in other non-model plants. It will strongly advance subsequent in-depth exploration of the biosynthesis and regulatory mechanisms of medicinally active compounds in A. githago.

The following are the revision suggestions.

1. The reasons for selecting these 7 specific candidate genes can be briefly explained (such as whether they were selected based on transcriptome data or through phylogenetic conservation analysis), to enhance the logical coherence.

2. In the experimental design, is it possible that gene expression could be influenced by the circadian rhythm or the annual cycle?

3. In the methods part, it is possible to provide additional explanations regarding the basis for the cDNA dilution factor (such as the range of template concentrations), and mention the specific weight calculation method of RefFinder.

4. The results section can include brief explanatory text to highlight the key findings, instead of relying solely on the charts.

5. Table 2: The content of this table is comprehensive, but it is quite lengthy. Consider splitting it or simplifying the presentation method.

figuers: Optimize image details (axis labels, font size).

6. The discussion section can compare the research results of more closely related species to deepen the ecological or evolutionary perspectives.

Author Response

Dear Reviewer,

We are very grateful for your careful evaluation of our manuscript and your thoughtful suggestions for improvement. In revising the manuscript, we have addressed all comments in detail and implemented the recommended changes to enhance the clarity, readability, and overall quality of the work. All revisions have been marked using the track changes option for transparency. A detailed point-by-point response to the reviewer’s comments is provided below, with our answers highlighted in red and references to the specific modifications made in the revised manuscript.

Comment 1: The reasons for selecting these 7 specific candidate genes can be briefly explained (such as whether they were selected based on transcriptome data or through phylogenetic conservation analysis), to enhance the logical coherence.

Response 1: We thank the reviewer for highlighting this critical point. To support downstream gene expression analyses in Agrostemma githago L., we have selected previously validated candidate reference genes from Silene vulgaris [37], Dianthus caryophyllus [38] and Dianthus broteri [39] and found their orthologous sequences in the A. githago transcriptome (published by our collaborators in Weise et.al. 2020). This allowed us to design primer pairs and systematically assessed the stability of the preselected genes under our experimental conditions. We deliberately refrained from including analyses of novel candidate reference genes because the study was not intended as a broad-scale discovery project. We believe that the main strength of our study lies in the rigorous evaluation of these preselected candidates across 40 independent A. githago cDNA libraries spanning diverse developmental stages and cultivation regimes. This application-oriented work builds on prior research to provide a practical framework for selecting appropriate reference genes for A. githago grown either in soil or in vitro.

A paragraph describing our methodological reasoning has been added to the Introduction section of the manuscript.

In detail, the reasons for selecting the seven specific potential reference genes from A. githago for stability evaluation can be explained as follows:

• Severely limited transcriptomic resources for A. githago

Although the A. githago genome has recently become available, publicly accessible RNA-seq data for this species are extremely scarce. At present, only one pooled RNA-seq dataset is available in NCBI/SRA, representing a mixture of multiple mature organs, which precludes organ-specific or condition-specific assessment of transcript abundance. A second dataset covers only a very young flower (and this one was used to retrieve the orthologous sequences for primer design).

• Lack of quantitative expression information suitable for stability screening

The available RNA-seq datasets do not provide enough amounts of comprehensive, comparable expression metrics (e.g., FPKM/TPM values across organs and conditions) required to screen large numbers of candidate genes for expression stability. Consequently, a transcriptome-wide search for novel reference genes was not technically feasible.

• Availability of annotated sequences for established housekeeping genes

Well-characterized housekeeping genes could be reliably retrieved from the annotated transcriptome published by Weise et al. (2020). This allowed us to confidently design primers and perform qRT-PCR-based stability assessments using validated gene models.

• Focus on applicability rather than discovery

Rather than conducting a broad discovery study aimed at identifying novel universal reference genes—an approach typically feasible only in model species with extensive transcriptomic resources—the study was designed as a targeted, application-oriented analysis. We have therefore selected seven previously validated housekeeping genes, informed by existing knowledge from Caryophyllaceae species.

• Strength of experimental validation across extensive biological material

The primary strength of the study lies in the rigorous experimental evaluation of these pre-selected candidate genes across 40 independent A. githago cDNA libraries encompassing multiple developmental stages and diverse cultivation conditions (soil-grown and in vitro). This comprehensive validation provides practical guidance for qRT-PCR normalization in this non-model species.

• Preservation of the study’s original objective

Expanding the analysis to include uncharacterized genes would have shifted the study toward transcriptome-wide gene discovery, fundamentally changing its scope and aims. The authors therefore deliberately limited the analysis to established candidates to maintain a clear, focused objective.

In summary, the selection of seven potential reference genes was driven by data availability constraints, feasibility considerations, and a deliberate focus on practical applicability, resulting in a robust and targeted evaluation of reference gene suitability for qRT-PCR analyses in A. githago.

Comment 2: In the experimental design, is it possible that gene expression could be influenced by the circadian rhythm or the annual cycle?

Response 2: Thank you for this insightful comment – indeed, the circadian rhythms might have influenced the expression of some of the genes, and this has not been studied. We have mentioned it in the revised discussion. However, the selection of several genes as reference should minimize the influence of such fluctuations. Moreover, in our experiment, the plant samples from the greenhouse were collected during morning hours (9-11) and it has been mentioned in the revised Methods section. As for the in vitro plant material, it is usually considered less influenced by external factors due to highly controlled environment. However, the light/darkness cycle has been applied and it could, of course, have some impact. Motivated by the Reviewer’s comment, we are planning to include such experiments in our future research. Within the current project, we aim at genes related to phenylpropanoid and terpenoid pathways, as well as RIP expression, which are not typically regarded as diurnally regulated but rather stress-related.

Regarding the annual cycle, as the plant is an annual (completes the life cycle within a single season), and we have studied ontogenetic stages as well as plants in tissue cultures, this influence has been accounted to.

Comment 3: In the methods part, it is possible to provide additional explanations regarding the basis for the cDNA dilution factor (such as the range of template concentrations), and mention the specific weight calculation method of RefFinder.

Response 3: Thank you for this suggestion. Indeed, as described in the methods section, each cDNA was synthesized from 5 µg of total RNA in a total volume of 40 µL – and this parameter represents the only constant value/starting point across all samples. Subsequently, all samples were treated the same way, including the dilution of each cDNA five times with nuclease-free water to maintain the same range of template concentrations across all samples. The paragraph 4.2 has been amended with this information accordingly.

In RefFinder, the “specific weight calculation method” refers to how the tool combines the results from several reference-gene stability algorithms into one final ranking. RefFinder integrates four commonly used methods: ΔCt, geNorm, NormFinder, BestKeeper. By assigning a weight (ranking value), each of these methods produces its own ranking of candidate reference genes based on different statistical principles. It is, however, true that the exact weighting scheme is not disclosed, making its internal processes opaque. This lack of transparency it is often called a “black box” issue. We have described in detail the limitations of RefFinder in our previous publication on Reynoutria [19] and we refer to it in this work, in the Discussion section.

The paragraph 4.5 in the Materials and Methods section has been amended accordingly.

Comment 4: The results section can include brief explanatory text to highlight the key findings, instead of relying solely on the charts.

Response 4: Wherever it was missing, we have summarized sections of the Results with a sentence highlighting key findings of a particular section.

Comment 5: Table 2: The content of this table is comprehensive, but it is quite lengthy. Consider splitting it or simplifying the presentation method.

Response 5: Following this recommendation, we have removed the ranking values from each subsection of the Table 2, which makes the table considerably shorter. The comprehensive ranking with values included is provided as the Supplementary Table S4.

figures: Optimize image details (axis labels, font size).

Figures have been optimised, wherever possible, including enlargement of font sizes and improving general clarity.

Comment 6: The discussion section can compare the research results of more closely related species to deepen the ecological or evolutionary perspectives.

Response 6: Thank you for this insightful suggestion. We have amended the Discussion with a paragraph comparing our results with studies published on more distant relatives, representing various families of the Caryophyllales order, such as Amaranthaceae, e.g. pseudocereals quinoa, amaranth and several halophytes – Halostachys, Salicornia, Salsola, Suaeda.

 

these responses are also included in the attached PDF file.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I appreciate the authors' efforts in revising the manuscript. However, not all my comments have been appropriately addressed or explained, and additional revisions are needed before the manuscript can be accepted for publication.

Major comment:

1. Sequence Information for Candidate Reference Genes

My previous comment: The authors should provide detailed sequence information for all candidate reference genes as supplementary file. This should include gene/transcript IDs (linked to the published genome assembly), genomic DNA sequences, cDNA sequences, protein sequences, and annotated features such as start/stop codons, exon/introns, and primer binding sites. Such information is critical for reproducibility, as researchers ofter struggle to cross-reference gene names with updated genome annotations.

New comment: (1) The authors have provided some additional information but not enough to fully address the request. Their explanation, that "genomic DNA sequences, exon-intron structures, predicted coding sequences, and protein sequences from the dcAgrGith1.1 assembly originate from a recently published genome resource and are not generated by the authors", is not convincing. Since the genome resource is already publicly available, the authors could easily integrate these sequences into the supplementary materials with proper citation. Readers should not be forced to retrieve this information independently, as errors may arise due to their lack of familiarity with the genome assembly compared to the authors.

(2) New concerns have emerged regarding Table S7 (Agrostemma githago contigs with BLAST score parameters and nucleotide sequences). Most reference genes align to multiple transcripts in the dcAgrGith1.1 genome assembly with very high identity (close to 100%). If this indicates multiple gene copies, these genes are unsuitable for use as internal controls in qRT-PCR, as researchers cannot distinguish which copy the primers target. Additionally, the header "According to the RNA-Seq data by Weise et al. 2020" should be placed above the columns "Isocontig/Isogroup/Number of contigs/Length" rather than above "Top blastx Hit/BLAST max score/%identity". For the sake of clarity, traceability, and reproducibility, the authors must address these issues thoroughly.

2. Agarose Gel Electrophoresis Results

My previous comment: The authors could supplement the melt curve analysis (Figure 1) with agarose gel electrophoresis results. Gel data would visually confirm the presence of a single amplicon of the expected size, providing more robust evidence of primer efficiency and specificity.

New comment: While the authors have added agarose gel results, new issues have been identified. As noted in the figure legend, amplicon sizes were estimated against the Perfect 100 bp DNA Ladder (EURx, Gdansk, Poland), which typically contains 13 bands (100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500 bp). However, only one marker band (500 bp) is visible in the gel figure, which is an obvious error. The authors should repeat the agarose gel electrophoresis using appropriate markers with clear, complete ladders. Additionally, the faint/blurred gel bands for β-tubulin, TEF1, and BAS suggest low expression, poor amplification efficiency, or insufficient specificity, further supporting that these genes are unsuitable as reference genes.

Minor comments:

1. Figure Legends in Figure 7 (now Figure 8)

My previous comment: Figure 7 uses figure legends to identify samples, forcing readers to cross-them with data points. To improve readability, I suggest placing sample identifiers directly on the x-aixs.

New comment: The authors' explanation, that "this particular figure has been produced by Maestro software (dedicated to CFX data analysis) and Maestro does not allow for such modifications", is unsatisfactory. Software should not be a limitation: multiple tools (R, GraphPad Prism, Origin, Excel) can generate this type of figure using raw data exported from Maestro.

2. Figures 3, 4, and 6 use only one color. For visual consistency and clarity, the color scheme should be standardized across these figures.
3. In Figure 8, "K, control" should be revised to "CK, control" (a more standard abbreviation in plant science).
4. Table 1: The first column would be clearer as "Gene name" (without adding "F/R" to entries). The second column could be split into "Forward (F):" and "Reverse (R):" to separate primer sequences.
5. Table 2: "Treatment and growth conditions" should be a separate column, distinct from "Method".
6. Several grammar and typographical errors remain and should be corrected thoroughly:

1. Line 477. "houskeeping" -> "housekeeping"
2. Line 512. "regardless from the plant organ" -> "regardless of the plant organ"
3. Line 642. "commerciall developmental y available soil"? "organic mater"?

Author Response

Reviewer1 – Review round 2

Dear Reviewer,

We would like to thank you for giving us a chance to revise and resubmit our manuscript. We do appreciate the time and effort spent on insightful evaluation of the submission that allowed us to improve our work. We did our best to suitably address all issues raised by performing appropriate changes into the manuscript and provided extensive explanations in the correspondence, where requested. We now submit a revised manuscript to be re-considered for publication in IJMS. All changes in the revised manuscript have been indicated by using tracked changes. Revised Figures 1, 4, 6, 8 and revised Supplementary table S7 are attached. Below, please find all issues raised outlined with our answers submitted in red. 

Major comment:

1. Sequence Information for Candidate Reference Genes

My previous comment: The authors should provide detailed sequence information for all candidate reference genes as supplementary file. This should include gene/transcript IDs (linked to the published genome assembly), genomic DNA sequences, cDNA sequences, protein sequences, and annotated features such as start/stop codons, exon/introns, and primer binding sites. Such information is critical for reproducibility, as researchers ofter struggle to cross-reference gene names with updated genome annotations.

New comment 1: The authors have provided some additional information but not enough to fully address the request. Their explanation, that "genomic DNA sequences, exon-intron structures, predicted coding sequences, and protein sequences from the dcAgrGith1.1 assembly originate from a recently published genome resource and are not generated by the authors", is not convincing. Since the genome resource is already publicly available, the authors could easily integrate these sequences into the supplementary materials with proper citation. Readers should not be forced to retrieve this information independently, as errors may arise due to their lack of familiarity with the genome assembly compared to the authors.

Response 1: We thank the reviewer for the detailed comments and for highlighting the importance of traceability and reproducibility in reference gene selection.

Regarding comment 1, we clarify that genomic DNA sequences, exon–intron structures, predicted coding sequences, and protein sequences for Agrostemma githago are derived from the publicly available dcAgrGith1.1 genome assembly and are not generated de novo in this study. In the revised version, Supplementary Table S7 has been updated to include stable gene identifiers together with direct links to the corresponding genome annotation pages, enabling immediate access to genomic, cDNA, and protein sequences (all these sequences can be easily downloaded using the ‘Download’ button in the right-hand menu). We deliberately did not duplicate these sequences as static FASTA files in the supplementary materials, as genome annotations are actively evolving and static copies may quickly become outdated. Direct linking to the primary genome resource ensures traceability while allowing readers to retrieve the most current annotations.

New comment 2: New concerns have emerged regarding Table S7 (Agrostemma githago contigs with BLAST score parameters and nucleotide sequences). Most reference genes align to multiple transcripts in the dcAgrGith1.1 genome assembly with very high identity (close to 100%). If this indicates multiple gene copies, these genes are unsuitable for use as internal controls in qRT-PCR, as researchers cannot distinguish which copy the primers target. Additionally, the header "According to the RNA-Seq data by Weise et al. 2020" should be placed above the columns "Isocontig/Isogroup/Number of contigs/Length" rather than above "Top blastx Hit/BLAST max score/%identity". For the sake of clarity, traceability, and reproducibility, the authors must address these issues thoroughly.

Response 2: Thank you for this important remark. We acknowledge that several candidate reference genes align with multiple highly similar transcripts in the dcAgrGith1.1 genome assembly, often with near-100% sequence identity. Importantly, a substantial fraction of these alignments corresponds to multiple transcript isoforms annotated for the same gene locus, rather than to independent gene copies. Such patterns reflect alternative splicing and transcript model redundancy in the current genome annotation and should not be interpreted as evidence for multiple functional gene copies per se. Yet, among the candidate reference genes, transcripts corresponding to ACT and βTUB may indeed originate from multiple gene loci. Based on its length (5,764 bp), the ENSOXJT00000045558.1 transcript likely represents a concatenation of several transcripts, one of which is homologous to βTUB; therefore, both its gene- and transcript-level annotations are not easy to confirm.

However, the presence of multiple transcripts originating from a single gene, or membership of a multigene family, does not automatically disqualify a candidate from use as an internal control in qRT-PCR. Classical housekeeping genes such as actin are well known to belong to multigene families and may include paralogs or pseudogenes. As discussed by Thellin et al. (1999), many housekeeping genes "belong to multigene families or present pseudogenes," and their use therefore requires careful evaluation rather than blind acceptance. Similarly, Vandesompele et al. (2002) demonstrated that no universal reference gene exists and that accurate normalization depends primarily on expression stability rather than genomic structure. Also, when genome structure information is missing for a given organism (a common situation for non-model species) reference gene selection must rely on experimental and methodological safeguards rather than genomic certainty. In such cases, genes representing multiple copies are not automatically disqualified, but they require more stringent validation.

These limitations and their implications for data interpretation are now explicitly discussed in the manuscript (please find the new paragraphs added in the Discussion section).

 

In the revised analysis, Supplementary Table S7 has been expanded to clearly distinguish gene-level and transcript-level annotations, and BLAST-based transcript assignment was re-evaluated using a ranking strategy prioritizing alignment score, alignment length, and sequence identity. This approach reduces the influence of short, perfect matches arising from conserved regions shared among transcript isoforms or related genes. The observed multi-transcript alignments therefore reflect the complexity of the current genome annotation rather than ambiguous primer targeting.

Finally, the placement of table headers in Supplementary Table S7 has been corrected to improve clarity and traceability, as suggested by the reviewer.

2. Agarose Gel Electrophoresis Results

My previous comment: The authors could supplement the melt curve analysis (Figure 1) with agarose gel electrophoresis results. Gel data would visually confirm the presence of a single amplicon of the expected size, providing more robust evidence of primer efficiency and specificity.

New comment 3: While the authors have added agarose gel results, new issues have been identified. As noted in the figure legend, amplicon sizes were estimated against the Perfect 100 bp DNA Ladder (EURx, Gdansk, Poland), which typically contains 13 bands (100, 200, 300, 400, 500, 600, 700, 800, 900, 1000, 1500, 2000, 2500 bp). However, only one marker band (500 bp) is visible in the gel figure, which is an obvious error. The authors should repeat the agarose gel electrophoresis using appropriate markers with clear, complete ladders. Additionally, the faint/blurred gel bands for β-tubulin, TEF1, and BAS suggest low expression, poor amplification efficiency, or insufficient specificity, further supporting that these genes are unsuitable as reference genes.

Response 3: As suggested, we repeated the agarose gel electrophoresis. To obtain comparable band intensities, the amount of template used in each PCR reaction was adjusted where possible. The lower intensity of the bands corresponding to the βTUB and BAS amplicons reflects the naturally lower expression levels of these genes in the analyzed cDNA sample, which is a key aspect of this study.
In this electrophoresis, amplicon sizes were estimated using two molecular weight markers: Perfect™ 100 bp DNA Ladder (EURx, Gdańsk, Poland) (left) and Perfect™ 100–1000 bp DNA Ladder (EURx, Gdańsk, Poland) (right). All bands in the 100–500 bp range are clearly visible and are now labelled, along with the 1000 bp and 2500 bp marker bands.

Minor comments:

1. Figure Legends in Figure 7 (now Figure 8)

My previous comment: Figure 7 uses figure legends to identify samples, forcing readers to cross-them with data points. To improve readability, I suggest placing sample identifiers directly on the x-aixs.

New comment: The authors' explanation, that "this particular figure has been produced by Maestro software (dedicated to CFX data analysis) and Maestro does not allow for such modifications", is unsatisfactory. Software should not be a limitation: multiple tools (R, GraphPad Prism, Origin, Excel) can generate this type of figure using raw data exported from Maestro.

Response: Figure 7 has been modified. In the revised version, sample identifiers have been placed directly on the x-axis.

2. Figures 3, 4, and 6 use only one color. For visual consistency and clarity, the color scheme should be standardized across these figures.

The color scheme has been unified across Figures 3, 4 and 6.

3. In Figure 8, "K, control" should be revised to "CK, control" (a more standard abbreviation in plant science).

It has been modified to “CK”, in both the Figure 8 and Fig. 8 legend.

4. Table 1: The first column would be clearer as "Gene name" (without adding "F/R" to entries). The second column could be split into "Forward (F):" and "Reverse (R):" to separate primer sequences.

Table 1 has been modified according to the comments above.

5. Table 2: "Treatment and growth conditions" should be a separate column, distinct from "Method".

Table 2 has been modified accordingly.

6. Several grammar and typographical errors remain and should be corrected thoroughly:

1. Line 477. "houskeeping" -> "housekeeping"
2. Line 512. "regardless from the plant organ" -> "regardless of the plant organ"
3. Line 642. "commerciall developmental y available soil"? "organic mater"?

All grammar and typographical errors have been corrected. The entire manuscript has been carefully revised for grammatical, stylistic, and typographical correctness.

 

Round 3

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed most of the concerns raised previously. Below are minor revisions proposed to enhance the consistency and accuracy of the manuscript. I confirm that the editor may make a final decision on the manuscript without further consultation with me.

1. Unify gene name notations (e.g. "TIF5A1-1", "TIF5A1-2", "TIF5A1", and "TIF5A2") throughout the manuscript, especially in Figures 4.
2. Standardize the abbreviation "EF1-α" to "EF1α" globally.
3. Correct the spelling error "developnemtal" (Line 674) to "developmental".

Author Response

Comment 1. The authors have addressed most of the concerns raised previously. Below are minor revisions proposed to enhance the consistency and accuracy of the manuscript. I confirm that the editor may make a final decision on the manuscript without further consultation with me.

Response: On behalf of all co-authors we appreciate  the Reviewer's effort. Your feedback has been invaluable in improving the clarity, readability, and overall quality of our work. In this revision round, we have carefully corrected all typographical mistakes listed below (and others) to enhance the consistency and accuracy of the manuscript.

Comment 2. Unify gene name notations (e.g. "TIF5A1-1", "TIF5A1-2", "TIF5A1", and "TIF5A2") throughout the manuscript, especially in Figures 4.

Response 2. We have unified the gene notation convention, including the figure 4. (new figure attached in the revision, separately)

Comment 3. Standardize the abbreviation "EF1-α" to "EF1α" globally.

Response 3. We have changed the spelling as suggested.

Comment 4. Correct the spelling error "developnemtal" (Line 674) to "developmental".

Response 4. This and any other spelling mistakes have been corrected.