Review Reports - RNA–Protein Dynamics of the Segmentation Clock and Its Implications in Developmental Disorders

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript entitled “Integrated RNA and protein dynamics of the segmentation clock: post-transcriptional control and implications for developmental disorders” addresses an important and rapidly evolving area of developmental biology. Its principal strength is the attempt to integrate transcriptional oscillations, RNA processing and turnover, translation, protein stability, intercellular synchronization, boundary formation, and human vertebral disorders within a single conceptual framework. The organization around initiation, maintenance, synchronization, boundary formation, and termination of the segmentation clock is potentially useful.

However, the manuscript in its current form is excessively broad and does not consistently distinguish direct evidence for RNA-mediated regulation from indirect associations, protein-level mechanisms, general signaling biology, and speculative epitranscriptomic hypotheses.

Several statements are insufficiently supported by the cited references, and there are important problems involving species-specific terminology, gene nomenclature, Notch/DLL biology, clinical genetics, and the interpretation of disease-associated variants. Substantial restructuring and expert re-evaluation of the human disease section are required before the manuscript can be considered for publication.

Major comments

The central scope and novelty of the review need to be defined more precisely.

The manuscript presents the segmentation clock as an “RNA-centered” or “RNA disease” system, yet large portions of the review concern transcription factors, signaling gradients, metabolism, membrane potential, protein degradation, cell proliferation, and tissue mechanics. These are relevant to somitogenesis, but they are not necessarily RNA-mediated mechanisms. In several places, any perturbation affecting a gene or its protein product is subsequently described as evidence for RNA dysregulation.

The authors should decide whether the review is primarily about:

direct post-transcriptional regulation of the segmentation clock;
integrated RNA–protein kinetics within the clock; or
the broader multilayer regulation of somitogenesis.

If the manuscript remains focused on post-transcriptional and epitranscriptomic regulation, the general developmental and signaling sections should be considerably shortened. Alternatively, the title, abstract, and conclusions should be broadened and should no longer position RNA regulation as the dominant mechanism. The authors should also explain clearly what this review adds beyond the recent reviews cited as references 14–16 and, particularly, the post-transcriptional review cited as reference 31.

The manuscript requires an explicit hierarchy of evidence.

Direct mechanistic observations, indirect developmental associations, findings from non-segmentation tissues, computational predictions, and hypotheses are frequently presented with similar certainty. This is especially problematic for RNA modifications, ribosome heterogeneity, metabolic regulation, and human disease.

I recommend that the authors classify the evidence as:

directly demonstrated in the segmentation clock;
demonstrated in somitogenesis or axial development but not in clock oscillations;
inferred from another tissue or biological system;
hypothetical and requiring experimental validation.

A summary table should include the molecular mechanism, molecule or gene, species/model, experimental approach, measured clock phenotype, and level of evidence. Preprints should be clearly identified and interpreted cautiously. Several central statements currently depend on preprints, including references 45, 46, 61, 92, 99, 125, 137, and 145. Reference 23 should also be updated to its recently published Cell Systems version where appropriate.

Even for a narrative review, the authors should provide a short description of the literature search, search date, databases, selection criteria, and policy regarding preprints.

Species-specific findings and orthologous genes are repeatedly conflated.

The manuscript moves rapidly between human, mouse, chick, zebrafish, Xenopus, goldfish, and eel data without consistently identifying which conclusions are species-specific. Mouse Lfng/Hes7, zebrafish her1/her7/deltaC/deltaD, chick c-hairy1/LFNG, and human HES7/LFNG should not be presented as interchangeable components.

A particularly important example occurs in Section 2.2, where the text suggests that “in mouse and zebrafish embryos” up to eight early waves of Lfng expression persist after Notch disruption. Reference 52 investigates Lfng dynamics in mouse embryos; it draws a comparison with earlier zebrafish findings involving other clock components, rather than demonstrating zebrafish Lfng oscillations. The corresponding developmental study should be represented accurately.

Similarly, Figure 2 combines HES7 and HER1/7 within a single generic Notch pathway without species annotation. The authors should introduce a comparative table identifying the principal clock genes and experimental evidence in each organism and should consistently qualify species throughout the text.

The description of LFNG and Notch/DLL biology requires correction and clarification.

The statement that LFNG glycosylation simply reduces Notch responsiveness to Delta-like ligands, particularly DLL1, is overly simplified and does not adequately represent the context-dependent effects of Fringe proteins. Reference 53 specifically describes Lfng-mediated trans-repression involving Dll1 in neighboring cells, rather than a universal reduction in receptor responsiveness.

The manuscript also states that Notch signaling “initiates and synchronizes” oscillations in Figure 2, whereas the text emphasizes that the initial oscillations can occur independently of canonical Notch signaling. This apparent contradiction should be resolved by separating oscillator initiation from subsequent coupling, stabilization, and synchronization.

DLL1, DLL3, DLL4, DeltaC, and DeltaD must also be discussed separately. DLL3 is an atypical Notch ligand with predominantly inhibitory intracellular/cis functions and should not be implicitly represented by the canonical signal-sending ligand depicted in Figure 2. Combining all these genes under the generic label “Dll” obscures major biological and clinical differences.

The human disease section and Table 3 require complete re-curation by clinical genetics standards.

Table 3 is currently not sufficiently reliable for publication. It combines variants from different human genes under generic labels such as “Dll,” “Mesp,” “Ripply,” “Fgfr,” and “Meox,” although the corresponding genes, mechanisms, inheritance patterns, and diseases differ substantially.

The revised table should provide:

the exact HGNC-approved human gene symbol;
the reference transcript;
complete HGVS c. and p. nomenclature;
zygosity and inheritance pattern;
variant classification and supporting evidence;
direct RNA consequence, when demonstrated;
protein or signaling consequence, when this is the actual mechanism;
associated phenotype and strength of the gene–disease relationship.

The title “Pathogenic variants” is not justified for every included entry. Some rows are described as “likely” disease-associated, functionally hypomorphic, or “not specified,” while others originate from individual case reports. These should not all be presented with equivalent pathogenic certainty.

The authors should also explain the TBX6 compound-inheritance model, including the interaction of a loss-of-function allele with a common hypomorphic haplotype, rather than describing TBX6-associated congenital scoliosis mainly as generic RNA dosage dysregulation. Likewise, the reported TBX6–MRKH association requires a more cautious discussion.

The Adams–Oliver syndrome description should be corrected and should include the characteristic terminal transverse limb defects. Furthermore, DLL4-associated Adams–Oliver syndrome should not automatically be interpreted as a segmentation-clock disorder without direct supporting evidence.

Several strong mechanistic claims are not supported by the cited references.

Examples include:

The statement that HIF1α “interacts with” Hes7 and Lfng transcripts is not demonstrated by reference 92. That preprint reports abnormal Hes7 and Lfng expression following HIF1α loss, not a physical interaction with these RNAs. Reference 98 is a general study of transcription factor–RNA interactions and does not establish this specific mechanism. Reference 92 also remains explicitly identified as a non-peer-reviewed preprint.
Reference 129 is a prenatal case report of a MESP2 variant. It does not, by itself, demonstrate that abnormal MESP transcript processing or delayed termination of MESP protein expression causes the reported phenotype.
RPL38 and RPL10A data support roles in selective translation and mesodermal development, but the manuscript acknowledges that neither has been directly demonstrated to regulate segmentation-clock oscillations. Their placement in the disease pathway of Figure 2 therefore overstates the evidence.
References 27–30 concern stress granules, neuronal RNA granules, Alzheimer-related cellular toxicity, and triple-negative breast cancer. Their relevance to segmentation-clock disorders is unclear and should either be explained or these references should be removed.

The authors should systematically verify that every cited article supports the specific sentence in which it is cited.

The epitranscriptomic component is substantially underdeveloped relative to its prominence in the title and abstract.

The discussion of m6A, pseudouridylation, RNA editing, and tRNA modifications is largely hypothetical. The tissue-specific tRNA modification study cited as reference 123 provides a useful general principle, but it does not directly demonstrate regulation of the segmentation clock or PSM biology.

The authors should either substantially reduce the prominence of epitranscriptomics in the title and abstract or explicitly present this topic as a hypothesis-driven future research area. A separate “Evidence gaps and testable hypotheses” section would be more appropriate than incorporating these mechanisms into established causal pathways.

This section could identify specific experimental approaches, such as transcript half-life measurements, modification-specific RNA profiling, ribosome profiling, live RNA imaging, and CRISPR perturbation of RNA modifiers in human somitoids. Section 3.3 should also be reorganized because its first part primarily discusses transcriptional bursting and noise rather than post-transcriptional or epitranscriptomic regulation.

The manuscript is excessively long and contains substantial conceptual repetition.

FGF/Wnt gradients, Hes/Her feedback, Notch synchronization, transcript kinetics, and disease implications are introduced repeatedly across Sections 2–7. Several subsections end with almost identical statements that precise RNA regulation is essential for development and that its disturbance can cause vertebral abnormalities.

The early embryology and extended ZEB1/ZEB2 discussion in Section 2.1 are disproportionately detailed and only indirectly connected to the principal topic. This material could be markedly shortened. Section 5, entitled “Stopping the clock: scaling, RNA decay and transition to differentiation,” primarily discusses FGF-mediated scaling and tailbud signaling, with very limited direct evidence concerning RNA decay. The title and content should therefore be aligned.

A more focused structure could separately address:

RNA kinetics of the core oscillator;
intercellular coupling and species-specific differences;
translation, proteostasis, and metabolic modulation;
conversion of oscillations into boundaries and clock termination;
human genetics and disease;
evidence gaps and future experiments.
Figures and tables need conceptual and technical revision.

Figures 1 and 2 are visually clear but partly redundant and portray speculative and established mechanisms with the same graphical certainty. Solid versus dashed arrows, or another explicit evidence code, should distinguish demonstrated mechanisms from hypotheses. Species should be indicated wherever specific genes are shown.

Table 1 combines in vivo embryos with stem-cell-derived models and includes an “average size of somite” column without comparable units or definitions. In vivo and in vitro systems should be separated.

Table 2 is entitled “Summary of RNAs role in the segmentation clock,” which should be grammatically corrected. More importantly, the table does not include several RNA modification mechanisms emphasized elsewhere and again combines direct and indirect evidence.

Table 3 should be reconstructed as discussed above. The authors should also disclose the source and licensing of all graphical elements and any software used to create the figures.

The ethics discussion should be updated and made more precise.

The statement that current ISSCR guidance restricts human embryo-model research according to a general 14-day limit is too broad. Current guidance distinguishes human embryos from different categories of stem-cell-based embryo models and emphasizes specialized oversight, scientific justification, jurisdiction-specific regulation, and the minimum duration required for the experiment. The authors should revise this section according to the current ISSCR guidance and avoid applying the embryo-culture rule indiscriminately to gastruloids, somitoids, segmentoids, and other model systems.

Minor comments

“Her” is incorrectly expanded as “human epidermal growth factor receptor.” In this context, zebrafish her1/her7 are hairy-related transcriptional repressors. This error appears in the abstract, tables, and abbreviation lists and must be corrected throughout.
Human gene symbols should be uppercase and italicized, mouse genes should use initial capitalization, and zebrafish genes should use lowercase italics. Protein names should not be italicized.
Correct typographical and language errors, including “Neuosurgical,” “traslation,” “Table1,” “Figure2,” “T-box transcription factor6,” and the awkward phrase “Summary of RNAs role.” The term “derealization” in the ZEB2 paragraph also appears inappropriate and should be checked against the cited source.
The abstract sentence stating that disruptions in RNA mechanisms “impair protein-level and signaling abnormalities” is grammatically and conceptually incorrect.
The term “RNA disease” is used too broadly and should either be explicitly defined or replaced with more specific descriptions of the molecular defect.
Abbreviations such as SCD, SCDO, SDV, CS, and KFS should be standardized. The manuscript should distinguish a clinical phenotype from a specific molecularly defined disorder.
Reference 49 and reference 164 appear to describe the same TBX6/MRKH study and should be deduplicated.
The reference list contains numerous formatting or transcription errors, including author-name errors, malformed journal titles, inconsistent DOI formatting, and apparent typographical errors in several DOIs. A complete reference audit is necessary.
The manuscript template still displays “Int. J. Mol. Sci. 2025,” although the bibliography contains multiple 2026 publications. The document metadata should be updated.
The final statement concerning future therapeutic strategies is too speculative for congenital structural abnormalities established during embryogenesis. Diagnostic and mechanistic implications are plausible, but therapeutic implications should be qualified.

Author Response

Reviewer1

Major comments

The central scope and novelty of the review need to be defined more precisely.

The authors should decide whether the review is primarily about:

direct post-transcriptional regulation of the segmentation clock;
integrated RNA–protein kinetics within the clock; or
the broader multilayer regulation of somitogenesis.

Response: We thank the reviewer for identifying this important conceptual issue. We agree that the original manuscript did not consistently distinguish RNA-mediated mechanisms from broader protein-, signaling-, metabolic-, and tissue-level mechanisms. We have therefore revised the central premise of the review and now frame the segmentation clock as an integrated RNA–protein–signaling oscillator. The revised manuscript emphasizes that transcriptional feedback establishes the core oscillatory architecture, while RNA processing, transcript turnover, translation, and protein degradation contribute to the kinetic delays, buffering, and temporal precision of the system. Notch-mediated intercellular coupling and developmental signaling pathways are discussed as complementary layers rather than being classified as intrinsically RNA-mediated mechanisms. We have also adjusted the broad term “RNA disease”. The title in line 2, the Introduction in lines 97-100, Sections 8 and 9 have been revised accordingly. We further clarify the contribution of this review by integrating RNA and protein kinetics across oscillator initiation, maintenance, synchronization, boundary formation, and termination, while explicitly separating established mechanisms from indirect evidence and testable hypotheses.

The manuscript requires an explicit hierarchy of evidence.

I recommend that the authors classify the evidence as:

directly demonstrated in the segmentation clock;
demonstrated in somitogenesis or axial development but not in clock oscillations;
inferred from another tissue or biological system;
hypothetical and requiring experimental validation.

Even for a narrative review, the authors should provide a short description of the literature search, search date, databases, selection criteria, and policy regarding preprints.

Response: We thank the reviewer for this important recommendation. We agree that the previous version did not consistently distinguish direct segmentation-clock evidence from developmental association, evidence extrapolated from other biological systems, computational inference, and hypothesis-generating mechanisms. We have therefore substantially revised the manuscript to introduce and consistently apply an explicit hierarchy of evidence.

We introduced a four-level evidence hierarchy in the Methodology in lines 187-202 and Table2. The revised manuscript now defines Level I as mechanisms directly demonstrated in the segmentation clock through experimental measurement or perturbation of oscillation period, amplitude, phase, synchronization, propagation, or immediate clock-to-boundary conversion; Level II as mechanisms demonstrated in somitogenesis or axial development without direct measurement of clock oscillations; Level III as mechanisms inferred from another tissue, developmental context, or biological system; and Level IV as hypotheses, unvalidated predictions, or mechanisms requiring direct experimental testing. We further clarify that experimentally constrained mathematical models are interpreted together with the underlying empirical evidence rather than being considered direct evidence independently.

We added a dedicated evidence-summary table. A new Table 2, “Evidence hierarchy for representative mechanisms discussed in this review,” now reports, for each representative mechanism, the molecular mechanism, relevant molecule/gene, species and model, experimental approach, measured clock or developmental phenotype, assigned evidence level, publication status, and supporting reference. This table includes examples spanning transcriptional bursting, Notch coupling, NICD turnover, metabolic regulation, cell-cycle coupling, clock-to-boundary conversion, ribosome-mediated translation, human congenital disease, tRNA modification, epitranscriptomic mechanisms, and late-axis termination.

We also strengthened the literature-search description. The added methodology section in lines 144-186 now states that the structured search included terms covering the segmentation clock, somitogenesis, PSM biology, clock genes, RNA processing/decay, translation, ribosome biology, metabolism, RNA modifications, epitranscriptomics, and vertebral segmentation disorders. We also specify inclusion and exclusion criteria, prioritization of primary peer-reviewed research, reference-list screening, and our policy for incorporating preprints.

We re-checked all cited preprints for subsequent peer-reviewed publication as of September 18, 2026. Previous reference 23 was replaced with its published version; References 42, 62-65, 69 and 79 remain available only as preprints and are explicitly labeled as such.

The clinically curated variants in the revised disease table establish human gene–disease relationships and vertebral segmentation phenotypes, but patient-level clock oscillations are generally not measured. We therefore classify these associations as Level II with respect to segmentation-clock causality unless the relevant perturbation has independently been shown to alter clock dynamics in an experimental model. Clinical pathogenicity and mechanistic clock evidence are now treated as distinct dimensions.

Species-specific findings and orthologous genes are repeatedly conflated.

Response: We thank the reviewer for this important comment. We have systematically revised the manuscript to identify the species and experimental system associated with each major mechanistic conclusion. We have revised the Abstract in lines 24-40, Introduction in lines 60-68, 79-88, 114-124, figures 1 and 2, tables, and mechanistic sections to distinguish the principal segmentation-clock components used in each organism. In particular, we now explicitly distinguish mouse Hes7/Lfng/Dll1, zebrafish her1/her7/deltaC/deltaD, chick c-hairy1/LFNG, and human HES7/LFNG rather than presenting these genes as interchangeable components of a generic vertebrate clock. We have also added an explicit statement that cross-species comparisons are used to identify conserved regulatory principles, but that analogous or orthologous genes are not assumed to have identical timing, paralog usage, ligand dependence, or regulatory function.

We also agree with the reviewer that the previous statement regarding early Lfng waves in “mouse and zebrafish embryos” was inaccurate. The revised text now states that, in mouse embryos, initial Lfng waves persist following inhibition or disruption of canonical Notch signaling and also persist in Hes7-deficient embryos. Accordingly, the revised manuscript no longer uses Reference 52 as evidence for zebrafish lfng dynamics. Zebrafish oscillator initiation and synchronization are discussed separately using the her1/her7 and deltaC/deltaD literature.

In response to the reviewer’s suggestion, we have substantially reconstructed Table 1 as:

“Species-specific segmentation-clock genes and experimental evidence represented in this review.”

The revised table now identifies, for each model system

the organism and experimental context;
the principal genes or molecular components discussed;
the specific experimental evidence contributed by that organism;
the relevant segmentation-clock period or developmental timing, where established; and
the supporting references.

Human iPSC-derived PSM models, mouse embryos/PSM explants, zebrafish embryos, chick embryos, Xenopus, goldfish, corn snake, and Japanese eel are listed separately. Importantly, the table explicitly identifies when a species is being used only for downstream boundary formation, signaling, scaling, or termination rather than as direct evidence for the core oscillator.

We also agree that the previous Figure 2 incorrectly combined HES7 and HER1/7 within a single generic Notch pathway. Figure 2 has therefore been revised so that the central Notch-coupling panel is explicitly labeled as a mouse PSM example. The figure now depicts Dll1–NOTCH coupling with mouse Hes7 and Lfng, rather than combining mouse and zebrafish genes in the same pathway. The figure legend now explicitly states that zebrafish her1/her7 and deltaC/deltaD constitute a separate species-specific system and are discussed independently in the text. The human disease panel continues to use HGNC-approved uppercase human gene symbols and is explicitly distinguished from the mouse mechanistic panel.

Several sections have been rewritten to avoid the generic “Hes/Her” formulation where it could imply mechanistic equivalence. The revised manuscript now describes Hes7 as a principal mouse oscillatory repressor, whereas her1 and her7 are discussed as the corresponding zebrafish paralogous oscillator system. Similarly, mouse Dll1-dependent coupling is treated separately from zebrafish deltaC/deltaD-dependent coupling. The discussion of the zebrafish developmental “timer,” her1 live imaging, tbx6/ripply boundary conversion, and ERK/ripply regulation is now explicitly identified as zebrafish evidence rather than generalized automatically to mouse or human somitogenesis.

We have similarly clarified the role of other organisms used in the review. Xenopus studies are now described specifically as supporting downstream Tbx6/Ripply/Bowline boundary-conversion mechanisms; goldfish data are identified as Wnt/somitogenesis perturbation evidence rather than direct core-clock evidence; comparative corn snake studies are used for somite scaling rather than gene-level equivalence; and the Japanese eel study is presented specifically as a late-axis termination/prolonged-segmentation phenotype. Human segmentoid, somitoid, and related stem-cell-derived models are also distinguished from animal data. To maintain the same species-specific framework throughout the manuscript, Table 2 has also been revised. The previous combined oscillator entries have been separated into individual rows for mouse Hes7, mouse Lfng, mouse Dll1, zebrafish her1/her7, zebrafish deltaC/deltaD, and the relevant downstream species-specific modules.

The description of LFNG and Notch/DLL biology requires correction and clarification.

Response: We thank the reviewer for identifying these important mechanistic distinctions. We agree that the previous wording oversimplified LFNG function and did not sufficiently distinguish oscillator initiation from subsequent Notch-dependent coupling and synchronization. We have therefore substantially revised the relevant text, Figure 2, and Table 2.

We agree that describing LFNG glycosylation as simply reducing Notch responsiveness to Delta-like ligands was overly generalized. We have revised Section 3.2 in lines 266-273 to state explicitly that Fringe effects are context dependent and should not be interpreted as a universal reduction in ligand-induced Notch activation. In particular, we now describe the findings of Okubo et al. more precisely: in the mouse PSM, Lfng contributes to segmentation-clock synchronization through a trans-repression mechanism, in which Lfng in a signal-sending cell modulates Dll1 function and reduces Notch activity in the neighboring receiving cell. Thus, Lfng is now presented as a context-dependent regulator of Notch coupling rather than as a simple inhibitor of receptor responsiveness. We have also incorporated subsequent evidence showing that Lfng and Dll3 can interact through cis-regulatory mechanisms to modulate Dll1/Notch1 signal-sending activity, further emphasizing that Lfng function depends on the ligand and cellular context.

We agree that the previous Figure 2 wording stating that Notch signaling “initiates and synchronizes” oscillations was inconsistent with the manuscript text describing early Notch-independent oscillatory activity. We have therefore revised the manuscript throughout to distinguish these processes explicitly. The revised model proposes that initial oscillatory transcription can arise before canonical Notch coupling is fully engaged, whereas Notch signaling subsequently contributes to intercellular coupling, stabilization, phase alignment, and synchronization of neighboring oscillators. Accordingly, we now state that Notch signaling should not be regarded as a universal trigger of the first oscillatory cycle. Section 4.2 in lines 517-539 has also been retitled and revised to focus specifically on the later coupling and synchronization functions of Notch signaling. This distinction is consistent with experimental evidence that Notch signaling is critical for maintaining synchrony among neighboring PSM oscillators.

Furthermore, Figure 2 has been revised accordingly. The previous generic label “DLL1/4 ligand” has been replaced with “DLL1 ligand”, because the schematic is intended to illustrate the canonical mammalian trans-activating interaction relevant to the mouse PSM coupling literature. The Figure 2 legend now explicitly states that the panel illustrates intercellular coupling after oscillator onset, and that Notch signaling stabilizes and synchronizes oscillations rather than universally initiating the earliest oscillatory transcription.

We also agree that grouping these proteins under a generic “Dll” designation obscures important biological differences. We have therefore reorganized Section 4.2 in lines 517-570, now entitled:

“Synchronizing oscillations through species-specific Notch ligand dynamics.”

In the revised section:

DLL1/Dll1 is described as the canonical mammalian trans-activating ligand with direct evidence for intercellular coupling in the mouse PSM.
DLL3/Dll3 is described separately as an atypical, predominantly intracellular/cis Notch regulator that can cooperate with LFNG to modulate DLL1/NOTCH1 signaling rather than acting as a conventional trans ligand.
DLL4 is no longer treated as interchangeable with DLL1 and is not incorporated into the mechanistic segmentation-clock schematic in the absence of comparable direct evidence.
DeltaC and deltaD are discussed explicitly as zebrafish ligands with distinct temporal and functional roles rather than as direct equivalents of mammalian DLL proteins.

Their RNA/protein functions, developmental roles, and consequences of dysregulation are described independently. In addition, to avoid retaining the same ambiguity elsewhere in the manuscript, the generic “Notch–Delta” wording in Figure 1 and its legend has been revised to emphasize Notch-mediated coupling and to note that the relevant Delta-like ligand is species- and context-dependent.

The human disease section and Table 3 require complete re-curation by clinical genetics standards.

The revised table should provide:

the exact HGNC-approved human gene symbol;
the reference transcript;
complete HGVS c. and p. nomenclature;
zygosity and inheritance pattern;
variant classification and supporting evidence;
direct RNA consequence, when demonstrated;
protein or signaling consequence, when this is the actual mechanism;
associated phenotype and strength of the gene–disease relationship.

Response: We thank the reviewer for this important and detailed comment. We agree that the previous version of the human disease section 8, 9 and previous Table 3 did not adequately distinguish between genes, variants, inheritance models, molecular mechanisms, and levels of clinical-genetic evidence. We have therefore substantially re-curated this section and reconstructed Table 3 (current Table 4).

Exact gene nomenclature and variant annotation

We have replaced generic gene-family labels such as “Dll,” “Mesp,” “Ripply,” “Fgfr,” and “Meox” with the HGNC-approved human gene symbols and now provide the corresponding reference transcript and HGVS c. and p. nomenclature for each selected variant. The revised table is intentionally representative rather than exhaustive and is restricted to variants for which these annotations can be stated clearly.

Zygosity and inheritance

We have added zygosity and inheritance pattern for each variant. The revised table focuses on HES7, LFNG, DLL3, MESP2, and RIPPLY2, all of which have established associations with autosomal-recessive spondylocostal dysostosis or closely related vertebral segmentation phenotypes.

Variant classification and strength of evidence

We have removed the previous blanket presentation of all entries as “pathogenic variants.” The revised table reports the classification supported by the individual primary study and/or condition-specific record, together with the type of supporting evidence.

RNA-level versus protein/signaling consequences

We agree that the previous table sometimes conflated distinct mechanistic levels. The revised table now contains separate columns for direct RNA consequences and protein/signaling consequences. Importantly, an RNA consequence is only listed when experimentally demonstrated or explicitly supported by the cited study.

Gene–disease relationship and phenotype

The revised Table (current Table 4) now separately reports the associated clinical phenotype and the strength/context of the gene–disease relationship. We have focused the table on well-established vertebral segmentation disorders rather than combining clinically distinct conditions under broad developmental labels.

DLL4 and Adams–Oliver syndrome

We agree with the reviewer that DLL4-associated Adams–Oliver syndrome should not automatically be interpreted as a segmentation-clock disorder. We have therefore removed Adams–Oliver syndrome and the DLL4 entry from the human disease discussion and the revised Table (current Table 4).

TBX6-associated disease

We also agree that a rigorous treatment of TBX6-associated congenital scoliosis would require discussion of its distinctive dosage-sensitive and compound-inheritance model, including the interaction of a loss-of-function allele with a common hypomorphic haplotype. Because this mechanism would require substantial additional clinical-genetic discussion and is not central to the principal focus of the present review, we have removed TBX6-associated congenital scoliosis, and TBX6 variant entries from the human disease section 8, 9 and Table 3. This approach also avoids overinterpreting TBX6-associated phenotypes as generic RNA-dosage disorders. TBX6 is retained only where necessary to describe its established developmental role in PSM specification and somite-boundary formation.

Reconstruction of Table

The revised Table (current Table 4) has therefore been completely reconstructed and retitled as:“Selected human variants in genes associated with spondylocostal dysostosis and vertebral segmentation defects: clinically curated molecular evidence.”

The revised table now includes the exact gene/transcript and HGVS nomenclature, zygosity/inheritance, variant classification and supporting evidence, demonstrated RNA consequence, protein/signaling consequence, associated phenotype, gene–disease relationship, and supporting references.

Integration with the revised conceptual structure

Finally, we revised the surrounding human-disease text so that clinically established genetic mechanisms are clearly separated from hypothesis-driven RNA and epitranscriptomic mechanisms. This is consistent with the newly reorganized Section 5, “Evidence gaps and testable hypotheses: RNA-mediated post-transcriptional control and emerging epitranscriptomic mechanisms,” and avoids presenting emerging RNA mechanisms with the same evidentiary certainty as established Mendelian disease mechanisms.

Several strong mechanistic claims are not supported by the cited references.

Examples include:

The statement that HIF1α “interacts with” Hes7 and Lfng transcripts is not demonstrated by reference 92. That preprint reports abnormal Hes7 and Lfng expression following HIF1α loss, not a physical interaction with these RNAs. Reference 98 is a general study of transcription factor–RNA interactions and does not establish this specific mechanism. Reference 92 also remains explicitly identified as a non-peer-reviewed preprint.
Reference 129 is a prenatal case report of a MESP2 variant. It does not, by itself, demonstrate that abnormal MESP transcript processing or delayed termination of MESP protein expression causes the reported phenotype.
RPL38 and RPL10A data support roles in selective translation and mesodermal development, but the manuscript acknowledges that neither has been directly demonstrated to regulate segmentation-clock oscillations. Their placement in the disease pathway of Figure 2 therefore overstates the evidence.
References 27–30 concern stress granules, neuronal RNA granules, Alzheimer-related cellular toxicity, and triple-negative breast cancer. Their relevance to segmentation-clock disorders is unclear and should either be explained or these references should be removed.

The authors should systematically verify that every cited article supports the specific sentence in which it is cited.

Response: We have therefore performed a sentence-level citation audit and revised the relevant passages. We agree that the previous wording implying that HIF1α “interacts with” Hes7 or Lfng transcripts was not supported by the cited study. The relevant preprint reports that loss of HIF1α perturbs Hes7 and Lfng expression/oscillatory behavior and that oxygen-related manipulation can modify the phenotype. We have therefore removed language implying direct HIF1α–RNA interaction and now describe the result more conservatively. We also no longer use the general transcription factor–RNA interaction study as evidence for a specific HIF1α–Hes7/Lfng interaction. The HIF1α study remains explicitly identified as a preprint and its conclusions are described as provisional.

In terms of MESP2, we have revised the text so that the prenatal MESP2 case report is used only as clinical genetic evidence, not as evidence for a specific RNA-processing or protein-termination mechanism.

We also agree that the previous placement of RPL38 and RPL10A within the disease pathway in Figure 2 overstated their relationship to the segmentation clock.

We have therefore:

removed RPL38/RPL10A from the direct disease/clock pathway in Figure 2;
retained them only in the text and evidence table as Level II developmental/translational evidence;
and explicitly state that they support the broader concept of selective translation during axial development rather than a demonstrated clock-specific mechanism.

We agree that the previous use of references concerning stress granules, neuronal RNA granules, Alzheimer-related toxicity, and triple-negative breast cancer was too remote from vertebrate segmentation disorders. We have therefore removed them from sentences describing human vertebral segmentation disorders and deleted them where they were not needed for a clearly labeled cross-system inference.

The epitranscriptomic component is substantially underdeveloped relative to its prominence in the title and abstract.

Response: We thank the reviewer for this important comment. We agree that the previous version did not sufficiently distinguish between RNA regulatory mechanisms that have been directly demonstrated to control segmentation-clock dynamics and epitranscriptomic mechanisms whose roles in the presomitic mesoderm (PSM) remain largely hypothetical. We have therefore substantially revised this aspect of the manuscript to avoid presenting emerging epitranscriptomic mechanisms with the same level of causal certainty as established transcriptional, RNA-processing, and transcript-turnover mechanisms.

First, we have reduced the prominence of epitranscriptomic regulation in the title and abstract, and now explicitly describe RNA modification and editing as “emerging and incompletely tested regulatory layers”. We also clarify that direct evidence linking m6A, pseudouridylation, RNA editing, or tRNA modifications to segmentation-clock periodicity is currently limited.

Second, we have removed Section 3.3 to more clearly separate established post-transcriptional mechanisms from speculative epitranscriptomic mechanisms. The discussion of transcriptional bursting and stochastic transcriptional noise, which primarily concerns the generation and kinetics of the core oscillator rather than epitranscriptomic regulation, has been reframed within “5. Evidence gaps and testable hypotheses” section addressing RNA kinetics of the oscillator. In this section, epitranscriptomic regulation is presented as a hypothesis-driven future research direction. We identify several experimentally testable questions, including whether specific RNA modifications alter the stability, translation, localization, or phase of oscillatory transcripts and whether perturbation of RNA-modifying enzymes changes clock periodicity or somite formation.

Third, we have revised the discussion of the tissue-specific tRNA modification study previously cited as Reference 68. We now state explicitly that this study provides a general principle that tRNA modification can be tissue- and developmental-state-specific but does not directly demonstrate a role for tRNA modification in the segmentation clock, PSM biology, or somitogenesis. We have therefore removed any wording that could imply direct mechanistic evidence.

In accordance with the reviewer’s suggestion, we also outline specific experimental approaches that could address these questions, including transcript half-life measurements, modification-specific RNA profiling, ribosome profiling, live RNA imaging, and CRISPR-mediated perturbation of RNA-modifying enzymes in human iPSC-derived somitoids/segmentoids. These approaches would allow future studies to distinguish correlative changes in RNA modification from direct causal effects on segmentation-clock dynamics.

The manuscript is excessively long and contains substantial conceptual repetition.

A more focused structure could separately address:

RNA kinetics of the core oscillator;
intercellular coupling and species-specific differences;
translation, proteostasis, and metabolic modulation;
conversion of oscillations into boundaries and clock termination;
human genetics and disease;
evidence gaps and future experiments.
Figures and tables need conceptual and technical revision.

Table 3 should be reconstructed as discussed above. The authors should also disclose the source and licensing of all graphical elements and any software used to create the figures.

Response: We thank the reviewer for this detailed and constructive comment. We agree that the previous version contained substantial repetition, particularly regarding FGF/Wnt gradients, Hes/Her negative-feedback loops, Notch-mediated synchronization, transcript kinetics, and disease implications. We have therefore undertaken a substantial structural revision of the manuscript rather than making only local textual changes. The revised organization separates mechanistically distinct aspects of the segmentation clock, removes repeated explanatory passages, and more clearly distinguishes established mechanisms from hypotheses and indirect evidence.

We systematically reviewed the manuscript for repeated explanations of FGF/Wnt gradients, Hes/Her feedback, Notch synchronization, RNA turnover, and vertebral disease consequences. Repetitive introductory descriptions have been removed or consolidated so that each concept is now introduced in detail only once and subsequently referred to only where mechanistically necessary. In particular, repeated concluding statements asserting that “precise RNA regulation is essential for development” or that its disturbance causes vertebral abnormalities have been removed from individual subsections.

We also agree that the previous discussion of early embryogenesis and ZEB1/ZEB2 was disproportionately detailed relative to the RNA–protein dynamics that constitute the main focus of the review. We have therefore markedly condensed this material. General embryological events are now limited to the minimum context required to explain establishment of the presomitic mesoderm, and the extended discussion of ZEB1/ZEB2 lineage regulation has been substantially shortened. ZEB-related findings are retained only insofar as they establish the developmental context in which the segmentation clock emerges, without presenting them as central components of the clock itself.

We adopted the reviewer’s recommendation to separate the manuscript into more focused conceptual modules. The revised structure now distinguishes:

RNA kinetics of the core oscillator, including transcription, processing, transcript stability, and delayed negative feedback;

intercellular coupling and species-specific mechanisms, including mouse Dll1/Lfng/Hes7 and zebrafish deltaC/deltaD/her1/her7 systems;

translation, proteostasis, and metabolic modulation, including NICD turnover, ribosomal regulation, glycolytic inputs, and other protein-level mechanisms;

conversion of oscillations into segment boundaries and clock termination, including Tbx6/Ripply/Mesp modules and the transition from oscillatory to differentiated states;

human genetics and developmental disease; and evidence gaps and future experiments, including post-transcriptional and emerging epitranscriptomic hypotheses.

In addition, the manuscript has been reorganized so that FGF-mediated scaling, tailbud signaling, oscillator arrest, boundary formation, and differentiation are discussed within the broader framework of conversion of oscillations into boundaries and clock termination. Direct evidence for RNA decay is now discussed only where transcript stability or degradation has been experimentally demonstrated. We no longer imply that RNA decay is the principal mechanism underlying clock termination where such evidence is lacking.

In response to the reviewer’s concerns regarding epitranscriptomics, we introduced a dedicated section entitled: “ 5. Evidence gaps and testable hypotheses: RNA-mediated post-transcriptional control and emerging epitranscriptomic mechanisms” in lines 600-689. This section now contains mechanisms for which direct segmentation-clock evidence remains limited, including m6A, pseudouridylation, RNA editing, tRNA modification, ribosome heterogeneity, and other emerging RNA-regulatory mechanisms.

To further prevent established and speculative mechanisms from being presented with equal weight, we introduced an explicit hierarchy of evidence in the Methodology lines 187-202 and a dedicated evidence-summary table (new Table2).

Figure 1 and its legend have therefore been revised to distinguish evidence-supported mechanisms from emerging or hypothetical relationships using an explicit evidence code, including different arrow styles where appropriate. Species are now indicated wherever species-specific genes or mechanisms are shown or discussed. The figure has also been simplified to reduce overlap with Figure 2, with Figure 1 serving primarily as the integrative overview of the segmentation-clock framework.

Figure 2 has been also revised to provide a more focused representation of RNA/protein regulation and Notch-mediated coupling. Importantly, the previous cross-species mixture of HES7 and HER1/7 has been removed. The central pathway is now explicitly identified as a mouse PSM example, using Dll1, Hes7, and Lfng, whereas zebrafish her1/her7 and deltaC/deltaD mechanisms are discussed separately. All graphical elements were created by the authors using BioRender, and no copyrighted third-party graphical material was reproduced/adapted.

For tables, we agree that the previous Table 1 inappropriately combined intact embryos and stem-cell-derived systems and that the “average size of somite” column was not quantitatively comparable across models. Table 1 has therefore been reconstructed as a species- and model-specific comparative table. In vivo vertebrate models and human stem-cell-derived somitogenesis systems are clearly distinguished, and the non-comparable “average somite size” parameter has been removed.

The former table 2 has been conceptually reorganized. Gene families are no longer grouped under ambiguous labels, species-specific genes are separated, and the table distinguishes RNA-level mechanisms from protein/signaling mechanisms.

The former table 3 has been completely re-curated according to clinical-genetics standards and now uses exact HGNC-approved gene symbols, reference transcripts, HGVS c. and p. nomenclature, zygosity/inheritance information, variant classification, direct RNA consequences where demonstrated, protein/signaling consequences, phenotype, and strength of the gene–disease relationship. Generic gene-family labels and clinically heterogeneous associations have been removed. We also removed Adams–Oliver syndrome/DLL4 and TBX6-associated congenital scoliosis/MRKH from the disease table to avoid overextending the segmentation-clock interpretation.

Collectively, these revisions have shortened repetitive portions of the manuscript, consolidated overlapping concepts, separated direct evidence from emerging hypotheses, and reorganized the review around discrete mechanistic questions.

The ethics discussion should be updated and made more precise.

Response: We thank the reviewer for this important correction. We agree that our previous wording was overly broad and incorrectly implied that the conventional 14-day/primitive-streak framework governing human embryo culture applies uniformly to stem cell-based embryo models. We have revised this section to reflect the current ISSCR Guidelines for Stem Cell Research and Clinical Translation, including the 2025 update concerning stem cell-based embryo models. (Lines 989-1002)

Minor comments

“Her” is incorrectly expanded as “human epidermal growth factor receptor.” In this context, zebrafish her1/her7 are hairy-related transcriptional repressors. This error appears in the abstract, tables, and abbreviation lists and must be corrected throughout.

Response: Error mentioned above has been corrected throughout the manuscript.

Human gene symbols should be uppercase and italicized, mouse genes should use initial capitalization, and zebrafish genes should use lowercase italics. Protein names should not be italicized.

Response: Appropriate correction has been made throughout the manuscript.

Correct typographical and language errors, including “Neuosurgical,” “traslation,” “Table1,” “Figure2,” “T-box transcription factor6,” and the awkward phrase “Summary of RNAs role.” The term “derealization” in the ZEB2 paragraph also appears inappropriate and should be checked against the cited source.

Response: Appropriate corrections have been made.

The abstract sentence stating that disruptions in RNA mechanisms “impair protein-level and signaling abnormalities” is grammatically and conceptually incorrect.

Response: We thank the reviewer for identifying this wording error. We agree that the phrase “impair protein-level and signaling abnormalities” is both grammatically incorrect and conceptually unclear. The revised wording now states that disruptions in RNA regulatory mechanisms can lead to protein-level and signaling abnormalities, rather than “impair” such abnormalities.

The term “RNA disease” is used too broadly and should either be explicitly defined or replaced with more specific descriptions of the molecular defect.

Response: We thank the reviewer for this important comment. We agree that the term “RNA disease” was used too broadly in the previous version and could imply a unified disease category that is not appropriate for the heterogeneous mechanisms discussed in this review. Where “RNA disease” was used as a broad descriptor, we replaced it with more precise terminology such as “RNA processing,” “post-transcriptional regulation,” “RNA-associated molecular defects,” or “vertebral segmentation disorders,” depending on the context.

Abbreviations such as SCD, SCDO, SDV, CS, and KFS should be standardized. The manuscript should distinguish a clinical phenotype from a specific molecularly defined disorder.

Response: We thank the reviewer for this important clinical-genetics clarification. We agree that the previous version used several overlapping abbreviations inconsistently and, in some places, blurred the distinction between a descriptive clinical phenotype and a molecularly defined disorder. We have therefore standardized the terminology throughout the manuscript, figures, and tables. We now use spondylocostal dysostosis (SCD) as the clinical disorder category and reserve SCDO1, SCDO2, SCDO3, SCDO4, and SCDO6 for molecularly defined subtypes associated with specific genes where appropriate.

Reference 49 and reference 164 appear to describe the same TBX6/MRKH study and should be deduplicated.

Response: Duplicated reference has been removed and referencing order corrected.

The reference list contains numerous formatting or transcription errors, including author-name errors, malformed journal titles, inconsistent DOI formatting, and apparent typographical errors in several DOIs. A complete reference audit is necessary.

Response: We have performed complete citation audit.

The manuscript template still displays “Int. J. Mol. Sci. 2025,” although the bibliography contains multiple 2026 publications. The document metadata should be updated.

Response: Year has been updated accordingly.

The final statement concerning future therapeutic strategies is too speculative for congenital structural abnormalities established during embryogenesis. Diagnostic and mechanistic implications are plausible, but therapeutic implications should be qualified.

Response: We thank the reviewer for this important comment. We agree that the previous wording overstated the near-term therapeutic implications of the mechanisms discussed in this review, particularly because many congenital vertebral abnormalities are established during embryogenesis and may not be amenable to postnatal molecular correction. The revised manuscript no longer suggests that the RNA-regulatory mechanisms discussed here directly point to readily translatable therapeutic strategies for established congenital structural abnormalities.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

The author attempts to reexamine the classic developmental biology model of the vertebrate segmentation clock from the emerging perspective of “dynamic RNA regulation.” The author systematically reviews RNA-related mechanisms across multiple levels—from transcriptional regulation, RNA processing, stability, and translation to epigenetic transcriptomic modifications—and links them to congenital spinal developmental disorders. This review features a novel topic, a comprehensive structure, up-to-date literature, and a wealth of information, making it a valuable review in the current interdisciplinary field of segmentation clocks and post-transcriptional regulation. However, it still has some shortcomings in terms of its core argument and critical analysis.

The central argument that “RNA regulation is the core mechanism of segmental clocks” lacks sufficient direct evidence to support it—while the paper extensively cites studies on transcriptional regulation, signaling pathways, and protein stability, direct evidence specifically addressing “regulation at the RNA level” is relatively limited.
At present, this paper reads more like a summary report on “progress in research on RNA regulation of segmental clocks” rather than a review with distinct viewpoints and a debate-oriented approach. We recommend that the authors clearly state 2–3 core arguments in the introduction (e.g., “RNA processing kinetics are the determining factor for differences in circadian clock periods across species,” or “epigenetic modifications may represent a new mechanism for clock synchronization”) and develop their arguments around these points in the subsequent sections.
The article’s structure is loose and its central theme is unclear—the content ranges from the ZEB transcription factor to RPL38 translation regulation to FBXW7 protein degradation. Although titled “RNA Dynamics,” some sections have only an indirect connection to RNA regulation.
The review is unbalanced overall, with an excess of descriptive content and a lack of critical analysis and unique insights. In most sections, the authors adopt a descriptive pattern such as “Study A shows… Study B found…” in most sections, but fail to provide sufficient critical analysis of the limitations, contradictions, and unresolved issues in these studies.
The section on epitranscriptomic modifications (m6A, RNA editing, etc.) consists largely of speculative discussion and lacks experimental evidence. While the authors candidly acknowledge that this area is “largely unexplored,” this weakens the persuasiveness of their arguments in the review. Furthermore, while the authors emphasize “emerging epitranscriptomic mechanisms” in the abstract, this topic occupies only a small portion of the main text and is primarily framed as “although direct evidence linking RNA editing and specific RNA modifications to segmentation clock function remains limited.” However, as a review focused on “RNA regulation,” such “self-contradiction” may lead readers to question the paper’s value proposition. It is recommended that the authors either strengthen this section (by supplementing it with more indirect evidence or speculative models) or adjust the wording in the abstract to avoid overpromising.
The logical relationship between Sections 3.1 and 3.3 is unclear. Section 3.1 primarily discusses the effects of protein stability (FBXW7-NICD) and metabolic state (glycolytic metabolites) on the clock, while Section 3.3 discusses RNA modifications, translational regulation, and buffering of transcriptional noise. Although both sections address “clock maintenance,” the former focuses more on signaling pathways and protein-level regulation, whereas the latter focuses more on the RNA level. The logical relationship between the two (why are they divided into separate sections? How do they work together?) is not clearly explained. It is recommended to move Section 3.3 forward and treat Section 3.1 as a supplement.
Section 5, “Stopping the Clock,” is relatively brief: compared to the detailed discussions in Sections 3 and 4, Section 5’s discussion of “how the clock stops” is relatively concise. The “shutdown” of segmental clocks during development is an important biological question; for example, how does the clock stop in late embryonic development after PSM depletion? This warrants more in-depth analysis.
Section 7, “Future Perspectives,” is too general. Aside from mentioning new technologies (organoids, single-cell transcriptomics, in vivo RNA imaging) and ethical discussions, Section 7 lacks specific, testable scientific hypotheses.
The reference list contains a significant number of preprints; in particular, some references (e.g., Ref. 23, 45, 92, 125) support key arguments. Extensive citation of preprints in a formal journal review weakens the “peer-reviewed” status of the conclusions. We recommend verifying whether each preprint has been formally published and updating the references to the formal versions as needed.
The definition of the abbreviation "Her" is incorrect; both the abstract and the notes in Tables 1 and 2 define "Her" as "human epidermal growth factor receptor" (EGFR). In zebrafish, Her1/Her7 are hairy and Enhancer of split-related (Hes-related) bHLH transcriptional repressors, not EGFR receptors.
Spelling errors

L6: “Neuosurgical” → should be “Neurosurgical”

L708: “traslation efficiency” → “translation”

L183: “derealization” is a psychiatric term; this is clearly inappropriate. The author must verify the original text.

Author Response

Please find the responses in attached word file, Reviewer 2 section. Thank you.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript provides a clear and well-organized review of RNA and protein dynamics in the vertebrate segmentation clock. The topic is interesting, and the literature coverage is broad. The manuscript presents the major regulatory mechanisms and their links to developmental disorders in a coherent way. I have a few minor comments. I recommend Minor Revision.

1. Table 3 summarizes different molecular consequences of genetic variants. However, it is not always clear whether these effects were directly demonstrated in the original studies or inferred from the variant itself. This is especially important for effects such as splicing defects and nonsense-mediated decay, which may be predicted but not directly tested at the RNA level. The authors could clarify this in the table or the table legend, this would make the evidence easier to interpret.
2. The manuscript discusses computational and quantitative methods such as eGRN and RAFL analyses. A short explanation of what these analyses measure or infer would help readers who are less familiar with these methods.
3. The main text states that early Lfng oscillations can occur independently of canonical Notch signaling and that clock initiation precedes the classical Notch–Hes/Her feedback loop. However, the Figure 2 legend states that Notch signaling “initiates and synchronizes” oscillations. The authors should clarify the difference between clock initiation and later Notch-dependent maintenance or synchronization.

Overall, the manuscript is well written, and only minor revisions are needed.

Author Response

Reviewer3

Table 3 summarizes different molecular consequences of genetic variants. However, it is not always clear whether these effects were directly demonstrated in the original studies or inferred from the variant itself. This is especially important for effects such as splicing defects and nonsense-mediated decay, which may be predicted but not directly tested at the RNA level. The authors could clarify this in the table or the table legend, this would make the evidence easier to interpret.

Response: We thank the reviewer for this important comment. We agree that the previous version of Table 3 did not consistently distinguish molecular consequences that were directly demonstrated experimentally from those that were inferred from variant type.

We have revised Table 3 (new Table 4 in page 24-25 the revised manuscript) so that the molecular consequence of each variant is now explicitly qualified according to the supporting evidence. Effects are labeled as “experimentally demonstrated,” “predicted/inferred,” or “RNA consequence not directly demonstrated,” as appropriate.

We have added an explanatory statement to the Table legend indicating that molecular consequences are reported according to the level of experimental support in the cited primary study and that predicted effects are explicitly distinguished from directly demonstrated RNA or protein consequences. The revised legend now states:

“Molecular consequences are classified according to the evidence reported in the cited studies. ‘Experimentally demonstrated’ indicates direct RNA-, protein-, or signaling-level validation. ‘Predicted/inferred’ indicates a consequence inferred from variant type, sequence context, or established molecular principles but not directly demonstrated in the cited study. In particular, predicted splice disruption and nonsense-mediated decay are not considered experimentally established unless supported by transcript-level analysis.”

The manuscript discusses computational and quantitative methods such as eGRN and RAFL analyses. A short explanation of what these analyses measure or infer would help readers who are less familiar with these methods.

Response: A short description has been added respectively in lines 480-482 and lines 581-582.

The main text states that early Lfng oscillations can occur independently of canonical Notch signaling and that clock initiation precedes the classical Notch–Hes/Her feedback loop. However, the Figure 2 legend states that Notch signaling “initiates and synchronizes” oscillations. The authors should clarify the difference between clock initiation and later Notch-dependent maintenance or synchronization.

Response: We thank the reviewer for pointing out this inconsistency. We agree that the previous wording in the Figure 2 legend could be interpreted as suggesting that canonical Notch signaling is required to initiate the segmentation clock, which is not consistent with the evidence discussed in the main text.

We have revised the manuscript to distinguish clearly between initial oscillator onset and subsequent Notch-dependent intercellular coupling and synchronization. The revised text now states that early oscillatory activity, including initial Lfng waves in the mouse presomitic mesoderm, can arise before canonical Notch-dependent coupling is fully engaged. The Figure legend has been revised to remove the phrase “initiates and synchronizes oscillations.” It now states that Notch signaling “couples and synchronizes neighboring oscillators after clock onset” and helps maintain coherent oscillatory propagation across the presomitic mesoderm.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have carefully and comprehensively addressed the comments raised during the previous round of review.

The revised manuscript is substantially improved in terms of structure, methodological transparency, interpretation of the available evidence, and distinction between experimentally established mechanisms and more speculative or indirect observations. In particular, the introduction of a clearly defined evidence hierarchy considerably strengthens the scientific rigor of the review and helps the reader distinguish direct segmentation-clock evidence from developmental, clinical-genetic, and hypothesis-generating data. ijms-4541627-peer-review-v2

The manuscript also now handles species-specific differences more appropriately and avoids overgeneralization across mouse, zebrafish, human stem-cell-derived models, and other vertebrate systems. The revised figures and accompanying explanations further improve conceptual clarity, particularly by distinguishing established mechanisms from proposed or indirect relationships.

Importantly, the clinical section is now more appropriately framed. The authors clearly distinguish pathogenic variants associated with vertebral segmentation disorders from direct experimental evidence of altered oscillator dynamics in affected patients, which addresses an important concern from the previous review. ijms-4541627-peer-review-v2

Overall, I consider the authors’ responses satisfactory and believe that the manuscript has been strengthened substantially through revision.

I have no further major comments. The manuscript is suitable for publication in its current form.

Reviewer 2 Report

Comments and Suggestions for Authors

Upon review of the revised manuscript, I am satisfied that the concerns raised in my earlier review have been adequately addressed, and I have no further comments.