JCI Insight Response to Reviewers: A Translational Revision Guide

Every response needs a page and line reference. For experiments, add the figure and panel; for analyses, add the table or statistical subsection; for resource concerns, add the reagent table or supplement.

Clinical and translational revisions propagate. If a cohort reanalysis changes an effect estimate, update the Results, figure, abstract, discussion, and any graphical summary. List all those locations in one response. A reviewer should not find the old estimate elsewhere in the marked PDF.

After compiling the final response-prefaced manuscript, verify the locations from that combined file. The response pages themselves can shift all later pagination if you cite the wrong PDF version.

Because the response precedes the marked manuscript, readers should see an immediate visual boundary between reviewer text and author response. Use bold or shaded reviewer blocks plus explicit labels. Keep author replies in regular text. Do not use red and green alone.

Within the manuscript, mark substantive additions consistently. A response that cites page 12 but leaves the new causal limitation visually indistinguishable from old text makes verification harder than it needs to be.

Do not treat "another cohort" as a universal answer. A second convenience cohort may repeat the same selection bias and leave the translational concern untouched.

Another reviewer's silence is not evidence. Answer each scientific uncertainty on its own merits.

In our pre-submission review work with JCI Insight manuscripts and revisions, we test whether the response, marked manuscript, clean manuscript, figures, statistical outputs, and supplementary evidence tell one translational story. This is where generic point-by-point advice stops being useful: a gracious answer can still leave mechanism, model, patient evidence, and claim boundary misaligned. The observations below are anonymized Manusights patterns, not access to internal JCI Insight decisions or private editorial data. Each can be checked directly from an author's own files.

Mechanism claimed from parallel movement. In JCI Insight revisions, the perturbation and phenotype may change together, but no rescue, temporal test, or pathway-specific control separates cause from correlation. Authors add pathway discussion in the response while leaving the abstract's causal verb intact. We flag the verb first because that is where reviewers see the unresolved inference, then trace it through the model figure, Results, and Discussion.

Human relevance reduced to one expression plot. A JCI Insight disease-model result can be paired with a public dataset showing differential expression, then described as clinically validated. The cohort lacks treatment context, stage, tissue handling, or an endpoint aligned with the claim. A credible revision defines what the human evidence actually establishes and puts that boundary in the abstract and figure legend, not only in a limitation paragraph.

Experimental units that change by figure. Animals, wells, images, and cells are counted inconsistently. A revised Methods paragraph cannot repair confidence intervals derived from the wrong unit. The response must state which analyses were rerun and where every corrected result appears.

A supplementary repair with an unchanged narrative. Authors add controls to the supplement but do not alter the model figure, Results hierarchy, or conclusion. Returning reviewers compare the central story with the new evidence. If the repair narrows the mechanism, the narrative must narrow too.

Our final check asks whether a skeptical reader could reconstruct the revised inference from the Methods and figures without reading the rebuttal. If not, the response is carrying scientific content that belongs in the manuscript. That cross-file test regularly catches incomplete revisions that a checklist of comments would mark as done.

Review the JCI Insight response and revised translational chain together before all authors approve the package.

Create one issue ledger shared by all coauthors. Merge comments that target the same claim even if they ask for different experiments. An editor's concern about disease relevance, a reviewer's request for a patient cohort, and another reviewer's objection to the model may all point to one translational gap.

Choose one evidence plan, then reference the same plan from each answer. Contradictory replies are particularly visible in a combined response-and-manuscript PDF.

Disagree when the experiment cannot identify the concern, would violate ethics or consent boundaries, or would imply a different study design. Explain the distinction and provide the closest valid evidence.

If a reviewer requests an intervention outcome from an observational cohort, do not imply causality from additional adjustment. Add the relevant sensitivity analysis, state the residual confounding, and remove treatment-effect language.

Most rejection-on-revision risk sits in unresolved central inference, not missing courtesy. A causal mechanism remains correlational, a model-to-patient leap remains unsupported, or a statistical correction changes the result without changing the claim. Those are scientific failures even when every response paragraph is polished.

A revision invitation is not an acceptance. If the new evidence no longer supports the original headline, revise the headline. Editors can work with a bounded conclusion; they cannot accept a response that conceals the boundary.

1. Put the Editors' controlling scientific issues first.
2. Connect mechanism, model, patient evidence, and claim boundary.
3. Define biological units, replicate types, effect sizes, and multiplicity.
4. Supply reagent identifiers and validation where relevant.
5. Reproduce every reviewer comment and answer in order.
6. Cite page, line, figure, panel, table, Methods, and supplement locations.
7. Make every promised change visible in the marked manuscript.
8. Reconcile abstract, figures, Results, and Discussion after reanalysis.
9. Obtain approval from every listed author before the deadline.
10. Re-run a JCI Insight journal-fit check if the revision materially narrows the translational claim.

This page was last reviewed on July 12, 2026. Official JCI Insight sources control the submission mechanics; Manusights analysis provides the translational revision map.