Nature Figure Guidelines: A Pre-Upload Decision Matrix
A practical figure audit for Nature submissions that separates editorial readability, the Source Data requirement, and final-artwork production rules.
Next step
Choose the next useful decision step first.
Use the guide or checklist that matches this page's intent before you ask for a manuscript-level diagnostic.
Quick answer: Prepare a Nature figure for two separate gates: an editor and reviewer must be able to interpret it at initial submission, and the underlying files must satisfy final-artwork and Source Data rules later if the paper is accepted. Verify the current stage-specific instructions before upload; do not treat an old dimension or resolution rule as timeless.
Evidence basis: We reviewed Nature's formatting guide, its initial- and final-submission pages, and the Nature Portfolio image-integrity policy on September 25, 2026. The journal's own pages set the actual rule at each stage; the checklist below is a pre-upload pass built from those rules, not a replacement for reading them.
Nature's editors and reviewers make the scientific call; this checklist only tests whether a figure holds up before it reaches them, so a stage-two rule cannot sink a stage-one submission and a scientific problem does not survive because the artwork looks finished.
Check | Ready to upload | Not yet |
|---|---|---|
Scientific claim | Each panel maps to a named result in the text | A panel is decorative or the text never references it |
Source Data | The raw numbers behind each chart already exist in a workbook | The chart exists but the numbers behind it do not, yet |
Visual encoding | Groups, scale, and units read without guessing | The reader needs color alone to tell groups apart |
Reduction | Text holds up at 89 mm (single column) or 183 mm (double) | Legible only at full zoom on a large screen |
File type | Line art is still editable vector; photos hold their resolution | Line art was flattened to a raster image somewhere along the way |
Stage one: initial submission is more flexible than authors assume
Nature encourages a single Word or PDF file with figures embedded in the text or grouped at the end, each legend beside its figure, up to about 30 MB. It does not require production-resolution artwork at this stage and explicitly asks authors not to upload high-resolution files until an editor requests them. Authors who spend hours perfecting final-artwork specifications before the first editorial read are solving a problem that does not exist yet at this stage.
What does matter at initial submission: line numbers, legible figures at a normal read, and a figure set that lets a triaging editor follow the paper's argument without opening a separate viewer. A figure that is scientifically clear but technically rough at this stage costs you nothing; a figure that is polished but hides the actual result costs you the submission.
Stage two: final artwork has real, specific rules
If the paper is accepted in principle, production figures separate from the review-stage file. Nature's standard figure widths are 89 mm for one column and 183 mm for two, and it prefers editable, unflattened vector artwork (EPS, AI, or PDF) for line drawings, graphs, and schematics rather than rasterized versions. Photographic images generally need meaningfully higher resolution than a screen-display export. RGB color is recommended for color artwork. These are final-artwork specifications; do not present them as blocking checks on an initial-review PDF.
Source data and unprocessed images: the part most authors leave late
Nature's initial-submission page encourages authors to provide source data in an unstructured public repository such as Dryad or figshare, or as Supplementary Information, and says that if full access to data is needed for peer review, authors must provide it. For life-science papers, the Nature Portfolio image-integrity policy goes further: unprocessed original images of gels and western blots must be submitted with the final accepted version and are published in the Supplementary Information, and editors may ask for unprocessed data files during review.
This is not a formatting nicety. It means the numbers behind every bar chart, scatter plot, and line graph, and the full scans behind every cropped blot, need to exist somewhere reviewable outside the rendered image, not reconstructed from the figure after the fact.
Build the Source Data workbook alongside the figure, not after a reviewer or editor asks for it. A figure whose underlying spreadsheet was assembled retroactively is where transcription errors between the plot and the claimed values tend to hide.
Where each piece of figure information actually belongs
A recurring mistake is putting the wrong information in the wrong place, then discovering the gap only when a reviewer asks a question the figure should have already answered.
This information | Belongs in | Symptom when it is misplaced |
|---|---|---|
Panel-by-panel meaning | The legend | A reader cannot tell what a panel is arguing without re-reading the Results |
How the data was collected and analyzed | Methods | The legend balloons into a paragraph of procedure text |
The actual numbers behind a chart | The Source Data workbook | A reviewer asks for values that were never exported anywhere |
The claim the figure supports | The Results narrative | The figure caption asserts something the body text never states |
Before upload, walk the chain from Source Data workbook to figure panel to legend to Results sentence to figure callout, in that order, and confirm each link matches the next. Extended Data exists to support the main Article online, not to relocate evidence a reader needs for the central claim somewhere less visible. Use the figure-quality check for general design work and a manuscript readiness check to test whether the text claims more than each cited figure shows.
Readiness check
Run the scan while the topic is in front of you.
See score, top issues, and journal-fit signals before you submit.
When a cleaner-looking figure is actually a weaker one
A figure redesign done late, under deadline pressure, sometimes fixes the visual problem and quietly breaks the evidentiary one. Watch for three specific tells. An axis or color scale gets rescaled so two genuinely different results look closer together than the data supports. A comparison that came back null gets dropped from the panel while the favorable comparisons stay, so the figure set no longer represents what was actually tested. And a group label that used to sit on the plot gets moved into the legend, so identifying which line is which now requires cross-referencing dense text instead of reading the chart.
None of these are formatting problems. They are places where the polish pass and the evidence diverged, and a reviewer who checks the underlying Source Data against the plotted claim will find the gap faster than a casual read of the figure ever would.
What we see before Nature-targeted submissions go out
Two figure gaps keep coming back in the Nature-bound manuscripts that reach our pre-submission review. The first is a legend that carries the methods, which tests were run and how, while the Methods statistics section stays thin. Nature asks legends to describe what a figure depicts rather than the methods used, wants error bars defined throughout the figures with exact n values, and wants the tests themselves described in a statistics section of the Methods. The second is a figure set with no source data prepared at all, which leaves the authors rebuilding values from their own plots when an editor or reviewer asks for them.
This is not a prediction of what Nature's editors will decide. The free readiness scan reads the manuscript text, including legends and Methods, but it does not see the rendered figures. The $39 Full Review adds a read of the figure images themselves, but matching every plotted value to its source spreadsheet is still a job for the lab before upload.
Sources accessed September 25, 2026.
- Formatting guide, Nature.
- Final submission, Nature.
- Initial submission, Nature.
- Image integrity and standards, Nature Portfolio.
Frequently asked questions
Not as one blanket rule on Nature's public author pages. Nature encourages source data in a public repository such as Dryad or figshare, or as Supplementary Information, and says authors must provide full data access if peer review requires it. For life-science papers, Nature Portfolio policy requires unprocessed original gel and western blot images with the final accepted version. Follow the fuller final-submission instructions Nature sends if the paper is accepted.
Nature does not ask for production-resolution files at initial submission and asks authors not to upload them until an editor requests them. High-resolution requirements (vector line art, roughly 300 dpi photographic images) apply at final artwork, not the first review round.
Nature asks for line art, graphs, charts, and schematics as vector formats (EPS, AI, or PDF) exported directly from the application that made them, not as rasterized images. Photographic images are handled separately and generally need higher resolution at final submission.
No. A legend can define symbols and conditions, but it cannot repair a missing unit, an ambiguous encoding, or a visual hierarchy that only makes sense at full zoom.
Before you upload
Choose the next useful decision step first.
Move from this article into the next decision-support step. The scan works best once the journal and submission plan are clearer.
Use the scan once the manuscript and target journal are concrete enough to evaluate.
Private API processing. Your manuscript is not used to train models.
Put the guidance to work
Choose the practical next step for your research.
Start with the question in front of you, or use the guided learning path when you want a clearer view of the full publishing process.