Nature Figure Guidelines: Size, DPI & Format (2026)

Preparing Figures for Nature, Science, and Cell: A Technical Reference

A manuscript accepted by Nature, Science, or Cell represents the highest tier of scientific publishing. Yet even original, rigorous research can stall at the editorial stage because of figures that fail to meet technical standards. Reviewers and editors form rapid impressions based on visual clarity, and a figure that looks unprofessional or is incorrectly formatted can trigger rejection before the science is even evaluated.

This guide details the precise technical specifications for all three journals: correct figure widths (89 mm and 183 mm for Nature), resolution minimums (300 DPI), acceptable file types, typographic rules, and pitfalls that commonly cause submissions to bounce back.

AI Scientific Image Generator

Create publication-quality scientific figures with AI. Export at 300 DPI for journal submission.

Try it free →

The Role Figures Play in Acceptance Decisions

Research published in PLOS ONE and editorial commentary from Elsevier both point to unclear or incomplete figures as a leading cause of manuscript rejection.

At journals with acceptance rates under 10%, editors handle far more submissions than they can evaluate in depth. Figures are often the entry point for that evaluation. A figure that is hard to read or inconsistently formatted signals insufficient care, which can color how the underlying data is perceived.

A publication-quality signaling pathway figure. Clear labels, intentional color choices, and a balanced layout demonstrate the kind of visual discipline top journals expect.

What strong figures communicate to reviewers:

• The research was conducted carefully and systematically
• The authors understand the visual conventions of their discipline
• Results have been presented with the reader's comprehension in mind
• Attention to detail extends across every aspect of the work

Specification Comparison: Nature, Science, and Cell

The three journals share many conventions but differ on key details. Knowing those differences before you build your figures saves significant revision time.

Figure Dimensions

Resolution Requirements

File Formats

Typography

Color Mode

All three journals accept RGB for initial submission. Color conversion to CMYK for print is handled by the journal's production team. Authors who want to preview the printed result can convert a copy to CMYK beforehand, though this is optional.

Accessibility requirements shared across all three journals:

• No red-green combinations, which are indistinguishable for colorblind readers
• Avoid rainbow or jet color scales for continuous data
• Text contrast ratio must exceed 4.5:1
• Colors must be distinguishable in grayscale reproductions

For thorough guidance on palette selection, see our Scientific Color Palette Guide.

Figure Types and Appropriate Tools

Each category of scientific figure calls for a different production approach. Matching the tool to the figure type prevents quality issues downstream.

1. Data Visualizations: Graphs, Charts, and Statistical Plots

Recommended tools: R with ggplot2, Python with matplotlib or seaborn, GraphPad Prism, OriginPro

These environments generate figures directly from raw data, which preserves reproducibility and makes it straightforward to update when reviewers request changes.

Key practices:

• Export as vector formats (PDF, EPS, SVG) wherever the software allows
• Apply a consistent visual theme across all data figures in a paper
• Include error bars, confidence intervals, and statistical significance markers as required
• Set font families to match the target journal before exporting

Sample export code in R:

# R example with ggplot2
library(ggplot2)
ggplot(data, aes(x, y)) +
  geom_point() +
  theme_classic() +
  theme(text = element_text(family = "Arial", size = 7))
ggsave("figure1.pdf", width = 89, height = 60, units = "mm")

2. Molecular Structures and Chemical Diagrams

Recommended tools: ChemDraw, PyMOL, UCSF Chimera, UCSF ChimeraX

Specialized chemistry and structural biology software ensures that bond angles, stereochemistry, and molecular geometry follow IUPAC conventions.

A reaction mechanism illustration showing correct bond representation and electron flow, built with chemical accuracy as the priority.

Key practices:

• Follow your journal's chemistry style guide for bond lengths and notation
• Export at 1000 DPI when submitting line art
• Explicitly depict stereochemistry whenever it is relevant to the result

3. Biological Schematics and Pathway Figures

Recommended tools: Adobe Illustrator, BioRender, Inkscape, Figviz

Conceptual figures, signaling pathway maps, and experimental workflow diagrams require tools that offer precise control over layout and iconography.

Key practices:

• Keep icon styles and stroke weights uniform across all schematic figures
• Assign color purposefully to convey meaning, not purely for aesthetics
• Use directional arrows that are easy to follow at publication size
• Keep text annotations brief and positioned so they do not overlap other elements

Rapid concept iteration: For generating initial diagram drafts before committing to final production, AI-assisted tools like Figviz can produce working schematics in seconds. This accelerates early ideation and lab-meeting presentations before the polished version is built in vector software.

4. Microscopy and Experimental Imaging

Recommended tools: ImageJ/FIJI, Adobe Photoshop, Imaris

Photomicrographs and fluorescence images have strict rules around processing to prevent data misrepresentation.

Key practices:

• Apply any brightness, contrast, or gamma adjustments identically across all comparable panels
• Scale bars are mandatory for nearly all journals: place them consistently, typically at the lower right
• Keep a written log of every processing step
• Select lookup tables (LUTs) that are perceptually uniform and colorblind-accessible
• Never manipulate images in ways that obscure, add, or remove features from the original data

Cell Press editorial guidelines state plainly: "Before you submit, check the original data and make sure you know how the figures were prepared from them."

5. Multi-Panel Composite Figures

Recommended tools: Adobe Illustrator, Inkscape, PowerPoint

Combining individual panels into a single submission figure requires precision alignment and consistent visual language.

A composite figure with uniform styling, legible labels, and well-organized spacing that makes efficient use of the available column width.

Key practices:

• Align all panel edges using guides or a grid; do not eyeball
• Keep inter-panel spacing uniform
• Label panels in reading order, using the convention specified by your target journal (lowercase for Nature, uppercase for Science and Cell)
• Set all annotation text to the same point size
• White space between panels aids readability and should not be eliminated to fit more content

Figure Production Workflow, Step by Step

A structured production process reduces the chance of errors and makes late-stage revisions less painful.

Step 1: Review Requirements Before Creating Anything

Before opening software:

1. Download the author instructions for your specific target journal, including any supplementary figure-preparation guides
2. Sketch a layout for each figure, noting what each panel needs to show and how the panels build a narrative
3. Classify each figure as a data plot, schematic, imaging figure, or composite, since this determines the toolchain
4. Collect all source materials including raw data files, images, and any icon libraries you plan to use

Step 2: Produce Each Element Independently

Build every panel using the most appropriate tool for that element type:

Workflow note: For schematic and conceptual panels, Figviz's diagram tools let you generate starting-point visuals quickly. These can be refined in vector software when preparing the final submission package.

Step 3: Assemble Panels in a Vector Environment

Import all elements into Adobe Illustrator or Inkscape:

1. Create an artboard at the exact target dimensions (for example, 89 mm wide for a Nature single-column figure)
2. Place each element as a linked or embedded object rather than copying and pasting as pixels
3. Arrange and align panels using the application's alignment tools
4. Add panel labels, annotations, and scale bars as vector text or shapes
5. Audit styling to confirm fonts, stroke weights, and colors are consistent throughout

Step 4: Verify Against the Specification Checklist

Before exporting:

• Artboard dimensions match the journal's column width requirement
• All text is between 5 and 8 pt at final output size
• Fonts are embedded or have been converted to outlines
• Color palette is accessible to colorblind readers
• No red and green appear as the sole distinguishing cues
• Panel labels follow the journal's specified case and weight
• Scale bars are present on all imaging panels
• Resolution meets or exceeds the minimum for the figure type

Step 5: Export in the Correct Format

For Nature:

File > Export > Export As > PDF
- Preserve Illustrator Editing Capabilities: ON
- Embed Page Thumbnails: OFF
- Optimize for Fast Web View: OFF

For Cell:

File > Export > Export As > TIFF
- Color Mode: RGB
- Resolution: 300 DPI (1000 DPI for line art panels)
- LZW Compression: ON

Step 6: Conduct a Final Quality Review

Open the exported file and confirm:

• Text is fully legible at 100% zoom
• No pixelation, compression artifacts, or rendering errors are visible
• Colors display as intended
• File size falls within the journal's upload limit (typically 10 to 20 MB per figure)

Eight Figure Problems That Lead to Rejection

Drawing on published editor feedback and official journal guidelines, these are the errors that most frequently cause figures to be returned or rejected.

1. Incorrect Resolution or Artboard Dimensions

Problem: The figure appears blurry in print or does not align with the journal's column grid.

Solution: Set your artboard to the target dimensions at the outset and work at 300 DPI minimum throughout. Scaling a figure down after the fact does not fix resolution problems introduced early in production.

2. Rasterized Text or Line Art

Problem: Text or diagram lines look soft or jagged because the element was saved as a bitmap.

Solution: Keep all text and line art in vector format until final export. In Illustrator, avoid flattening or rasterizing text layers.

3. Text That Cannot Be Read at Print Size

Problem: Labels that appear legible on a large monitor become unreadable when printed at journal column width.

Solution: Use the application's actual-size view to preview your figure at its true output dimensions. If you cannot read the annotations comfortably, reviewers will not be able to either.

Annotation text sized to remain legible at the column width where it will appear in print.

4. Red-Green Color Combinations

Problem: Approximately 8% of men cannot distinguish red from green, making figures that rely on this contrast uninterpretable for a meaningful portion of readers.

Solution: Substitute blue-orange, blue-yellow, or purple-green pairings. Validate your palette using a colorblind simulator before finalizing.

5. Inconsistent Visual Style Across Figures

Problem: Different figures use different fonts, stroke weights, or labeling conventions, giving the paper a disorganized appearance.

Solution: Define a style guide covering typeface, type size, color values, and line weights before production begins. Apply it rigidly across every figure in the manuscript.

6. Absent or Poorly Placed Scale Bars

Problem: Imaging figures without scale bars lack a critical piece of quantitative information.

Solution: Include a scale bar in every imaging panel. Position it consistently, with the lower-right corner being the most widely accepted convention.

7. Overcrowded Panel Layouts

Problem: Fitting too much content into a single panel reduces clarity and makes it harder for reviewers to extract the key finding.

Solution: Distribute content across additional figures or move supporting data to supplementary panels. Negative space within a figure is not wasted: it reduces cognitive load and highlights what matters most.

8. JPEG Artifacts on Vector-Style Content

Problem: Saving graphs or schematic figures as JPEG files introduces block-shaped compression artifacts that are especially visible around text and thin lines.

Solution: Use PDF, EPS, or SVG for figures containing vector elements. Reserve TIFF or PNG for photographic or imaging content.

Tool Comparison for Journal-Quality Figures

Suggested Workflows by Budget

Cost-free workflow:

1. R or Python for all data plots
2. ImageJ for image processing and adjustment
3. Inkscape for panel assembly and annotation
4. Figviz for initial schematic drafts

Full professional workflow:

1. GraphPad Prism or R for statistical figures
2. ChemDraw for chemical and structural diagrams
3. BioRender or Illustrator for biological schematics
4. Adobe Illustrator for all final assembly and annotation

Journal-Specific Tips

Nature

• Consult their Research Figure Guide as a primary reference: it is detailed and regularly updated
• JPEG, TIFF, and PNG are not accepted for main article figures; use vector formats
• Extended Data and Supplementary figures may use JPEG or TIFF
• Panel labels must be 8 pt bold lowercase (a, b, c)

Science

• Lower-resolution figures (150 DPI) are acceptable at the initial submission stage
• Final accepted manuscripts require full-resolution files
• The standard typeface is Times New Roman
• Position panel labels at the upper-left corner of each panel

Cell

• A graphical abstract is required, sized at 1200 x 1200 pixels at 300 DPI
• Accepted typefaces are Avenir and Arial
• Line art panels must be supplied at 1000 DPI
• Consider submitting Figure360 video supplements for figures depicting dynamic processes

Putting It Into Practice

The technical demands of top-journal figure preparation are significant, but the underlying principles are consistent across all three journals:

1. Confirm your target journal's requirements before any production work begins
2. Select the right tool for each category of figure element
3. Apply a unified style across every figure in the manuscript
4. Validate accessibility, particularly for colorblindness and print contrast
5. Export in the specified format and verify the output file before submitting

For concept exploration, lab meetings, and educational materials, tools like Figviz compress the time needed to produce working diagrams. For final submission to Nature, Science, or Cell, those concepts should be finished in professional vector software to satisfy the exact technical requirements.

Figures that communicate clearly and meet every specification do more than avoid rejection: they build confidence in the research itself.

A data visualization that presents results without visual noise, demonstrating the economy and clarity that characterizes publication-quality figures.

Related Resources

• Scientific Color Palette Guide
• How to Make Scientific Diagrams for Research Papers
• Best Free BioRender Alternatives
• Best Free Adobe Illustrator Alternatives
• Science Diagram Tools

FAQ

Q: What resolution should figures be for Nature? A: Nature requires 300 DPI for all submitted figures, with individual file sizes kept under 10 MB. For figures built entirely from vector elements (graphs, schematic diagrams), submit as PDF, EPS, or SVG: these formats are resolution-independent and will reproduce cleanly at any output size.

Q: Can I use PowerPoint to make figures for Science or Nature? A: PowerPoint is a viable option for straightforward figures, provided you export carefully to preserve quality. Nature accepts PowerPoint files converted to PDF. For complex or multi-panel figures, a dedicated vector application such as Illustrator or Inkscape gives you tighter control over final output.

Q: What is the difference between RGB and CMYK, and which should I use for submission? A: RGB (red, green, blue) is the native color model for screens and offers a broader gamut. CMYK (cyan, magenta, yellow, black) is used for offset printing. Current guidelines from Nature, Science, and Cell all specify RGB for submission: the journal's production team handles CMYK conversion. If you want to preview print colors before submitting, convert a copy of your file to CMYK and compare.

Q: How do I make figures accessible to colorblind readers? A: The most common form of color vision deficiency makes red and green appear similar, affecting roughly 8% of men. Replace red-green pairings with blue-orange, blue-yellow, or purple-green combinations. Check your final palette with a simulator such as Coblis or Color Oracle. Our Scientific Color Palette Guide covers this in depth.

Q: Should I include a graphical abstract? A: Cell Press titles make a graphical abstract mandatory (1200 x 1200 pixels, 300 DPI). Nature and Science do not require one, but a well-designed abstract graphic increases discoverability and can improve social sharing of your published paper.

Q: Can AI tools be used to generate figures for journal submission? A: AI tools are well suited to early-stage concept generation, schematic drafting, and educational illustrations. Before final submission, verify that every element is scientifically accurate and consider finishing the figure in professional vector software to meet exact journal specifications. Follow your journal's disclosure policy: Nature's guidance on AI is available at nature.com/nature-portfolio/editorial-policies/ai.

Q: What file format is best for a figure that combines photographs and text? A: During production, work in a layered format (PSD or AI) so that raster and vector elements remain editable independently. For submission, export to PDF to preserve vector quality alongside any embedded images, or to TIFF if your journal requires it. Confirm the journal's format preference in their author instructions.

Q: My figure contains too much information. How should I handle it? A: Distribute content across additional figures or move secondary data to supplementary panels. Strip out any element that does not directly serve the figure's central message. Ample white space is a feature, not a flaw: it reduces visual complexity and directs the reader's attention to what matters most.