generate-publication-plots - SKILL.md Agent Skill

name: generate-publication-plots

description: Create publication-quality figures using ggplot2 and other R visualization libraries, with support for multiple export formats and journal-specific formatting guidelines.

Generate Publication-Quality Plots

A specialized sub-skill for creating professional, publication-ready figures in R that meet journal standards and scientific visualization best practices.

Overview

This sub-skill generates high-quality figures using ggplot2 and other R visualization libraries. It supports multiple export formats (PDF, PNG, SVG, TIFF) and follows common journal formatting guidelines including resolution, color space, and font specifications.

Use this sub-skill when the user wants to:

Create publication-quality figures from data
Export plots in specific formats for journals
Generate multi-panel composite figures
Apply journal-specific formatting guidelines
Create consistent figure styles across publications

What This Sub-Skill Does

When invoked, this sub-skill will:

Analyze data and requirements
- Examine data structure and variables
- Determine appropriate plot type
- Identify journal formatting requirements
- Check for categorical vs continuous variables
Create ggplot2 visualizations
- Apply theme customization
- Use color-blind friendly palettes
- Set appropriate figure dimensions
- Add proper annotations and labels
Export to publication formats
- PDF (vector format for journals)
- PNG (raster, 300-600 DPI)
- SVG (scalable vector graphics)
- TIFF (high-quality raster)
Follow journal guidelines
- Nature (single column, 89mm or 183mm width)
- Science (specific dimensions)
- PLOS ONE (300+ DPI TIFF/PNG)
- IEEE (vector preferred)

Supported Plot Types

| Plot Type | Description | Use Case |

|-----------|-------------|----------|

| Scatter plots | Point clouds with regression lines | Correlation analysis |

| Bar plots | Grouped, stacked, or error bars | Group comparisons |

| Box/Violin plots | Distribution visualization | Statistical summaries |

| Line plots | Time series, trends | Longitudinal data |

| Heatmaps | Color-coded matrices | Expression data |

| Volcano plots | -log10(p) vs fold change | Differential expression |

| PCA plots | Principal components | Dimensionality reduction |

| Forest plots | Effect sizes | Meta-analysis |

| Survival curves | Kaplan-Meier | Time-to-event data |

| Multi-panel figures | Combined plots | Complex results |

Example User Requests

"Create a publication-quality scatter plot with regression line"
"Generate a volcano plot for my differential expression results"
"Make a bar plot with error bars for Nature journal format"
"Export this plot as 600 DPI PNG and vector PDF"
"Create a multi-panel figure combining PCA and heatmap"

Journal Formatting Guidelines

Nature

Single column: 89mm width
Double column: 183mm width
Preferred: PDF, EPS, SVG (vector)
Minimum: 600 DPI for raster
Fonts: Arial, Helvetica, or Symbol

Science

Width: 5.5cm (single) or 11.4cm (double)
Height: Maximum 22.5cm
Preferred: TIFF, EPS
Resolution: 300-600 DPI
Fonts: Helvetica or Arial

PLOS ONE

Width: 150-300 pixels per inch
Preferred: TIFF, PNG, PDF
Resolution: Minimum 300 DPI
Color: RGB or CMYK accepted

IEEE

Column width: 3.5 inches (single column)
Page width: 7.5 inches (double column)
Preferred: PDF, EPS
Fonts: Times New Roman, Helvetica, Arial

Color Palettes

Color-blind Friendly Palettes


# Okabe-Ito palette (8 colors)

c("#E69F00", "#56B4E9", "#009E73", "#F0E442",

  "#0072B2", "#D55E00", "#CC79A7", "#999999")



# Viridis (perceptually uniform)

viridis::viridis(8)



# ColorBrewer palettes

RColorBrewer::brewer.pal(8, "Set2")

Diverging Palettes

Red-Blue (for heatmaps, volcano plots)
RdYlBu, RdBu, BrBG

Sequential Palettes

Blues, Greens, PuBuGn

Template Functions

Base Publication Theme


theme_publication <- function(base_size = 12, base_family = "Arial") {

  theme_bw(base_size = base_size, base_family = base_family) +

  theme(

    legend.position = "top",

    panel.border = element_rect(size = 0.5),

    panel.grid.major = element_line(color = "grey90"),

    panel.grid.minor = element_blank(),

    axis.text = element_text(size = rel(0.8)),

    legend.key = element_rect(color = "grey90"),

    legend.key.size = unit(0.4, "cm"),

    legend.margin = margin(0, 0, 0, 0, "cm")

  )

}

Export Function


save_publication_figure <- function(plot, filename,

                                    width = 89, height = 89,

                                    dpi = 600, units = "mm") {

  ggsave(filename, plot = plot,

         width = width, height = height,

         dpi = dpi, units = units,

         device = "pdf")

}

Parameters

| Parameter | Description | Default |

|-----------|-------------|---------|

| plot_type | Type of plot to create | Required |

| data | Data frame for plotting | Required |

| x_var | Variable for x-axis | Required |

| y_var | Variable for y-axis | Required |

| journal | Target journal format | generic |

| output_format | Export format (pdf, png, svg, tiff) | pdf |

| width | Figure width in mm | 89 |

| height | Figure height in mm | 89 |

| dpi | Resolution for raster formats | 600 |

| color_palette | Color scheme | Okabe-Ito |

Example Workflows

Simple Scatter Plot


# Input data: df with columns x, y, group

# Output: scatter_with_regression.pdf



p <- ggplot(df, aes(x = x, y = y, color = group)) +

  geom_point(size = 3, alpha = 0.7) +

  geom_smooth(method = "lm", se = TRUE) +

  scale_color_okabe_ito() +

  theme_publication() +

  labs(x = "X Variable", y = "Y Variable")



save_publication_figure(p, "scatter_with_regression.pdf",

                       width = 89, height = 89, dpi = 600)

Volcano Plot


# Input data: de_results with log2FC, pvalue

# Output: volcano.pdf



p <- ggplot(de_results, aes(x = log2FC, y = -log10(pvalue))) +

  geom_point(aes(color = significant), alpha = 0.6) +

  scale_color_manual(values = c("grey", "#D55E00", "#0072B2")) +

  theme_publication() +

  geom_vline(xintercept = c(-1, 1), linetype = "dashed") +

  geom_hline(yintercept = -log10(0.05), linetype = "dashed") +

  labs(x = expression(log[2]~Fold~Change),

       y = expression(-log[10]~p~value))



save_publication_figure(p, "volcano.pdf")

Multi-Panel Figure


# Combine multiple plots using patchwork

library(patchwork)



combined <- (p1 | p2) / (p3 | p4)



ggsave("multipanel.pdf", combined,

       width = 183, height = 183,

       dpi = 600)

Output Specifications

PDF (Vector)

Recommended for journals
Scalable without quality loss
File extension: .pdf

PNG (Raster)

300-600 DPI
Good for web/PPT
File extension: .png

SVG (Vector)

Web-friendly vector
Editable in Inkscape/Illustrator
File extension: .svg

TIFF (Raster)

High-quality for publication
Lossless compression
File extension: .tiff

Font Options

| Font | Usage | Availability |

|------|-------|--------------|

| Arial | Most common | System font |

| Helvetica | Classic | System font |

| Times New Roman | Traditional | System font |

| CM Roman | LaTeX-style | Extrafont package |

| Latin Modern | LaTeX-style | Extrafont package |

Best Practices

Use vector formats when possible - PDF for publication
Check color blindness - Use color-blind friendly palettes
Maintain aspect ratio - Avoid distorted figures
Label clearly - Self-explanatory axis labels and legends
Keep it simple - Remove chart junk and unnecessary decorations
Use consistent style - Same theme across all figures in paper
Export at correct DPI - 600 for most journals

Notes

Color-blind friendly palettes used by default
Font family can be customized per journal requirements
Multi-panel figures use patchwork package
All plots use publication-quality theme by default
Resolution and dimensions match common journal standards

Related Sub-Skills

create-project: Set up projects before creating plots
run-analysis: Generate analysis results for plotting