R Single Cell Plotting

R Single-Cell Plotting Conventions

Seurat / UMAP Plots

cluster_colors <- unname(pals::polychrome())

# UMAP — always square, no raster
p <- DimPlot(obj, reduction = "umap", group.by = "seurat_clusters",
             order = TRUE, repel = TRUE, cols = cluster_colors, raster = FALSE) +
  theme_bw() +
  theme(aspect.ratio = 1)
print(p)
ggsave("umap.png", p, width = 8, height = 8, dpi = 300, bg = "white")

# Feature plot
p <- FeaturePlot(obj, features = "GENE", raster = FALSE) +
  theme(aspect.ratio = 1)

# Violin (no points)
p <- VlnPlot(obj, features = "nFeature_RNA", group.by = "Sample", pt.size = 0)

# DotPlot — genes must be unique
p <- DotPlot(obj, features = unique(genes), group.by = "cell_type") + coord_flip()

Key rules:

• Always theme(aspect.ratio = 1) for UMAP/dim reductions
• Always raster = FALSE (prevents pixel artifacts in PDF/HTML)
• Always order = TRUE, repel = TRUE on DimPlot
• Cluster colors: unname(pals::polychrome())
• Always print() plots in Rmd chunks
• Save at 300 DPI PNG

Volcano Plots

library(EnhancedVolcano)
EnhancedVolcano(degs,
  lab = degs$gene,
  x = "avg_log2FC",
  y = "p_val_adj",
  pCutoff = 0.05,
  FCcutoff = 0.25)

Statistical Comparisons (publication-ready)

# Wilcoxon + Bonferroni
stat_df |>
  rstatix::wilcox_test(value ~ group) |>
  rstatix::adjust_pvalue(method = "bonferroni") |>
  rstatix::add_significance()

# Annotate ggplot directly
p + ggpubr::stat_compare_means(
  comparisons = list(c("GroupA", "GroupB")),
  method = "wilcox.test"
)

General ggplot

# Save any plot at 300 DPI
ggsave("plot.png", p, width = 8, height = 6, dpi = 300, bg = "white")

# Theme for publications
p + theme_bw() + theme(
  text = element_text(size = 12),
  axis.text = element_text(size = 10)
)