Bio Data Visualization Ggplot2 Fundamentals

library(ggplot2)

# data + aes + geom is the minimum
ggplot(df, aes(x = condition, y = expression)) +
    geom_boxplot() +
    geom_jitter(width = 0.2, alpha = 0.5) +
    # scales
    scale_y_continuous(trans = 'log10', labels = scales::label_log()) +
    scale_color_manual(values = c('#0072B2', '#D55E00')) +
    # labels
    labs(x = NULL, y = 'Expression (log10)',
         title = 'Gene X across conditions',
         caption = 'Source: ...') +
    # facets
    facet_wrap(~ tissue, ncol = 3, scales = 'free_y') +
    # theme
    theme_classic(base_size = 10) +
    theme(panel.grid = element_blank(),
          strip.background = element_blank(),
          strip.text = element_text(face = 'bold'))

geom_point(alpha = 0.7, size = 1, rasterize = TRUE)   # rasterize: ggplot2 3.5+ inline OR ggrastr::rasterize()
geom_line(linewidth = 0.5)                             # linewidth replaces size for lines (ggplot2 3.4+)
geom_col()                                              # bar with y values (use this; geom_bar(stat='identity') is older)
geom_bar()                                              # bar with counts
geom_boxplot(outlier.shape = NA)                       # always suppress when overlaying jitter
geom_violin(bw = 'SJ', trim = FALSE)                   # Sheather-Jones bandwidth; show full tails
geom_histogram(bins = 30)                              # bins NOT binwidth for control
geom_density(alpha = 0.5)
geom_tile(aes(fill = z))                               # heatmap building block
geom_text(aes(label = label), check_overlap = TRUE)
geom_text_repel(aes(label = label), max.overlaps = Inf)   # ggrepel; max.overlaps = Inf prevents silent label drops

aes(x, y, color, fill, shape, size, alpha, linetype, linewidth, group)

# Color vs fill: color = stroke; fill = interior (boxplot, bar, area, polygon)
# Use both when needed: geom_point(aes(color = group, fill = group), shape = 21)

geom_point(color = 'red')             # constant: every point red
geom_point(aes(color = group))        # mapping: color varies with group

# Continuous
scale_x_continuous(limits = c(0, 10), breaks = seq(0, 10, 2),
                    labels = scales::label_number(scale = 1e-6, suffix = 'M'))
scale_y_log10()
scale_y_continuous(trans = 'sqrt')

# Discrete
scale_x_discrete(limits = c('Control', 'Treatment', 'Vehicle'))   # explicit order
scale_color_manual(values = c(Control = '#0072B2', Treatment = '#D55E00'))

# Colormap (sequential, diverging, cyclic) -- see color-palettes
scale_color_viridis_c(option = 'viridis')
scale_color_scico(palette = 'batlow')                              # Crameri
scale_fill_gradient2(low = '#0072B2', mid = 'white', high = '#D55E00', midpoint = 0)

# Date / time
scale_x_date(date_breaks = '1 year', date_labels = '%Y')

facet_wrap(~ var, ncol = 3, scales = 'free_y')
facet_grid(rows = vars(condition), cols = vars(timepoint), scales = 'free_x')
facet_grid(condition ~ timepoint)                                  # formula syntax

# Publication baseline
theme_pub <- theme_classic(base_size = 10) +
    theme(
        panel.grid = element_blank(),
        axis.text = element_text(color = 'black'),
        axis.ticks = element_line(color = 'black', linewidth = 0.3),
        axis.line = element_line(color = 'black', linewidth = 0.3),
        legend.position = 'right',
        legend.key.size = unit(0.4, 'cm'),
        strip.background = element_blank(),
        strip.text = element_text(face = 'bold', size = 9),
        plot.title = element_text(face = 'bold', size = 11),
        plot.tag = element_text(face = 'bold', size = 11))

# Save as a function for re-use across project

# Pass variable name as a string
plot_var <- function(df, x_var, y_var) {
    ggplot(df, aes(x = .data[[x_var]], y = .data[[y_var]])) +
        geom_point()
}
plot_var(df, 'PC1', 'PC2')

# Alternative: bare names via embracing
plot_var2 <- function(df, x_var, y_var) {
    ggplot(df, aes(x = {{ x_var }}, y = {{ y_var }})) +
        geom_point()
}
plot_var2(df, PC1, PC2)

library(ggtext)
ggplot(df, aes(x, y)) + geom_point() +
    labs(x = 'log<sub>2</sub> fold change',
         y = '\\u2212log<sub>10</sub>(*p*)') +
    theme(axis.title.x = element_markdown(),
          axis.title.y = element_markdown())

# cairo_pdf for TrueType embedded; portable across systems
ggsave('figure.pdf', plot = p,
       width = 89, height = 70, units = 'mm',
       device = cairo_pdf)

# Vector + raster mix via ggrastr (for large scatter)
library(ggrastr)
ggplot(df, aes(x, y)) +
    rasterise(geom_point(alpha = 0.5), dpi = 300) +
    theme_pub
ggsave('out.pdf', device = cairo_pdf)

# PNG for raster
ggsave('figure.png', p, width = 89, height = 70, units = 'mm', dpi = 300)

# TIFF for some journals
ggsave('figure.tiff', p, width = 89, height = 70, units = 'mm', dpi = 300,
       compression = 'lzw')