Pivot a wide data frame and plot with facets.

boxplots_for_B.R

library(tidyverse)

#### MAKING MOCK DATA ####

# Some parameters for amount of data.
n_metabs <- 5
n_tissues <- 10
tissues <- purrr::map_chr(1:n_tissues, ~ paste0(sample(letters, 3), collapse=""))

# Make a wide table of data, adding on columns per metabolite
dfs <- tibble(tissue = tissues)
for (metab in LETTERS[1:n_metabs]) {
    df <- tibble(vals = rnorm(n_tissues))
    colnames(df) <- metab
    dfs <- bind_cols(dfs, df)
}

# Add sensitive and insensitive labels.
dfs$sensitive <- c(rep("sens", n_tissues/2), rep("insens", n_tissues/2))

# Show final data frame.
dfs

#> # A tibble: 10 x 7
#>    tissue      A      B      C       D       E sensitive
#>    <chr>   <dbl>  <dbl>  <dbl>   <dbl>   <dbl> <chr>    
#>  1 edi     1.35   1.89   0.712  2.86    0.251  sens     
#>  2 vxo     0.958  0.844  0.743 -0.915   2.34   sens     
#>  3 cji    -0.596  1.00   0.583  0.0501 -1.35   sens     
#>  4 oit     1.36  -0.626  1.49  -0.900   0.537  sens     
#>  5 tdj     1.47  -0.383 -2.59   1.22   -1.92   sens     
#>  6 dru    -0.841  2.11  -0.555  0.705   0.512  insens   
#>  7 rhc    -0.231  0.585  0.405  0.729   1.36   insens   
#>  8 eqm    -0.171  0.660  0.741 -0.241  -0.165  insens   
#>  9 sop     0.115 -0.175 -0.601 -0.519   0.0214 insens   
#> 10 fvo     1.03   0.984 -0.839 -0.430  -1.33   insens   




#### PLOTTING ####

# Make a boxplot per metabolite.
dfs %>% 
    pivot_longer(c(-tissue, -sensitive), names_to = "metab", values_to = "values") %>% 
    ggplot(aes(x = sensitive, y = values)) +
    facet_wrap(~ metab) +
    geom_boxplot(outlier.shape = NA) +
    geom_jitter(size = 1, color = "dodgerblue") +
    theme_bw(base_size = 7, base_family = "Arial")

ggsave("metab_boxplots.svg", width = 10, height = 10, unit = "cm")