Plot DNA nucleotide data using loop and PDF graphics in R

user_18486979 0
Viewed 61

My boss asked me to plot a matrix of DNA nucleotides using the pdf graphics function in R. I have a bit of code I'm working with but I can't figure it out and have spent way too much time trying! I know there are likely other methods/packages out there to visualize this genetic data, and I am absolutely interested in hearing them, but I also need to do this the way it was assigned to me.

I have sequence data in R like so:

> head(b)

Sequence                            X236 X237 X238 X239 X240 X241 X242 X244 X246 X247 X248 X249 X250 X251 X252 X253 X254 X255 X256 X257 X258 X259
1    L19088.1                         G    G    G    G    G    A    G    A    C    C    A    A    G    A    T    G    G    C    C    G    A    A   
2    chr1_43580199_43586187           ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    ·    g    g

There are a total of 1040 rows and 483 columns, with character possibilities A, a, G, g, T, t, C, c, mid-dot, or X.

I want to color the different characters and plot them in a way that is similar to a heatmap. The dots and the Xs don't need to be colored. The code I am working with so far is:

pdf( 
  sprintf( 
    "%s/L1.pdf", 
    out_dir),
  width = 8.5, height = 11 )
par(omi = rep(0.5,4))
par(mai = rep(0.5,4))
par(bg  = "#eeeeee")
plot( NULL, 
      xlim = c(1,100), ylim = c(1,140), 
      xlab = NA,       ylab = NA, 
      xaxt = "n",      yaxt = "n",
      bty = "n",       asp = 1 )

plot_width <- 100
w <- plot_width / 600


genome_colors <- list()
genome_colors[["A"]] <- "#ea0064"
genome_colors[["a"]] <- "#ea0064"
genome_colors[["C"]] <- "#008a3f"
genome_colors[["c"]] <- "#008a3f"
genome_colors[["G"]] <- "#116eff"
genome_colors[["g"]] <- "#116eff"
genome_colors[["T"]] <- "#cf00dc"
genome_colors[["t"]] <- "#cf00dc"

I <- nrow(b)
J <- ncol(b)
for ( i in 1:I ){
 for ( j in i:J ){
 # plot nucleotide as rectangle with color and text label, something like:

 # plot nucleotides with genome_colors
 # rect( (j-1)*w, top-(i-1)*w, j*w, top-i*w, col = color, border = NA )
 }
 # text( (j+1)*w, top-(i-1)*w, labels = i, cex = 0.05, col = "#dddddd" )
}
dev.off()

If anyone can help me with the plotting loop or point me in a helpful direction I will be so thankful!

heatmap dna-sequence r for-loop pdf
1 Answers

assuming df is your dataframe in wide format (one column per position, one row per sequence), example:

df <- structure(list(sequence = c("L19088.1", "chr1_43580199_43586187"
), X236 = c("G", "."), X237 = c("G", "."), X238 = c("A", "a"), 
    X239 = c("T", "C"), X240 = c("A", "c"), X241 = c("G", "G"
    )), class = "data.frame", row.names = 1:2)

## > df
##                 sequence X236 X237 X238 X239 X240 X241
## 1               L19088.1    G    G    A    T    A    G
## 2 chr1_43580199_43586187    .    .    a    C    c    G

... you can use packages ggplot2 and tidyr from the tidyverse like so:

library(tidyr)
library(ggplot2)

df %>%
  ## reshape to long table
  ## (one column each for sequence, position and nucleotide):
  pivot_longer(-sequence, ## stack all columns *except* sequence
               names_to = 'position',
               values_to = 'nucleotide'
               ) %>%
  ## create the plot:
  ggplot() +
  geom_tile(aes(x = position, y = sequence, fill = nucleotide),
              height = .9 ## adjust to visually separate sequences
            ) +
  scale_fill_manual(values = c('A'='#ea0064', 'a'='#ea0064', 'C'='#008a3f',
                              'c'='#008a3f', 'G'='#116eff', 'g'='#116eff',
                              'T'='#cf00dc', 't'='#cf00dc', '.'='#a0a0a0'
                              )
                    ) +
  labs(x = 'x-axis-title', y='y-axis-title') +
  ## remove x-axis (=position) elements: they'll probably be too dense:
  theme(axis.title.x = element_blank(),
        axis.text.x = element_blank(),
        axis.ticks.x = element_blank()
        )

^^^ for easy styling see e.g. ggplot themes

save plot with convenience wrapper ggsave:

  ggsave(filename = 'my_plot.pdf',
       width = 12, ## inches; to fill DIN A4 landscape
       height = 8
       )

When using the pdf() function, don't forget to explicitly print your plot:

pdf(file = 'my_plot.pdf',
    ## ... other parameters
)
print( ## you need to print the plot
  qplot(data = cars, x = speed, y = dist, geom = 'point')
)
dev.off()
user_183097110
Related