Barplot to plot frequency of DNA in different sequence lengths

Question

I have this dataframe df below read as df<- read.table("WT1.txt", header= TRUE) . I want to plot the histogram labelling ACGT frequency for each length value. Is there a better way to plot this?

df

 length      A      C      G       T
     17  95668  73186 162726  730847
     18 187013  88641 120631  334695
     19 146061 373719 152215  303973
     20 249897  73862 115441  343179
     21 219899  82356 109536  636704
     22 226368 101499 111974 1591106
     23 188187 112155  98002 1437280

Answer 1

You could melt the data frame into long format by the variable length and plot a stacked bar plot with ggplot2 :

df <- read.table(text=
    "length      A      C      G       T
     17  95668  73186 162726  730847
     18 187013  88641 120631  334695
     19 146061 373719 152215  303973
     20 249897  73862 115441  343179
     21 219899  82356 109536  636704
     22 226368 101499 111974 1591106
     23 188187 112155  98002 1437280", header=T)
library(reshape2)
df <- melt(df, id.vars = "length")
library(ggplot2)
ggplot(df)+
  geom_bar(aes(x=length, y=value, fill=variable), stat="identity")

Answer 2

Use dplyr to calculate frequency for each base and ggplot2 to plot bar plot. I prefer using stat = "identity", position = "dodge" instead of only stat = "identity" as it gives better sense what data looks like.

library(tidyverse)

gather(df, Base, value, -length) %>%
    group_by(length) %>%
    mutate(frequency = value / sum(value)) %>%
    ggplot(aes(factor(length), y = frequency, fill = Base))+
        geom_bar(stat = "identity", position = "dodge",
                 color = "black", width = 0.6) +
        labs(x = "Base pairs", 
             y = "Frequency",
             fill = "Base") +
        scale_y_continuous(limits = c(0, 1)) +
        scale_fill_brewer(palette = "Set1") +
        theme_classic()