boxplot using ggplot2 in R

Question

[raavi21198]

Hello members!!

I have a raw data which consists of 45 samples and their intensities. This is a microarray data expression. I have comverted this into a dataframe. However, I am confused how to plot a boxplot of all these 45 samples and also group them as "normal" and "tumor". Please help me out with this The code i used is as follows

read_data <- ReadAffy() ##read the raw .CEL files

ph=read_data@phenoData#annotation of the data

ph$sample
ph@data
ph@data[,1]=c("NB","ND","TB","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB","TC","NB","ND","TB")

sampleNames=vector()
logs=vector()
for (i in 1:45) 
{
  sampleNames=c(sampleNames,rep(ph@data[i,1],dim(pmexp)[1]))
  logs=c(logs,log2(pmexp[,i]))
}
logdata <- data.frame(logint=logs,sampleName=sampleNames)

the structure of this dataframe is as follows

    > str(logdata)
'data.frame':   11155455 obs. of  2 variables:
 $ logint    : num  8.79 9.74 11.09 12.38 12.36 ...
 $ sampleName: chr  "NB" "NB" "NB" "NB" ...
> head(logdata)
     logint sampleName
1  8.791163         NB
2  9.736402         NB
3 11.091435         NB
4 12.376125         NB
5 12.363587         NB
6 11.574594         NB
> p <- ggplot(logdata,aes(sampleName,logint))
> p+geom_boxplot()

Can someone please guide me how to create a boxplot using ggplot2 in R, of these 45 samples, by grouping them as normal and tumor samples, as the above code gives me the boxplot of only four samples. I need to print them all together

Thank you

[cpad0112]

in the absence of data, i suggest following:

1. Convert the data frame from wide format to long format. (dplyr/tidyr)
2. Attach grouping information for each sample (dplyr)
3. Draw box plot (ggplot)
4. Add jitter (ggplot)
5. Facet by group (ggplot)

Instead of boxplot, consider using violin plot with jitter.

[raavi21198]

Thank you for your response. I have edited to repost the data. Could you now let me know where am i going wrong

Answer 1

[cpad0112]

here is an example i built from https://bioconductor.org/packages/devel/workflows/vignettes/arrays/inst/doc/arrays.html

library(affy)   # Affymetrix pre-processing
library(limma)  # two-color pre-processing; differential
phenoData <- read.AnnotatedDataFrame(system.file("extdata", "pdata.txt", package="arrays"))
celfiles <- system.file("extdata", package="arrays")
eset <- justRMA(phenoData=phenoData,celfile.path=celfiles)
df=as.data.frame(exprs(eset))
pdata=pData(eset)

library(dplyr)
library(tidyr)
library(tibble)
library(ggplot2)


df %>% 
    pivot_longer(everything(),names_to = "cels", values_to ="vals") %>% 
    inner_join(., rownames_to_column(pdata),by = c("cels" = "rowname")) %>% 
    ggplot(., aes(cels,vals, fill=Sensitivity)) +
    geom_boxplot()+
    facet_wrap(~IVT, scales = "free")+
    xlab("")+
    ylab("")+
    theme_bw()+
    theme(axis.text.x = element_text(angle = 90),
          axis.text = element_text(size=18),
          strip.text = element_text(size=18),
          legend.text = element_text(size=18),
          legend.title = element_text(size = 18)
          )

[ATpoint]

Code suggestions:

a) use theme_set() to both define a theme and set a base size for all relevant parts (axis, theme, labels) in a single command, that saves the multiple arguments in theme().

b) rotate x-axis labels with guides rather than angle as guides ensures proper alignment in horizontal and vertical directions even using angles such as 45°, see here, and

c) put legend on top so its large size does not shrink the plot itself. Again, the sizes of all fonts and labels are auto-adjusted to look decent based on the base_size in the theme_set() command on top.

theme_set(theme_bw(base_size = 15))
df %>% 
  pivot_longer(everything(),names_to = "cels", values_to ="vals") %>% 
  inner_join(., rownames_to_column(pdata),by = c("cels" = "rowname")) %>% 
  ggplot(., aes(cels,vals, fill=Sensitivity)) +
  geom_boxplot()+
  facet_wrap(~IVT, scales = "free")+
  xlab("")+
  ylab("")+
  guides(x = guide_axis(angle = 45))+
  theme(legend.position="top")

By the way, the code example you use requires the arrays package to be installed to have access to their extdata, BiocManager::install("arrays").

[raavi21198]

Thank you a so much for proving this example