boxplot with jitter

Question

[archana.bioinfo87]

Hi,

I am trying to make a box plot with the following codes.

 NDVI_ts <- read.table("merge_2_NC.csv", header = TRUE)
NDVI_ts
library(ggplot2)
library(scales)
library(tidyverse)
library(ggpubr)
NormalvsCancer<-interaction(NDVI_ts$Data, sep="\t")

######Outfile name as input file
pdf("boxplot.pdf")
##########
xlabs <- paste(levels(NDVI_ts$Data),"\n(N=",table(NDVI_ts$Data),")",sep="")
#ggplot(df,aes(x=group,y=x,color=group))+geom_boxplot()+scale_x_discrete(labels=xlabs)
p2 = ggplot(NDVI_ts, aes(x=NormalvsCancer, y=Read_count)) +
  geom_point(aes(fill=NormalvsCancer), size=5, shape=21, colour="grey20",
             position=position_jitter(width=0.2, height=0.1)) +
  geom_boxplot(aes(fill = Data), width = 0.6, outlier.colour=NA, fill=NA) + 
  scale_x_discrete(labels=xlabs) +
  theme_bw() + theme(axis.text.x = element_text(angle = 360, hjust = 1)) +
  stat_compare_means(aes(group = Data), label = "p.format")
print(p2)
dev.off()

I am getting the output image like this:

Everything is ok except in bladder_normal(19). Here N=0, whereas it is showing 1.

The csv input file is this.

Data    No_matched  Read_count
Bladder_tumor(414)  1   2
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   10
Bladder_tumor(414)  1   24
Bladder_tumor(414)  1   3
Bladder_tumor(414)  1   8
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   2
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   2
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_tumor(414)  1   1
Bladder_normal(19)  0   0

Any help is much appreciated.

Thanks

[ATpoint]

Changed the link to display the image properly. What is the actual question now, I do not really get it?

[Ram]

OP is bypassing factors using a logic doesn't make sense to me. OP, instead of

NDVI_ts <- read.table("merge_2_NC.csv", header = TRUE)
..
..
xlabs <- paste(levels(NDVI_ts$Data),"\n(N=",table(NDVI_ts$Data),")",sep="")

try

NDVI_ts <- read.table("merge_2_NC.csv", header = TRUE, stringsAsFactors = FALSE)
..
..
xlabs <- paste(unique(NDVI_ts$Data),"\n(N=",table(NDVI_ts$Data),")",sep="")

Even better, split the first column into two columns (category and count) so you're dealing with atomic data.

[archana.bioinfo87]

Ok, thanks. I will try this.

[archana.bioinfo87]

I want to get the count of normal and tumor samples as N. If the number of matched with N is "0" then it should get reflected that Bladder normal(19) N=0 and tumor N=17.

[Asaf]

I think you have the header wrong, shouldn't it be NormalvsCancer. Anyhow, N=1 means one sample

[archana.bioinfo87]

Actually, there is no normal sample but I want to show the number of normal sample is "0".

[Ram]

Your xlabs don't match up with the actual data. Split the first column so you have a clean data.frame and build the axis labels in a better manner.

[Asaf]

I got lost. table(NDVI_ts$Data) shouldn't return 1 for Bladder_normal(19)?

[ATpoint]

It does and should because table does simply quantifies characters :)

[Asaf]

I don't understand the use of NormalvsCancer<-interaction(NDVI_ts$Data, sep="\t") Why don't you just use Data? interaction function returns unordered values and is irrelevant here.

[archana.bioinfo87]

Thanks for pointing it out. I will correct it.

Answer 1

[ravipatel4]

Change your code that creates xlabs to following. It should fix your problem:

xlabs <- paste(levels(NDVI_ts$Data),"\n(N=",

table(NDVI_ts[ NDVI_ts$No_matched!=0 ,]$Data),")",sep="")

[archana.bioinfo87]

Thank you Ravi finally solved without much modification.

[Ram]

Be aware that this will only work when NDVI_ys$Data is a factor. For character columns, levels(df$col_name) will be NULL and table(subset_of_df) will not include values in a column that are not part of that subset. In your case, Bladder_normal(19) will be excluded from the table if Data is a character column.

Answer 2

[ATpoint]

Don't use table the way you do it as this simply counts presence of strings or characters regardless of the No_matched column. Do something like sum(as.numeric(NDVI_ts $No_matched)) to use that column and its numeric content.