# Question about the dots on Quartile groups in boxplot

I have Microarray Normalized Expression data for a specific Gene. It looks like below in a dataframe B

SampleID    Gene    Type
Sample1     5.02    Tumor
Sample2     5.06    Tumor
Sample3     5.1     Tumor
Sample4     5.11    Tumor
Sample5    5.127    Normal
Sample6     5.12    Normal
Sample7    5.138    Normal
Sample8    5.149    Normal


I see that the minimum expression value in the table is 5.0 and maximum expression value is around 5.9.

I wanted to show the expression between two conditions with a boxplot and used following code.

q <- ggboxplot(B, x = "Type", y = "Gene",
color = "black", palette = "npg",
add = "jitter", ylab = 'Gene expression', xlab=FALSE,
order=c("Normal", "Tumor"))
q + stat_compare_means(method = "t.test") +
geom_point() +
stat_n_text()


This gave a plot like this

But when I remove the jitter from the code,

q <- ggboxplot(B, x = "Type", y = "Gene",
color = "black", palette = "npg",
ylab = 'Gene expression', xlab=FALSE,
order=c("Normal", "Tumor"))
q + stat_compare_means(method = "t.test") +
geom_point() +
stat_n_text()


I see many black dots in the middle of the boxplot like this

May I know why I see those dots in the second boxplot after removing the jitter. Is there a way to avoid that?

• I don't know from where does the ggboxplot function comes from or stat_compare_means? Also please, continue editing the question to clarify the question. It is still unclear to me if you want to remove some points, add them or something else – llrs Nov 23 '18 at 13:52
• from the package ggpubr – beginner Nov 23 '18 at 14:22
• I am confused by the wording in "May I know why I see those dots in the second boxplot after removing the jitter. Is there a way to avoid that?" Are you trying to ask "Why did the points not disappear after I remove jitter?" – winni2k Nov 23 '18 at 17:46

You add the points with geom_point(). Just remove it and you will get your "empty" boxplot.
q <- ggboxplot(B, x = "Type", y = "Gene",

Unfortunately I couldn't use stat_compare_means(method = "t.test") and stat_n_text(), but this two just add the labels.