Hello,
I don't know if I can ask this here but I have a small problem with R Studio box plot creation, maybe it's a silly question but anyways.
I'm working with different lines of plants that have different values of over expression of a gene (this really doesn't matter a lot but only giving the context) so my data set is organized the following way:
Lines OverExpression levels
1P 1233, 12232, 3268
2P. etc.
3P
etc.
The problem is that I want to create a boxplot from this data, like for exemple all the values from line 1P regrouped in a box plot.
By using the function boxplot() R Studio tells me that X values aren't numerical which is right, they are line names. My questions is the following, how can I arrange my data in order to make boxplot from each line with the different values (and of course all in 1 graph).
Thank you very much and sorry for my poor English,
Here's an example of taking "wide" format data (which it sounds like you have) and reshaping it to produce something that's easier for either base graphics boxplot() or ggplot2 to work with. (This is also an example of how you can create a reproducible example data set for other people to work with, which helps clarify your question ).
library(tidyverse)
# Create some imaginary sample data
set.seed(42) # to make this reproducible
gene_overexpr <- data.frame(
line = c("1P", "2P", "3P"),
sample_1 = sample.int(12000, size = 3),
sample_2 = sample.int(12000, size = 3),
sample_3 = sample.int(12000, size = 3),
sample_4 = sample.int(12000, size = 3),
sample_5 = sample.int(12000, size = 3)
)
gene_overexpr
#> line sample_1 sample_2 sample_3 sample_4 sample_5
#> 1 1P 10978 9966 8840 8461 11217
#> 2 2P 11244 7701 1616 5493 3065
#> 3 3P 3434 6229 7883 8628 5547
# Note that `line` is a factor (default for character values when using `data.frame`)
# In this case, that makes good sense since `line` is a categorical variable
str(gene_overexpr)
#> 'data.frame': 3 obs. of 6 variables:
#> $ line : Factor w/ 3 levels "1P","2P","3P": 1 2 3
#> $ sample_1: int 10978 11244 3434
#> $ sample_2: int 9966 7701 6229
#> $ sample_3: int 8840 1616 7883
#> $ sample_4: int 8461 5493 8628
#> $ sample_5: int 11217 3065 5547
# Convert to "long" format using `tidyr`
gene_overexpr_long <- gene_overexpr %>%
gather(
# names for new columns
key = "sample", value = "overexpr_level",
# gather all the columns *other than* `line`
-line
)
gene_overexpr_long
#> line sample overexpr_level
#> 1 1P sample_1 10978
#> 2 2P sample_1 11244
#> 3 3P sample_1 3434
#> 4 1P sample_2 9966
#> 5 2P sample_2 7701
#> 6 3P sample_2 6229
#> 7 1P sample_3 8840
#> 8 2P sample_3 1616
#> 9 3P sample_3 7883
#> 10 1P sample_4 8461
#> 11 2P sample_4 5493
#> 12 3P sample_4 8628
#> 13 1P sample_5 11217
#> 14 2P sample_5 3065
#> 15 3P sample_5 5547
# Boxplot using base graphics. The formula should be: y ~ group
boxplot(
overexpr_level ~ line, data = gene_overexpr_long,
xlab = "Line", ylab = "Overexpression Level"
)
# Boxplot using `ggplot2`
ggplot(gene_overexpr_long, aes(x = line, y = overexpr_level)) +
geom_boxplot() +
labs(x = "Line", y = "Overexpression Level")
).
