Analysis in R: PCA results in ggplot2 displayed in Biplot “ggbiplot” package

This is an introduction to the ggplot2 package for displaying the results of a principal component analysis in Biplot.

Package version is 0.55. Checked with R version 4.2.2.

＜おすすめのRに関する書籍です＞

生命科学・生物工学のための間違いから学ぶ実践統計解析: R・Pythonによるデータ処理事始め (New paradigm! On-demand publishing) | 川瀬雅也, 松田史生

Amazonで川瀬雅也, 松田史生の生命科学・生物工学のための間違いから学ぶ実践統計解析: R・Pythonによるデータ処理事始め (New paradigm! On-demand publishing)。アマゾンならポイント還元本が多数。

www.amazon.co.jp

Install Package

Run the following command.

#Install Package
install.packages("devtools")
devtools::install_github("vqv/ggbiplot")

Example

See the command and package help for details.

#Loading the library
library("ggbiplot")

###Creating Data#####
set.seed(1234)
TestData <- matrix(rnorm(1000), 200)
TestData <- as.data.frame(TestData)
#Data Adjustment
TestData[1:67,] <- TestData[1:67,] + 1
TestData[68:135,] <- TestData[68:135,] + 2
TestData[136:200,] <- TestData[136:200,] + 3
#Assign group information
TestData <- cbind(c(rep("Group1", 67),
                    rep("Group2", 67),
                    rep("Group3", 66)),
                  TestData)
colnames(TestData) <- c("Group", paste0("ColName", seq(5)))
########

#Package stat: Principal component analysis with prcomp command
TestPrc <- prcomp(TestData[, 2:6], scale. = FALSE)

#Creating a biplot: ggbiplot command
#Specifies the result of prcomp() or princomp(): pcobj option
#Specify principal components to plot: choices option
#Specify group information: groups option
#Draw probability ellipses for each group: ellipse option
#Plot correlated circles: circle option
#The ggplot2 command is available
ggbiplot(pcobj = TestPrc, choices = 1:2, obs.scale = 1, var.scale = 1,
         groups = TestData[, 1], ellipse = TRUE, circle = TRUE) +
  scale_colour_manual(values = c("#FF0000", "black", "#00FF00")) +
  theme(legend.direction = "horizontal", legend.position = "top")