R脚本--PCA分析系列1_v1.0

PCA分析系列-基于prcomp函数

摘要

[1. 数据准备](#1. 数据准备)

[2. R脚本](#2. R脚本)

[3. 结果示例](#3. 结果示例)

[3.1 结果结构](#3.1 结果结构)

[3.2 原始矩阵展示](#3.2 原始矩阵展示)

（1）gene_TPM2.txt

（2）group2.txt

[3.3 图展示](#3.3 图展示)

（1）p_ggbiplot.png

（2）p_ggbiplot_lable.png

（3）p_ggplot2.png

（4）p_ggplot2_lable.png

摘要

PCA分析R包有很多，不同的包作图时，如果不注意某些参数设置就会出现：画出图不一致情况。今天主要基于prcomp函数进行主成分分析，然后分别用ggbiplot包和ggplot2包进行可视化。注：如果要展示置信圈，要求每组组内样本数≥4；如果每组组内样本数=3，是画不了置信圈的，只能是画分组椭圆（一般也没必要，意义不大）；如果每组组内样本数＜3，尽量不要自找麻烦弄PCA了。

1. 数据准备

表达矩阵：第一列为基因信息，其余列为样本对应的表达数据

分组信息：第一列为样本，第二列为分组

注：两个表样本要对称一致。

2. R脚本

复制代码

rm(list = ls())

# 加载所需包
library(openxlsx)
library(ggbiplot)
library(ggplot2)
library(ggrepel)
getwd()
# 设置工作目录
setwd("E:/SCRIPTS/PCA") #换成你实际的位置进行

# 读取数据
##如果是xlsx文件
#dat <- read.xlsx("gene_TPM2.xlsx", colNames = TRUE, rowNames = TRUE)
#SampleDat <- read.table("group2.xlsx", colNames = TRUE, rowNames = TRUE)
dat <- read.table("gene_TPM2.txt", sep = '\t', row.names = 1, header = T, stringsAsFactors = F)
SampleDat <- read.table("group2.txt", sep = '\t', header = T)
#View(dat)
#View(SampleDat)
#View(SampleDat)

# 数据处理
#datNew <- dat[, -1] # 前面加载数据时候设置row.names = 1，这里就不用
datNew <- dat
group <- as.character(SampleDat[match(colnames(datNew), SampleDat[, 1]), 2])

#datNewT <- t(as.matrix(datNew))
#datNewT <- na.omit(datNewT)
#建议去掉常数列（全 0 或全相同值，全是0或相同值并不会提供变异值，对结果无贡献）
datNewT <- t(log2(as.matrix(dat) + 1))              # 先 log
datNewT <- datNewT[, apply(datNewT, 2, sd) > 0]     # 再踢常数列

# 计算PCA
#PCADat <- prcomp(datNewT)
#建议进行中心化和标准化处理，让所有基因处在"同一量尺"，使PCA不会被高表达基因绑架。
PCADat <- prcomp(datNewT, scale. = TRUE, center = TRUE)

# 映射分组信息
df_pca <- as.data.frame(PCADat$x)
df_pca$group <- SampleDat$group
#View(df_pca)

# 提取解释度
summ <- summary(PCADat)
#View(summ)
x_lable <- paste0("PC1(", round(summ$importance[2, 1]*100, 2), "%)")
y_lable <- paste0("PC2(", round(summ$importance[2, 2]*100, 2), "%)")


# 创建保存图形的文件夹
if (!dir.exists("PCA_results/ggbiplot_method")) {
  dir.create("PCA_results/ggbiplot_method")
}
if (!dir.exists("PCA_results/ggplot2_method")) {
  dir.create("PCA_results/ggplot2_method")
}


par(family = "Times New Roman")  # 设置全局字体

# 设置主题样式
theme_custom <- theme(
  plot.background = element_rect(fill = "white"),  # 设置背景为白色
  panel.background = element_rect(fill = "white"),  # 设置面板背景为白色
  panel.grid = element_line(color = "lightgray"), # 设置主要网格线颜色
  axis.text = element_text(color = "black", size = 12, face = "plain"), # 设置刻度值字体颜色、大小和粗细
  axis.title = element_text(color = "black", size = 14, face = "bold"), # 设置轴标签字体颜色、大小和斜体
  legend.text = element_text(color = "black", size = 10, face = "bold"), # 设置图例字体颜色、大小和样式
  legend.title = element_text(color = "black", size = 12, face = "bold"), # 设置图例标题字体颜色、大小和样式
  panel.border = element_rect(color = "black", fill = NA, linewidth = 0.5)  # 设置面板边框为黑色
)

################################# 无置信圈 #####################################
# -------------------- ggbiplot 图形 --------------------
# 基础图形
p_ggbiplot <- ggbiplot(PCADat, obs.scale = 1, var.scale = 0, 
                       ellipse = TRUE, circle = FALSE, var.axes = F, groups = group) +
  theme_custom + 
  coord_fixed(ratio = 1.4) #调整x轴和y轴长度比例
print(p_ggbiplot)

# 保存图形
ggsave("PCA_results/ggbiplot_method/p_ggbiplot.png", 
       plot = p_ggbiplot, width = 6, height = 6, dpi = 300)
ggsave("PCA_results/ggbiplot_method/p_ggbiplot.pdf", 
       plot = p_ggbiplot, width = 6, height = 6)

dev.off()

# 添加标签的图形
p_ggbiplot_lable <- p_ggbiplot + 
  geom_text(aes(label = rownames(datNewT)), 
            vjust = 1.5, size = 3, color = "black") + 
  coord_fixed(ratio = 1.4) #调整x轴和y轴长度比例
print(p_ggbiplot_lable)

ggsave("PCA_results/ggbiplot_method/p_ggbiplot_lable.png", 
       plot = p_ggbiplot_lable, width = 6, height = 6, dpi = 300)
ggsave("PCA_results/ggbiplot_method/p_ggbiplot_lable.pdf", 
       plot = p_ggbiplot_lable, width = 6, height = 6)

dev.off()


################################# 有置信圈 #####################################
# -------------------- ggplot2 图形 --------------------
# 基础图形
p_ggplot2 <- ggplot(df_pca, 
                    aes(x = PC1, 
                        y = PC2,
                        colour = group)
                    ) + 
  stat_ellipse(aes(fill = group),
               type = "norm", 
               geom = "polygon", 
               alpha = 0.2, 
               level = 0.68,
               color = NA, 
               linetype = 2, 
               show.legend = FALSE
               ) + 
  geom_point() + 
  labs(x = x_lable, 
       y = y_lable, 
       color = "Group"
       ) + 
  guides(fill = "none") + 
  theme(legend.key = element_rect(colour = "black",  # 黑色边框
                                  size = 0.2  # 边框线宽
                                  )
        ) + 
  theme_custom + 
  coord_fixed(ratio = 1.4) #调整x轴和y轴长度比例
                    
print(p_ggplot2)

# 保存图形
ggsave("PCA_results/ggplot2_method/p_ggplot2.png", 
       plot = p_ggplot2, width = 6, height = 6, dpi = 300)
ggsave("PCA_results/ggplot2_method/p_ggplot2.pdf", 
       plot = p_ggplot2, width = 6, height = 6)

dev.off()

# 添加标签的图形
p_ggplot2_lable <- ggplot(df_pca,
                    aes(x = PC1,
                        y = PC2,
                        colour = group)) +
  stat_ellipse(aes(fill = group),
               type = "norm",
               geom = "polygon",
               alpha = 0.2,
               color = NA,
               level = 0.68, 
               linetype = 2,
               show.legend = FALSE) +
  geom_point() +
  geom_text_repel(aes(label = rownames(df_pca)),
                  size = 3,
                  box.padding = 0.35,
                  point.padding = 0.5,
                  segment.color = "grey50") +
  labs(x = x_lable,
       y = y_lable,
       color = "Group") +
  theme(legend.key = element_rect(colour = "black", 
                                  size = 0.2
                                  )
        ) + 
  guides(fill = "none") + 
  theme_custom + 
  coord_fixed(ratio = 1.4) #调整x轴和y轴长度比例

print(p_ggplot2_lable)

ggsave("PCA_results/ggplot2_method/p_ggplot2_lable.png", 
       plot = p_ggplot2_lable, width = 6, height = 6, dpi = 300)
ggsave("PCA_results/ggplot2_method/p_ggplot2_lable.pdf", 
       plot = p_ggplot2_lable, width = 6, height = 6)

# 关闭图形设备（根据实际需求选择是否保留）
dev.off()