如何为数据集各值生成独立数据集？求解释R语言gridExtra绘图脚本

Hey there! Let's break down your provided code piece by piece, fix a small bug in it, and then show you straightforward ways to create individual plots for each of your 21 genes.

原代码的详细解释 #

First, let's walk through what each part does:

1. Load required packages

   • library(ggplot2): This is the go-to package for creating polished, customizable plots in R.
   • library(gridExtra): The star here is its grid.arrange() function, which lets you combine multiple ggplot objects into a single layout (great for putting all your gene plots on one page or PDF).

2. Prepare the data

   • data=read.csv("tubgal4crosses.csv"): Reads your CSV dataset into R.
   • str(data): Prints the structure of your data (variable types, sample values) to help you verify it loaded correctly.
   • datasum=aggregate(Count~Gene+Sex+Markers,FUN=sum,data=data): Aggregates your data by grouping on Gene, Sex, and Markers, then sums up the Count values. This ensures you're plotting total counts per group instead of raw, possibly duplicated data.
   • genes=unique(data$Gene): Extracts all distinct gene names from your dataset—this is what we'll loop over to create a plot for each gene.

3. The plot list (p) and loop

   • p<-list(): Initializes an empty list to store each ggplot object. Using a list is perfect here because we need to hold multiple plot objects (one per gene) to combine later.
   • The while loop runs once for each gene:
     • p[[i]]<-ggplot(...): Creates a grouped bar plot for the i-th gene. We filter the aggregated data to only include the current gene, set x=Markers, y=Count, and use fill=Sex to color-code bars by sex. geom_bar(stat="identity", position=position_dodge()) makes grouped bars (side-by-side instead of stacked), and theme_minimal() applies a clean, simple theme.
     • Important bug fix: Your original code has two lines (p[[i]] + scale_fill_manual(...) and p[[i]] + ggtitle(genes[i])) that don't actually modify the plot stored in p[[i]]. You need to reassign these changes back to p[[i]], like this:

       p[[i]] <- p[[i]] + scale_fill_manual(values=c('#999999','#E69F00'))
       p[[i]] <- p[[i]] + ggtitle(genes[i])
       

       Without this reassigning, the custom colors and gene title won't show up in your final plots!

4. Save all plots to a single PDF

   • pdf("scop.pdf"): Opens a PDF file to save your plots.
   • do.call(grid.arrange,p): Uses grid.arrange() to arrange all plots in the list p into a single layout (it automatically decides how many rows/columns to use based on the number of plots). do.call() is needed here because grid.arrange() doesn't accept a list directly—this function passes each plot in the list as a separate argument to grid.arrange().
   • dev.off(): Closes the PDF device, finalizing your saved file.

拆分绘制独立图表的方法 #

If you want each gene's plot saved as a separate file (or as separate pages in a PDF), here are two easy approaches:

Method 1: Save each plot as an individual file (PDF/PNG) #

This uses ggsave() (ggplot's built-in save function) to create a unique file for each gene:

library(ggplot2)

# Load and prepare data
data <- read.csv("tubgal4crosses.csv")
datasum <- aggregate(Count~Gene+Sex+Markers, FUN=sum, data=data)
genes <- unique(data$Gene)

# Loop through each gene to create and save a plot
for (gene in genes) {
  # Create the plot for the current gene
  plot <- ggplot(data=subset(datasum, Gene==gene), aes(x=Markers, y=Count, fill=Sex)) +
    geom_bar(stat="identity", position=position_dodge()) +
    theme_minimal() +
    scale_fill_manual(values=c('#999999','#E69F00')) +
    ggtitle(gene)
  
  # Save as PDF (change extension to .png for image files)
  ggsave(
    filename = paste0(gene, ".pdf"),  # Filename uses the gene name
    plot = plot,
    width = 6,  # Adjust width/height as needed
    height = 4
  )
}

This will generate files like GeneX.pdf, GeneY.pdf, etc., each containing only the plot for that gene.

Method 2: Save all plots as separate pages in one PDF #

If you prefer having all plots in a single PDF but on separate pages (instead of a grid), use this approach:

library(ggplot2)
library(gridExtra)

# Load and prepare data
data <- read.csv("tubgal4crosses.csv")
datasum <- aggregate(Count~Gene+Sex+Markers, FUN=sum, data=data)
genes <- unique(data$Gene)

# Open a PDF file that will hold all plots (one per page)
pdf("scop_individual_pages.pdf", onefile = TRUE)

# Loop through each gene and add its plot to the PDF
for (gene in genes) {
  plot <- ggplot(data=subset(datasum, Gene==gene), aes(x=Markers, y=Count, fill=Sex)) +
    geom_bar(stat="identity", position=position_dodge()) +
    theme_minimal() +
    scale_fill_manual(values=c('#999999','#E69F00')) +
    ggtitle(gene)
  
  grid.arrange(plot)  # Adds the plot as a new page
}

dev.off()  # Close the PDF device

The onefile = TRUE argument ensures all plots are saved to the same PDF, each on its own page.

内容的提问来源于stack exchange，提问作者Kodewings