安装

Functions and data relevant to the tutorials published on the WeChat official account MedBioInfoCloud are encapsulated in this R package.

Install package¶

This guide provides step-by-step instructions to install all required and optional dependencies for the R package, including distinctions between Bioconductor packages (common in bioinformatics) and CRAN packages (general R packages). Follow the steps below to resolve dependency-related errors during installation.

1. Prerequisites¶

Before installing dependencies:

Ensure your R version is 4.2.0 or higher (compatible with the latest versions of Bioconductor packages).
Install BiocManager (the official tool for Bioconductor package installation) if you haven’t already:

if (!requireNamespace("BiocManager", quietly = TRUE)) {
  install.packages("BiocManager")  # Install from CRAN
}
library(BiocManager)  # Load the package

2. Install Dependencies by Type¶

Dependencies are split into two categories for clarity. Run the code blocks sequentially to avoid missing packages.

2.1 Bioconductor Packages (Bioinformatics-Specific)¶

These packages are not available on CRAN and require installation via BiocManager::install(). Copy and run the entire code block:

# List of Bioconductor dependencies
bioc_packages <- c(
  "DESeq2",          # Differential expression analysis (RNA-seq)
  "edgeR",           # Differential expression analysis (RNA-seq)
  "GenomicFeatures", # Retrieve genomic annotations (e.g., gene coordinates)
  "GseaVis",         # GSEA result visualization
  "limma",           # Differential expression analysis (microarray/RNA-seq)
  "BiocGenerics",    # Core utility functions for Bioconductor packages
  "GSVA",            # Gene set variation analysis
  "miRBaseVersions.db", # miRBase annotation database
  "TCGAbiolinks",    # Access TCGA (The Cancer Genome Atlas) data
  "aPEAR",           # Alternative polyadenylation analysis
  "clusterProfiler", # Functional enrichment analysis (GO/KEGG)
  "org.Hs.eg.db",    # Human gene annotation database (Entrez ID ↔ Symbol)
  "msigdbr"          # Access MSigDB gene sets
)

# Install Bioconductor packages (skip already installed ones)
BiocManager::install(
  packages = bioc_packages,
  update = FALSE,    # Set to TRUE if you want to update existing packages
  ask = FALSE        # Auto-confirm installation (no manual prompts)
)

2.2 CRAN Packages (General R Packages)¶

# List of CRAN dependencies
cran_packages <- c(
  "dplyr",           # Data manipulation (filter, group_by, etc.)
  "ggplot2",         # Core visualization package
  "ggpubr",          # Publication-ready ggplot2 plots
  "ggrepel",         # Avoid text overlap in plots
  "ggsignif",        # Add significance labels (e.g., * p<0.05) to plots
  "purrr",           # Functional programming (e.g., map() for iteration)
  "magrittr",        # Pipe operator (%>%) for readable code
  "RColorBrewer",    # Color palettes for visualization
  "reshape2",        # Data reshaping (melt(), dcast())
  "data.table"       # Fast data frame manipulation
)

# Install CRAN packages (skip already installed ones)
install.packages(
  pkgs = cran_packages,
  repos = "https://cloud.r-project.org/",  # Use CRAN mirror for fast installation
  dependencies = TRUE,                     # Install required sub-dependencies
  quiet = FALSE                            # Show installation progress
)

3. Verify Successful Installation¶

After installation, run the following code to confirm all dependencies are installed and loadable. If no errors appear, the dependencies are ready:

# Function to check if a package is installed and loadable
check_package <- function(pkg) {
  if (requireNamespace(pkg, quietly = TRUE)) {
    message(sprintf("✓ %s is installed and loadable", pkg))
  } else {
    warning(sprintf("✗ %s is missing or failed to load", pkg), call. = FALSE)
  }
}

# Check all dependencies
all_packages <- c(bioc_packages, cran_packages)
invisible(lapply(all_packages, check_package))

4. Troubleshooting Common Errors¶

If you encounter installation failures, try these fixes:

4.1 "Package Not Available" Error¶

For Bioconductor packages: Ensure BiocManager is up-to-date with BiocManager::install(update = TRUE).
For CRAN packages: Use a different CRAN mirror (e.g., repos = "https://cran.r-project.org/").

4.2 Compilation Errors (e.g., "gcc not found")¶

Windows: Install Rtools (required for compiling C/C++ code in packages like data.table).
macOS: Install Xcode Command Line Tools with xcode-select --install (in Terminal).
Linux: Install system libraries via sudo apt-get install r-base-dev (Debian/Ubuntu) or sudo dnf install R-devel (Fedora).

4.3 Out-of-Memory Errors¶

Close unused R sessions or applications to free memory.
Install packages one at a time (e.g., BiocManager::install("DESeq2") followed by BiocManager::install("edgeR")) to reduce memory usage.

5. Proceed to Install the Target R Package¶

Once all dependencies are successfully installed, you can install the target R package (e.g., from a local file or GitHub):

Example 1: Install from a Local `.tar.gz` File¶

install.packages(
  pkgs = "path/to/your/package.tar.gz",  # Replace with your package path
  repos = NULL,
  type = "source",
  dependencies = TRUE  # Ensure dependencies are linked
)

Example 2: Install from GitHub (if applicable)¶

# Install devtools first (if needed)
if (!requireNamespace("devtools", quietly = TRUE)) {
  install.packages("devtools")
}
devtools::install_github("username/package-name")  # Replace with GitHub repo

6. Install MedBioInfoCloud¶

devtools::install_github("BioInfoCloud/MedBioInfoCloud")

Some functions with the same parameters are uniformly explained as follows:

save - A boolean value that determines whether the data should be preserved. If save is set to TRUE, the data will be saved. Conversely, if it's set to FALSE, the data will unsaved.

folder - A string that designates the destination directory for the saved data. This parameter is applicable only when save is set to TRUE. If the specified directory does not exist, it will be automatically created. Kindly note that the folder path should not terminate with a slash ("/"). If the function does not have a save parameter but includes a folder parameter, it indicates that the function will definitely save data, and may not necessarily return valid data. The folder parameter defaults to the current working directory, and it is recommended to set a specific data output directory.

For more learning materials, please refer to MedBioInfoCloud - 生物信息云.

=========================================================================

安装