GSE76008 Processing Pipeline

Publication

Nature communications (2022) — PMID 35781533

Dataset

A 17-Gene Stemness Score for Rapid Identification of High-Risk AML Patients [Illumina]

1. 1

   The data were normalized with variance stabilization and quantile normalization using the lumi (v2.16.0) package in R (v3.1.0).

   R v3.1.0 GitHub

   $ Bash example

   # Install R (v3.1.0) if not already installed. Specific old versions of R might require manual installation or using tools like `r-env` or `conda`.
   # For example, using conda:
   # conda create -n r3.1 r-base=3.1.0 -y
   # conda activate r3.1
   
   # Install BiocManager (if not already installed)
   # R -e 'if (!requireNamespace("BiocManager", quietly = TRUE)) install.packages("BiocManager", repos="https://cran.rstudio.com")'
   
   # Install lumi package (v2.16.0) compatible with R 3.1.0 (Bioconductor 2.14)
   # R -e 'BiocManager::install("lumi", version="2.14", ask=FALSE)'
   
   # Create an R script for normalization
   cat << 'EOF' > normalize_lumi_data.R
   # Load the lumi package
   library(lumi)
   
   # --- Input data loading ---
   # This script assumes you have a LumiBatch object named 'lumi_object'
   # available in your R environment or can load it from a file.
   # Replace 'path/to/your/input_lumi_object.RData' with the actual path
   # to your LumiBatch object saved from a previous step, or
   # load raw data using lumiR() or similar.
   
   # Example: Load a pre-saved LumiBatch object
   # load("path/to/your/input_lumi_object.RData")
   # If starting from raw data, use lumiR:
   # lumi_object <- lumiR("path/to/your/raw_data.txt")
   
   # For demonstration, creating a dummy LumiBatch object if not loaded
   if (!exists("lumi_object")) {
       message("Input 'lumi_object' not found. Creating a dummy object for demonstration.")
       # Simulate expression data (e.g., 100 probes, 5 samples)
       exprs_data <- matrix(rnorm(100 * 5, mean = 1000, sd = 200), nrow = 100, ncol = 5)
       rownames(exprs_data) <- paste0("Probe", 1:100)
       colnames(exprs_data) <- paste0("Sample", 1:5)
       # Simulate pheno data
       pheno_data <- data.frame(row.names = colnames(exprs_data),
                                Group = c("A", "A", "B", "B", "A"))
       # Create a LumiBatch object
       lumi_object <- new("LumiBatch", exprs = exprs_data, phenoData = new("AnnotatedDataFrame", data = pheno_data))
       message("Dummy LumiBatch object created.")
   }
   
   # Perform variance stabilization transformation (VST)
   # The 'lumiT' function with method="vst" performs variance stabilization.
   message("Performing variance stabilization transformation...")
   vst_lumi_object <- lumiT(lumi_object, method = "vst")
   
   # Perform quantile normalization
   # The 'lumiN' function with method="quantile" performs quantile normalization.
   message("Performing quantile normalization...")
   normalized_lumi_object <- lumiN(vst_lumi_object, method = "quantile")
   
   # --- Output data saving ---
   # Save the normalized expression matrix to a tab-separated file
   write.table(exprs(normalized_lumi_object),
               file = "normalized_expression_data.txt",
               sep = "\t", quote = FALSE, row.names = TRUE, col.names = TRUE)
   message("Normalized expression data saved to normalized_expression_data.txt")
   
   # Optionally, save the entire normalized LumiBatch object for further R analysis
   save(normalized_lumi_object, file = "normalized_lumi_object.RData")
   message("Normalized LumiBatch object saved to normalized_lumi_object.RData")
   EOF
   
   # Execute the R script using the specified R version
   # Ensure R 3.1.0 is in your PATH or specify its full path if needed.
   Rscript normalize_lumi_data.R

   Copy View on GitHub

Tools Used