M-values

In this document, all CpG Beta values were converted into M values using a logit-like transformation, log2(beta / (1 - beta)), to stabilize variance and improve the suitability of the data for statistical modeling.

#Load require packages
library(dplyr)
library(ggplot2)
library(kableExtra)

#Load data set
GSE55763_100_imputed <- read.csv("GSE55763_100_imputed.csv", header = TRUE, stringsAsFactors = FALSE)
head(GSE55763_100_imputed,3)

# Identify CpG columns
cpg_columns <- grep("^cg", colnames(GSE55763_100_imputed), value = TRUE)

# Copy the dataset in object called GSE55763_M
GSE55763_M <- GSE55763_100_imputed

# Extract Beta values
beta_matrix <- GSE55763_M[, cpg_columns]

beta_matrix[beta_matrix == 0] <- 1e-6
beta_matrix[beta_matrix == 1] <- 1 - 1e-6

# Transform Beta to M values (logit transformation)
M_matrix <- log2(beta_matrix / (1 - beta_matrix))

# Replace CpG columns with M values
GSE55763_M[, cpg_columns] <- M_matrix

#Verify first 3 rows
head(GSE55763_M,3)

# Save file in CSV format
write.csv(GSE55763_M, "GSE55763_M_values.csv", row.names = FALSE)