| Title: | Summarizing Regional Methylation with Regional Principal Components Analysis |
|---|---|
| Description: | Functions to summarize DNA methylation data using regional principal components. Regional principal components are computed using principal components analysis within genomic regions to summarize the variability in methylation levels across CpGs. The number of principal components is chosen using either the Marcenko-Pasteur or Gavish-Donoho method to identify relevant signal in the data. |
| Authors: | Tiffany Eulalio [aut, cre] |
| Maintainer: | Tiffany Eulalio <[email protected]> |
| License: | MIT + file LICENSE |
| Version: | 1.5.0 |
| Built: | 2024-10-31 04:32:08 UTC |
| Source: | https://github.com/bioc/regionalpcs |
Combine results dataframes across regions
combine_results(res, df_name)combine_results(res, df_name)
res |
List of lists; contains summarized region results |
df_name |
String; name of result being combined (sig_pcs or percent_var) |
Data Frame containing results
# Create example data for 'sig_pcs' and 'percent_var' sig_pcs_example <- data.frame(pcs = c("PC1", "PC2"), value = c(0.2, 0.4)) percent_var_example <- data.frame(pcs = c("PC1", "PC2"), value = c(0.7, 0.3)) # Create 'res' list containing both 'sig_pcs' and 'percent_var' res <- list(region = "Region1", sig_pcs = sig_pcs_example, percent_var = percent_var_example) # Example function use: Combine 'sig_pcs' across regions combined_sig_pcs <- combine_results(res, df_name = "sig_pcs") print(combined_sig_pcs)# Create example data for 'sig_pcs' and 'percent_var' sig_pcs_example <- data.frame(pcs = c("PC1", "PC2"), value = c(0.2, 0.4)) percent_var_example <- data.frame(pcs = c("PC1", "PC2"), value = c(0.7, 0.3)) # Create 'res' list containing both 'sig_pcs' and 'percent_var' res <- list(region = "Region1", sig_pcs = sig_pcs_example, percent_var = percent_var_example) # Example function use: Combine 'sig_pcs' across regions combined_sig_pcs <- combine_results(res, df_name = "sig_pcs") print(combined_sig_pcs)
Compute significant dimensions of a matrix using the Marchenko-Pastur or Gavish-Donoho methods
compute_dimension( x, var_explained, noise_select, pc_method = c("gd", "mp"), verbose = FALSE )compute_dimension( x, var_explained, noise_select, pc_method = c("gd", "mp"), verbose = FALSE )
x |
A data frame or matrix of methylation values; rows = features, columns = samples |
var_explained |
A numeric vector containing the variance explained by successive PCs, sorted in decreasing order. (Used for PCAtools) |
noise_select |
Numeric scalar specifying the variance of the random noise (Used for PCAtools) |
pc_method |
String indicating the method for estimating dimension; "gd" = Gavish-Donoho, "mp" = Marchenko-Pastur |
verbose |
Boolean indicating whether to print statements while running, default = FALSE |
Numeric scalar representing the optimal number of PCs to retain using the specified method
x <- diag(4) pca_res <- PCAtools::pca(x) # Run PCA eig_sq <- pca_res$sdev^2 # Compute variance explained compute_dimension(x, eig_sq, 1, "gd")x <- diag(4) pca_res <- PCAtools::pca(x) # Run PCA eig_sq <- pca_res$sdev^2 # Compute variance explained compute_dimension(x, eig_sq, 1, "gd")
Compute regional principal components for methylation data
compute_regional_pcs( meth, region_map, pc_method = c("gd", "mp"), verbose = FALSE )compute_regional_pcs( meth, region_map, pc_method = c("gd", "mp"), verbose = FALSE )
meth |
Data frame of methylation beta values, with CpGs in rows and samples in columns |
region_map |
Data frame mapping CpGs to gene regions |
pc_method |
Method to use for PC computation, either 'gd' (Gavish-Donoho) or 'mp' (Marchenko-Pastur) |
verbose |
Logical, should progress messages be displayed? |
A list containing several elements, including the regional PCs, percent variance, and other information
# Create synthetic methylation data meth_data <- matrix(rnorm(1000), nrow = 100, ncol = 10) rownames(meth_data) <- paste0("CpG", 1:100) colnames(meth_data) <- paste0("Sample", 1:10) # Create a synthetic region map region_map_data <- data.frame( region_id = rep(c("Gene1", "Gene2"), each = 50), cpg_id = rownames(meth_data) ) # Run the function compute_regional_pcs(meth_data, region_map_data, pc_method = 'gd')# Create synthetic methylation data meth_data <- matrix(rnorm(1000), nrow = 100, ncol = 10) rownames(meth_data) <- paste0("CpG", 1:100) colnames(meth_data) <- paste0("Sample", 1:10) # Create a synthetic region map region_map_data <- data.frame( region_id = rep(c("Gene1", "Gene2"), each = 50), cpg_id = rownames(meth_data) ) # Run the function compute_regional_pcs(meth_data, region_map_data, pc_method = 'gd')
This function generates a map that assigns CpG sites to gene regions, establishing a linkage based on their genomic coordinates and providing a foundation for subsequent region-specific analyses.
create_region_map(cpg_gr, genes_gr, verbose = FALSE)create_region_map(cpg_gr, genes_gr, verbose = FALSE)
cpg_gr |
A |
genes_gr |
A |
verbose |
Boolean; print output statements |
A data.frame with mappings between gene IDs and CpG IDs,
facilitating associating CpG sites with their corresponding gene regions for
downstream analyses.
library(GenomicRanges) # Creating dummy GRanges objects for CpG sites and gene regions cpg_gr <- GRanges(seqnames=c("chr1", "chr1", "chr2"), ranges=IRanges(start=c(100, 200, 150), end=c(100, 200, 150))) genes_gr <- GRanges(seqnames=c("chr1", "chr2", "chr2"), ranges=IRanges(start=c(50, 100, 130), end=c(150, 180, 160))) # Creating a region map using the function region_map <- create_region_map(cpg_gr, genes_gr)library(GenomicRanges) # Creating dummy GRanges objects for CpG sites and gene regions cpg_gr <- GRanges(seqnames=c("chr1", "chr1", "chr2"), ranges=IRanges(start=c(100, 200, 150), end=c(100, 200, 150))) genes_gr <- GRanges(seqnames=c("chr1", "chr2", "chr2"), ranges=IRanges(start=c(50, 100, 130), end=c(150, 180, 160))) # Creating a region map using the function region_map <- create_region_map(cpg_gr, genes_gr)
Get significant principal components
get_sig_pcs(x, pc_method = c("mp", "gd"), verbose = FALSE)get_sig_pcs(x, pc_method = c("mp", "gd"), verbose = FALSE)
x |
A data frame or matrix of methylation values; rows = features, columns = samples |
pc_method |
String indicating the method for estimating dimension; "gd" = Gavish-Donoho (default), "mp" = Marchenko-Pastur |
verbose |
Boolean; print output statements |
List containing four elements; sig_pcs = significant PCs, percent_var = percent variance explained, loadings = PC loadings, est_dim = estimated dimension
x <- diag(4) get_sig_pcs(x, "gd")x <- diag(4) get_sig_pcs(x, "gd")
Summarize a region using regional principal components
summarize_region(region, region_map, meth, pc_method, verbose = FALSE)summarize_region(region, region_map, meth, pc_method, verbose = FALSE)
region |
String; name of region being processed |
region_map |
Data frame; Mapping of CpGs to regions, column 1 should be regions, column 2 should be CpGs with the same names as the rows of meth |
meth |
Data frame or matrix; Methylation values to summarize; rows=CpGs, columns=samples |
pc_method |
String; indicating the method for estimating dimension; "gd"=Gavish-Donoho (default), "mp"=Marchenko-Pastur |
verbose |
Boolean; print output statements |
list containing PC results
# Create the region map with just one region containing 10 CpGs region_map <- data.frame(region_id = rep(1, 10), cpg_id = seq(1, 10)) # Create methylation data frame set.seed(123) meth <- as.data.frame(matrix(runif(10 * 20, min = 0, max = 1), nrow = 10)) rownames(meth) <- seq(1, 10) # Call the function summarize_region(1, region_map, meth, 'gd')# Create the region map with just one region containing 10 CpGs region_map <- data.frame(region_id = rep(1, 10), cpg_id = seq(1, 10)) # Create methylation data frame set.seed(123) meth <- as.data.frame(matrix(runif(10 * 20, min = 0, max = 1), nrow = 10)) rownames(meth) <- seq(1, 10) # Call the function summarize_region(1, region_map, meth, 'gd')