SVG (version 1.0.0)

binarize_expression: Binarize Gene Expression

Description

Converts continuous gene expression values to binary (0/1) using various methods. Used by the binSpect method.

Usage

binarize_expression(
  expr_matrix,
  method = c("kmeans", "rank", "median", "mean"),
  rank_percent = 30,
  n_threads = 1L,
  verbose = FALSE
)

Value

Binary matrix with same dimensions as input.

Arguments

expr_matrix

Numeric matrix of gene expression. Rows are genes, columns are spots/cells.

method

Character string specifying binarization method.

• "kmeans" (default): Use k-means clustering (k=2) to separate high and low expression

• "rank": Binarize based on expression rank percentile

• "median": Values above median are set to 1

• "mean": Values above mean are set to 1

rank_percent

Numeric. For method = "rank", the percentile threshold (0-100). Values in the top rank_percent percent are set to 1. Default is 30.

n_threads

Integer. Number of threads for parallel computation. Default is 1.

verbose

Logical. Whether to print progress. Default is FALSE.

Details

K-means method: For each gene, k-means clustering with k=2 is applied. The cluster with higher mean expression is labeled as 1, the other as 0.

Rank method: For each gene, spots are ranked by expression. The top rank_percent percent are labeled as 1.

Examples

# Create example expression matrix
expr <- matrix(rpois(1000, lambda = 10), nrow = 10, ncol = 100)
rownames(expr) <- paste0("gene_", 1:10)

# Binarize using k-means
bin_kmeans <- binarize_expression(expr, method = "kmeans")

# Binarize using rank (top 20%)
bin_rank <- binarize_expression(expr, method = "rank", rank_percent = 20)