SparCC correlation matrix (Rcpp implementation)

Fast C++ implementation of SparCC for inferring correlations from compositional count data. Input: samples x taxa matrix. Output: taxa x taxa correlation matrix.

Usage

sparcc_matrix_rcpp(data, iter = 20, inner_iter = 10, th = 0.1, nthreads = 0L)

Arguments

data

Numeric matrix. Rows = samples, columns = taxa (ASVs/OTUs).

iter

Integer. Number of outer iterations (Dirichlet resampling). Default 20.

inner_iter

Integer. Max inner iterations for pair exclusion. Default 10.

th

Numeric. Threshold for excluding highly correlated pairs. Default 0.1.

nthreads

Integer. Number of OpenMP threads for parallel sparccinner (0 = use default). Only effective when OpenMP is available.

Value

Numeric matrix of taxa x taxa correlations (median across iterations).

Examples

if (FALSE) { # \dontrun{
# data: samples x taxa count matrix
cor_mat <- sparcc_matrix_rcpp(asv_mat, iter = 20, inner_iter = 10, th = 0.1)
# With 4 threads (when OpenMP available)
cor_mat <- sparcc_matrix_rcpp(asv_mat, iter = 20, nthreads = 4)
} # }