SparCC p-value matrix (Rcpp implementation) — sparcc_pvalue

Fast C++-accelerated SparCC p-values via bootstrap and permutation. Uses sparcc_matrix_rcpp internally for speed. Input: samples x taxa matrix. Output: taxa x taxa p-value matrix.

Usage

sparcc_pvalue_rcpp(
  data,
  iter = 20,
  inner_iter = 10,
  th = 0.1,
  R = 20,
  ncpus = 1,
  seed = NULL
)

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.
R: Integer. Number of bootstrap/permutation replicates. Default 20.
ncpus: Integer. Number of CPUs for parallel boot. Default 1.
seed: Integer or NULL (default). If supplied, makes the bootstrap / permutation reproducible. When ncpus > 1 the L'Ecuyer-CMRG generator is used so parallel workers get independent, deterministic RNG streams; the previous RNG kind is restored on exit.

Value

Numeric matrix of taxa x taxa p-values. Diagonal = 0. NaN indicates observed correlation outside bootstrap CI.

Examples

if (FALSE) { # \dontrun{
p_mat <- sparcc_pvalue_rcpp(asv_mat, R = 20, ncpus = 4, seed = 1115)
} # }