Fit kit-stratified rCLR centering sets.

Usage

fit_kit_stratified_rclr(
  expr_matrix,
  sample_meta,
  kit_label_col,
  biofluid_col = "biofluid",
  trim_upper = 0.1,
  trim_lower = 0.05,
  min_centering_size = 8L,
  min_samples_per_stratum = 5L,
  exclude_features = c("hsa-miR-451a", "hsa-miR-16-5p", "hsa-miR-486-5p",
    "hsa-miR-144-3p", "hsa-miR-223-3p")
)

Arguments

expr_matrix: Numeric matrix, samples x features, counts or non-negative abundance values.
sample_meta: data.frame with one row per sample.
kit_label_col: Column in `sample_meta` containing library-kit family. Missing/unknown kit values fall back to `biofluid_col`, then global.
biofluid_col: Optional fallback column. Default "biofluid".
trim_upper, trim_lower: Feature trimming fractions used to choose centering features from the training stratum.
min_centering_size: Minimum centering set size before global fallback.
min_samples_per_stratum: Minimum training samples needed for a stratum-specific centering set.
exclude_features: Features excluded from denominator selection.

Value

A fit object consumed by `score_kit_stratified_rclr()`.