Score StableMate stable-predictor discriminator (single-sample)

Scores each row of X independently with the frozen stable-feature logistic learned by fit_sel_stablemate. For one specimen its per-sample rCLR is computed over the frozen candidate universe (present features only), restricted to the frozen selected stable features, and the frozen logistic linear predictor $$\eta = intercept + \sum_{j \in S} \beta_j \, z_j$$ is returned (NOT plogis() of it); larger is more case-like. Here \(z_j\) is the specimen's own rCLR coordinate for selected feature \(j\). Because rCLR centres each specimen by its own geometric mean over present parts, the score is EXACTLY invariant to per-sample positive scaling and uses only that specimen's own values, so the method is single-sample deployable and row-equivariant. Scoring uses no random numbers and no scored-batch statistics.

Missing selected features contribute 0 (their rCLR coordinate is absent from the specimen): the linear predictor is summed only over selected features present in X. If fewer than model$min_selected selected features are present in a specimen, or the model selected nothing at fit time (model$degenerate), that specimen receives the documented neutral score 0.

Usage

score_sel_stablemate(model, X, meta = NULL)

Arguments

model

A sel_stablemate_model object returned by fit_sel_stablemate.

X

Numeric matrix (samples \(\times\) features) of non-negative abundance values. Columns must be named feature ids.

meta

Optional per-sample metadata. Accepted for interface uniformity and ignored by this method (the transfer/environment information is used only at fit time).

Value

Plain finite numeric vector of length nrow(X). Scores use only each row's own values plus the frozen model, with no scored-batch centering, quantiles, or renormalization.

References

Deng Y, Liao Y, Wong NKP, Bhuva DD, Lin Y, Cao Y. (2024) StableMate: a statistical method to select stable predictors in omics data. bioRxiv.

Examples

if (FALSE) { # \dontrun{
set.seed(1)
n <- 180L; p <- 12L
env <- rep(c("e1", "e2", "e3"), length.out = n)
X <- matrix(stats::rgamma(n * p, shape = 5), n, p,
            dimnames = list(NULL, paste0("f", seq_len(p))))
lin <- log(X[, 1]) + log(X[, 2]) + log(X[, 3])
y <- stats::rbinom(n, 1, plogis(scale(lin)))
meta <- data.frame(accession = env, stringsAsFactors = FALSE)
model <- fit_sel_stablemate(X, y, meta_train = meta)
score_sel_stablemate(model, X[1, , drop = FALSE])
} # }