CpG-predicted-expression negative-control model (SCENIC case)

Simulates spatially correlated CpG methylation whose signal is partly driven by the captured confounders, then forms each negative control as a linear prediction from the methylation sites ("CpG-predicted expression"). The controls therefore carry confounder information only to the extent the methylation does, mediated through a realistic spatial methylation layer.

Usage

nc_cpg(U, covariates, params)

Arguments

U

n x k confounder matrix.

covariates

Covariate data frame (unused; kept for the NC contract).

params

List with n_features, and optionally coverage (confounder->methylation strength, default 0.7), captured (confounder indices, default all), n_cpg (methylation sites, default 60), rho (AR(1) spatial correlation across sites, default 0.6), MMCpG (methylation-confounding multiplier, default 1), MMCon (default 1), and noise_cor (a p x p correlation matrix for correlated idiosyncratic noise, or NULL for independent noise).

Value

n x n_features matrix of CpG-predicted negative controls.

Details

When noise_cor is supplied (a p x p correlation matrix), the idiosyncratic noise added to each control is drawn from a multivariate normal with that correlation structure, so the controls retain realistic cross-feature correlations conditional on the confounder.

Examples

U <- matrix(rnorm(100), 100, 1)
W <- nc_cpg(U, covariates = NULL,
  params = list(n_features = 10, coverage = 0.7, n_cpg = 20))
dim(W)
#> [1] 100  10