admix.simulate.binary_pheno#

admix.simulate.binary_pheno(dset: Dataset, hsq: float = 0.5, cor: float | None = None, n_causal: int | None = None, case_prevalence: float = 0.1, beta: ndarray | None = None, cov_cols: List[str] | None = None, cov_effects: List[float] | None = None, n_sim=10, method='probit')[source]#

Simulate under liability threshold. First simulate quantative traits in the liability scale. With the given case_prevalence, convert liability to binary traits.

Parameters:

  • hsq (if method is “probit”, this corresponds to the proportion of variance explained)

  • by the genotype in the liability scale.

  • if method is “logit”, this corresponds to the variance explained of (Xeta),

  • and hsq can be larger than 1, in the case.