R/bs_covar_est.R

In fboehm/limmbo2: Covariance Matrices Estimation for Multivariate Linear Mixed Effects Models in Genetics Studies

#' Perform covariance estimates for each of multiple bootstrap samples of traits
#'
#' @param limmbo_out object produced by call to limmbo$core$vdbootstrap$LiMMBo
#' @param cpus number of cpus to use (as an R object)
#' @param seed integer for pseudo random number seed (as an R object)
#'
#' @export
#' @examples \dontrun{
#' prep_data(pheno = matrix(rnorm(300), nrow = 100), kinship = diag(100)) -> dat
#' make_limmbo(dat, timing = TRUE, iterations = 10, subset_size = 2) -> limmbo_out
#' bs_covar_est(limmbo_out, cpus = 1, seed = 100)
#' }
bs_covar_est <- function(limmbo_out, cpus, seed){
  # change inputs to np objects
  cpus_np <- np$int(cpus)
  seed_np <- reticulate::np_array(seed, dtype = "int64")
  # call python function
  out <- limmbo$core$vdbootstrap$LiMMBo$runBootstrapCovarianceEstimation(limmbo_out,
                                                                  cpus = cpus_np,
                                                                  seed = seed_np
                                                                    )
  return(out)
}