lassosum2 — snp_lassosum2 • bigsnpr

lassosum2

snp_lassosum2(
  corr,
  df_beta,
  delta = signif(seq_log(0.001, 3, 6), 1),
  nlambda = 20,
  lambda.min.ratio = 0.01,
  dfmax = 2e+05,
  maxiter = 500,
  tol = 1e-05,
  ncores = 1
)

Arguments

corr	Sparse correlation matrix as an SFBM. If `corr` is a dsCMatrix or a dgCMatrix, you can use `as_SFBM(corr)`.
df_beta	A data frame with 3 columns: `$beta`: effect size estimates `$beta_se`: standard errors of effect size estimates `$n_eff`: sample size when estimating `beta` (in the case of binary traits, this is `4 / (1 / n_control + 1 / n_case)`)
delta	Vector of shrinkage parameters to try (L2-regularization). Default is `c(0.001, 0.005, 0.02, 0.1, 0.6, 3)`.
nlambda	Number of different lambdas to try (L1-regularization). Default is `20`.
lambda.min.ratio	Ratio between last and first lambdas to try. Default is `0.01`.
dfmax	Maximum number of non-zero effects in the model. Default is `200e3`.
maxiter	Maximum number of iterations before convergence. Default is `500`.
tol	Tolerance parameter for assessing convergence. Default is `1e-5`.
ncores	Number of cores used. Default doesn't use parallelism. You may use nb_cores.

Value

A matrix of effect sizes, one vector (column) for each row in attr(<res>, "grid_param"). Missing values are returned when strong divergence is detected.