Compute SCAD-Penalized Weights from Summary Statistics

Fits SCAD-penalized regression on the RSS objective, searching over a shrinkage grid s and lambda path. Model selection uses LD-quadratic pseudovalidation by default.

Usage

scadRssWeights(
  stat,
  LD,
  s = c(0.2, 0.5, 0.9, 1),
  gamma = 3.7,
  alpha = 1,
  selection = c("ld_quadratic", "min_fbeta"),
  ...
)

Arguments

stat

A list with $b (effect sizes) and $n (per-variant sample sizes).

LD

LD correlation matrix R (single matrix, NOT pre-shrunk).

s

Numeric vector of LD shrinkage parameters. Default: c(0.2, 0.5, 0.9, 1.0).

gamma

SCAD concavity parameter. Default 3.7.

alpha

Elastic-net mixing (1 = pure L1). Default 1.

selection

Selection strategy: "ld_quadratic" (default) or "min_fbeta".

...

Additional arguments passed to penalizedRss().

Value

A numeric vector of SNP coefficient weights.

Examples

data(eqtlRegionExample)
X <- eqtlRegionExample$X[, 1:30]
y <- eqtlRegionExample$yRes
ss <- lapply(
  seq_len(ncol(X)), function(j) coef(summary(lm(y ~ X[, j])))[2, 1:2])
stat <- list(
  b = vapply(ss, `[`, numeric(1), 1L),
  seb = vapply(ss, `[`, numeric(1), 2L),
  n = rep(nrow(X), ncol(X))
)
LD <- cor(X)
scadRssWeights(stat, LD)
#>  [1] -0.048565027 -0.048563995 -0.023504257  0.005937847  0.006035339
#>  [6] -0.022037522 -0.016458707  0.030890109  0.009653341  0.009390947
#> [11] -0.017635699 -0.037745517 -0.029085130  0.000000000 -0.017680721
#> [16]  0.030919611  0.005418932  0.007347711  0.092705318  0.007547093
#> [21]  0.007911081 -0.021719584  0.008182288 -0.016166490 -0.028844757
#> [26] -0.021718775 -0.111457739 -0.028777774  0.093059199  0.023114605
#> attr(,"penalized_rss_selection")
#>                  mode                 index               penalty 
#>        "ld_quadratic"                   "8"                "SCAD" 
#>                     s                lambda 
#>                 "0.2" "0.00127427498570313"