generate\(algorithm\)PRScs

PRS-CS algorithm specification

Description

Configure a single PRS-CS algorithm specification.

Usage

`generate$algorithm$PRScs`(reference.panel, variant.space = NULL, pval = 1, phi = NULL, a = 1, b = 0.5, n.iter = 1000, n.burnin = 500, thin = 5, beta.std = FALSE, write.psi = FALSE, write.pst = FALSE, ld.block.size = 1000, python = NULL, seed = NULL, ncores = 1)

Arguments

reference.panel

Character scalar reference-panel name used for LD.

variant.space

Optional variant-space name used to derive the effective reference panel before LD construction.

pval

Numeric scalar in ⁠[0, 1]⁠; GWAS p-value filter before fitting.

phi

Optional numeric global shrinkage parameter. NULL enables PRS-CS auto-phi behavior.

a, b

Numeric PRS-CS prior parameters.

n.iter, n.burnin, thin

Integer PRS-CS MCMC controls.

beta.std

Logical; forwarded to PRS-CS.

write.psi

Logical; forwarded to PRS-CS.

write.pst

Logical; posterior-sample output is not converted into a PolyGenius model and is therefore currently unsupported for model generation.

ld.block.size

Integer number of reference variants per PRS-CS LD block when PolyGenius builds neutral block files.

python

Optional Python executable. When NULL, PolyGenius uses python3 or python from PATH.

seed

Optional integer random seed.

ncores

Integer worker cores to request for LD construction and the PRS-CS rule.

Details

PRS-CS consumes ld.prscs storage built from the requested reference panel. When variant.space is supplied, PolyGenius first resolves a restricted reference panel and then builds PRS-CS LD blocks from that panel. The ld.prscs rule itself only receives the effective reference panel; it does not interpret the user-facing variant.space option.

PRS-CS fitting is delegated to the package Python wrapper, which expects the upstream PRS-CS Python modules to be importable in the selected Python environment.

Value

ResourceSpecSet of generate algorithm resources.