splitvalidate: Function to perform split-validation using output from...

In SeojinHwang/scadsum: nonconvex penalty with summary statistics and a reference panel

Function to perform split-validation using output from lassosum.pipeline with external phenotype

Description

Function to perform split-validation using output from lassosum.pipeline with external phenotype

Usage

splitvalidate(...)

## S3 method for class 'lassosum.pipeline'
splitvalidate(
  ls.pipeline,
  test.bfile = NULL,
  keep = NULL,
  remove = NULL,
  pheno = NULL,
  covar = NULL,
  trace = 1,
  split = NULL,
  rematch = !is.null(test.bfile),
  ...
)

Arguments

...	parameters to pass to validate.lassosum.pipeline
ls.pipeline	A lassosum.pipeline object
test.bfile	The (PLINK bfile for the test dataset
keep	Participants to keep (see lassosum for more details)
remove	Participants to remove
pheno	A vector of phenotype OR a data.frame with 3 columns, the first 2 columns being headed "FID" and "IID", OR a filename for such a data.frame
covar	A matrix of covariates OR a data.frame with 3 or more columns, the first 2 columns being headed "FID" and "IID", OR a filename for such a data.frame
trace	Controls amount of output
rematch	Forces a rematching of the ls.pipline beta's with the new .bim file

Details

Performs split-validation. Randomly split the test data into half for validation and half for prediction. Standardize the best cross-predicted pgs and stack together.

SeojinHwang/scadsum documentation built on June 30, 2023, 10:52 p.m.