predict_multisnpnet: Predict from the Fitted Object or File

In junyangq/multiSnpnet: Fit a sparse reduced rank regression model on large-scale SNP data and multivariate responses

Predict from the Fitted Object or File

Description

Predict from the Fitted Object or File

Usage

predict_multisnpnet(
  fit = NULL,
  saved_path = NULL,
  new_genotype_file,
  new_phenotype_file,
  idx = NULL,
  covariate_names = NULL,
  split_col = NULL,
  split_name = NULL,
  binary_phenotypes = NULL,
  zstdcat_path = "zstdcat"
)

Arguments

fit	List of fit object returned from multisnpnet.
saved_path	Path to the file that saves the fit object. The full path is constructed as $saved_path$idx.RData.
new_genotype_file	Path to the new suite of genotype files. genotype_file.pgen, psam, pvar.zst. must exist.
new_phenotype_file	Path to the phenotype. The header must include FID, IID.
idx	Lambda indices where the coefficients are requested.
covariate_names	Character vector of the names of the adjustment covariates.
split_col	Name of the split column. If NULL, all samples will be used.
split_name	Vector of split labels where prediction is to be made.
binary_phenotypes	Vector of names of the binary phenotypes. If training split is provided, logistic regression will be refitted on the covariates and linear prediction score (from multivariate fit) and the final prediction updated. In addition, AUC will be computed for binary phenotypes.
zstdcat_path	Path to zstdcat program, needed when loading variants

Value

A list containing the prediction and the resopnse for which the prediction is made.

junyangq/multiSnpnet documentation built on Oct. 19, 2023, 8:22 p.m.