Skip to contents

Format covariates for differential analysis

Source: R/fetch_prep_data.R
fix_covariates.Rd

If fewer than 5% of values are missing from a continuous covariate, replace missing values with the mean in meta. Otherwise, remove the covariate from covar. If values are missing from a factor covariate, remove the covariate from covar. If a covariate has constant values, remove it from covar. Note that covariates are only removed from covar, not meta.

Usage

fix_covariates(covar, meta, center_scale = FALSE)

Arguments

covar

string or character vector of covariate names that correspond to column names of meta

meta

sample by variable data frame of metadata

center_scale

boolean, whether to center and scale continuous variables

Value

named list of two items:

meta

data frame input meta with covariates imputed and/or centered and scaled as necessary

covariates

character vector input covar after removing covariates as necessary

See also

transcript_prep_data()

Examples

meta = data.frame(V1 = c(rnorm(19), NA),
                  V2 = c(rnorm(12), rep(NA, 8)))
covar = c("V1","V2")
result = fix_covariates(covar, meta)
#> Warning: Numeric variable of interest V1 has 1 missing values. Replacing missing values with mean.
#> Warning: Numeric variable of interest V2 has 8 missing values. Removing.
result = fix_covariates(covar, meta, center_scale = TRUE)
#> Warning: Numeric variable of interest V1 has 1 missing values. Replacing missing values with mean.
#> Warning: Numeric variable of interest V2 has 8 missing values. Removing.