Validate GWAS Data for B Estimation

Source: R/get_B_data_loader.R

Validates that GWAS summary statistics meet the requirements for B estimation. Checks for required columns, valid value ranges, and data consistency.

validate_B_training_data(data, method = "auto", verbose = 1)

Arguments

data

data.frame with GWAS summary statistics

method

Character string specifying B estimation method: "auto", "beta", or "pvalue". Determines which columns are required

verbose

Integer verbosity level (0=silent, 1=messages)

Value

List with elements:

  • valid: Logical indicating if data is valid

  • errors: Character vector of error messages (empty if valid)

  • warnings: Character vector of warning messages

  • summary: Summary statistics about the data

Details

Validation Checks:

Required for all methods:

  • MAF column present and values in (0, 1)

  • P column present and values in (0, 1)

  • N column present and values > 0

Additional for beta method:

  • BETA column present and numeric

Data consistency checks:

  • No duplicate rsIDs (warning if present)

  • Reasonable sample sizes (warning if very small or very large)

  • MAF harmonized to <= 0.5

Computational complexity: O(n x c) where n = number of variants, c = number of columns to validate. Current implementation makes multiple passes through the data.

Examples

if (FALSE) { # \dontrun{
# Validate data
validation <- validate_B_training_data(my_data, method = "pvalue")

if (!validation$valid) {
  stop("Validation failed: ", paste(validation$errors, collapse = "; "))
}
} # }