Skip to contents

Define a discrete fixed-class effect in a phenotype model

Source: R/define_effect_fixed_class.R
define_effect_fixed_class.Rd

Inserts a row into phenotype_effects representing a discrete fixed effect: each level of source_column maps to a numeric shift on the phenotype scale. The null_class_action argument controls what happens when an individual has a NULL value for source_column at phenotyping time.

Usage

define_effect_fixed_class(
  pop,
  phenotype_name,
  effect_name,
  source_column,
  levels,
  source_table = "ind_meta",
  null_class_action = c("skip", "error", "zero"),
  overwrite = FALSE
)

Arguments

pop

A tidybreed_pop object.

phenotype_name

Character. Name of an existing phenotype in phenotype_meta.

effect_name

Character. Unique label for this effect within the phenotype. Must be a valid SQL identifier.

source_column

Character. Column in source_table holding the group membership. Must be a valid SQL identifier.

levels

Named numeric vector mapping level strings to shifts, e.g. c(M = 30, F = 0). Every level present in the data at phenotyping time must appear here.

source_table

Character. Database table containing source_column. Default "ind_meta".

null_class_action

Character. Behaviour when source_column is NULL for an individual:

  • "skip" (default) — individual is excluded from ind_phenotype for this phenotype; a warning lists the count.

  • "error" — hard stop when any NULL is encountered. Use for effects like sex that should never be missing.

  • "zero" — treat NULL as a zero-shift contribution (reference level).

overwrite

Logical. Replace an existing effect with the same name.

Value

The modified tidybreed_pop (invisibly).

See also

define_effect_intercept(), define_effect_fixed_cov(), define_effect_random(), add_phenotype()

Examples

if (FALSE) { # \dontrun{
pop <- pop |>
  define_effect_fixed_class("ADG", "sex",
    source_column = "sex",
    levels = c(M = 30, F = 0))
} # }