Skip to contents

Define a continuous covariate (regression) effect in a phenotype model

Source: R/define_effect_fixed_cov.R
define_effect_fixed_cov.Rd

Inserts a row into phenotype_effects representing a polynomial regression term. At phenotyping time the contribution per individual is slope * (source_column_value - center)^poly_order. When center is NULL the mean of source_column is computed and stored automatically.

Usage

define_effect_fixed_cov(
  pop,
  phenotype_name,
  effect_name,
  source_column,
  slope,
  center = NULL,
  poly_order = 1L,
  source_table = "ind_meta",
  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.

source_column

Character. Column in source_table holding the continuous predictor values.

slope

Numeric scalar. Regression coefficient.

center

Numeric scalar or NULL. Value subtracted before raising to poly_order. If NULL, auto-computed as the column mean.

poly_order

Integer ≥ 1. Polynomial order. Default 1 gives the standard linear slope * (x - center) term. Use 2 for quadratic, etc.

source_table

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

overwrite

Logical. Replace an existing effect with the same name.

Value

The modified tidybreed_pop (invisibly).

See also

define_effect_intercept(), define_effect_fixed_class(), define_effect_random(), add_phenotype()

Examples

if (FALSE) { # \dontrun{
# Linear regression on age (center auto-computed)
pop <- pop |>
  define_effect_fixed_cov("ADG", "age",
    source_column = "age_days",
    slope = 2.5)

# Quadratic regression on DIM (dairy)
pop <- pop |>
  define_effect_fixed_cov("milk_td", "dim_sq",
    source_column = "dim",
    slope  = -0.02,
    center = 150,
    poly_order = 2L)
} # }