Predict temporal or geographic genomic offset

Source: R/genomic_offset.R

The genomic_offset function allows the user to predict genomic offset from a RDA model. The genomic_offset function can estimate both temporal or spatial genomic offset and accommodates raster data or discrete populations.

genomic_offset(
  RDA,
  K,
  env_pres,
  env_fut,
  env_mask = NULL,
  env_gar,
  method = "loadings"
)

# S4 method for class 'rda,ANY,SpatRaster,ANY,ANY,missing'
genomic_offset(RDA, K, env_pres, env_fut, env_mask = NULL, method = "loadings")

# S4 method for class 'rda,ANY,data.frame,ANY,missing,missing'
genomic_offset(RDA, K, env_pres, env_fut, method = "loadings")

# S4 method for class 'rda,ANY,data.frame,missing,missing,ANY'
genomic_offset(RDA, K, env_pres, env_gar, method = "loadings")

Arguments

RDA

a RDA model from which to extract loci and environmental variable scores

K

an integer specifying the number of RDA axes to use for the projection

env_pres

a RasterStack object or a data.frame with the environmental conditions in the present

env_fut

a RasterStack object or a data.frame with the environmental conditions in the future

env_mask

(optional, default NULL)
a Raster object to limit the projection to a specific area

method

(default 'loadings')
a character defining whether the function is to use weighted averages (scaling type 1, loadings) or linear combinations (scaling type 2, predict) of the projected environmental variables to predict site scores (i.e., adaptive index)

Value

A list containing :

  • genomic_offset : a RasterStack or a data.frame containing the genomic offset predictions for the K first RDA axes, as well as the overall genomic offset prediction

  • weights : the weights associated with each RDA axis used for the predictions

Details

This RDA-based method to predict genomic offset is relatively simple. RDA is first used to predict the optimal adaptive genetic composition for each environmental pixel under consideration (see adaptive_index function), using both current and future environmental conditions. The euclidean distance between these two predictions in the RDA space provides an estimate of the change in genetic composition that would be required to track climate change.

See also

adaptive_index

Author

Thibaut Capblancq