Skip to contents

Rank resistance surfaces by how well they explain genetic distance

Source: R/rank.R
rank_resistance.Rd

Compares competing resistance surfaces, each a hypothesis about what slows movement, by how well the least cost distance through it explains a pairwise genetic distance such as FST. For every surface, the cost of the cheapest route between each pair of sites is related to the genetic distance with a mixed model that has source and destination site as random effects (Maier et al. 2022). Models are fit by maximum likelihood, so they can be compared by log-likelihood and AIC. Straight-line distance is added as a baseline: a surface that cannot beat it adds nothing.

Usage

rank_resistance(
  surfaces,
  sites,
  genetic,
  id = "site",
  pairs = NULL,
  max_dist = Inf,
  barrier = 1e+06,
  straight = TRUE
)

Arguments

surfaces

Resistance surfaces to compare: a named list of single-layer SpatRasters, or one multi-layer SpatRaster with a layer per hypothesis. Names label the hypotheses.

sites

An sf object of points or polygons with an ID column.

genetic

A square matrix of pairwise genetic distance (for example from pairwise_fst()) with site IDs as row and column names.

id

Name of the ID column in sites.

pairs

Optional data.frame with columns from and to. By default every unordered pair of sites closer than max_dist.

max_dist

Only use pairs whose straight-line distance is below this (meters). Ignored when pairs is given.

barrier

Resistance at or above which a cell cannot be crossed, see make_transition().

straight

Add straight-line distance as a baseline hypothesis.

Value

A data.frame with one row per hypothesis, sorted from best to worst: hypothesis, logLik, AIC, delta_AIC (difference from the best model), R2m (marginal R2, the share of variance explained by distance alone; Nakagawa & Schielzeth 2013) and slope (the effect of one standard deviation of distance). Attribute distances holds the pairs and their distance under each hypothesis.

Details

Distances are scaled to mean 0 and standard deviation 1 before fitting, so the slopes are comparable across surfaces. Pairs that cannot be connected through every surface (for example because a barrier isolates a site) are left out of all models, so that every model is fit to the same pairs.

References

Maier PA, Vandergast AG, Ostoja SM, Aguilar A, Bohonak AJ (2022) Landscape genetics of a sub-alpine toad: climate change predicted to induce upward range shifts via asymmetrical migration corridors. Heredity 129:257-272.

Nakagawa S, Schielzeth H (2013) A general and simple method for obtaining R2 from generalized linear mixed-effects models. Methods in Ecology and Evolution 4:133-142.

Examples

# \donttest{
ex <- corridor_example()
fst <- pairwise_fst(read_genotypes(ex$files$structure))
surfaces <- list(slope = ex$resistance_slope, cover = ex$resistance_cover,
                 slope_80_cover_20 = ex$resistance)
rank_resistance(surfaces, ex$sites, fst, max_dist = 15000)
#>          hypothesis   logLik       AIC delta_AIC        R2m       slope
#> 1             slope 2288.989 -4567.978   0.00000 0.03622447 0.009864794
#> 2 slope_80_cover_20 2275.221 -4540.442  27.53579 0.05508055 0.011909024
#> 3             cover 2121.487 -4232.974 335.00414 0.10355020 0.016687517
#> 4     straight_line 2115.328 -4220.656 347.32251 0.01264270 0.006066638
# }