landgenreport {PopGenReport}

Create a landscape genetic report

Description

This function is the landscape genetic version of the popgenreport function. It needs to be provided with a genind object with spatial coordinates, a friction map (raster) and a specification which type of genetic distance should be used. Once all three type of input are provided with the necessary input, a landscape genetic analysis using least cost path analysis is computed (see Cushman et al. 2010, Landguth et al. 2010). Depending on the genetic distance meassurement this is done on a subpopulation basis (D, Gst.Hedrick, Gst.Nei=Fst) or on an individual basis (Kosman, Smouse).

Usage

landgenreport(
  cats,
  fric.raster,
  gen.distance = "Gst.Nei",
  NN = NULL,
  pathtype = "leastcost",
  plotpath = TRUE,
  theta = 1,
  mk.resistance = TRUE,
  mapdotcolor = "blue",
  mapdotsize = 1,
  mapdotalpha = 0.4,
  mapdottype = 19,
  mapzoom = NULL,
  mk.custom = FALSE,
  fname = "LandGenReport",
  foldername = "results",
  path.pgr = NULL,
  mk.Rcode = FALSE,
  mk.complete = FALSE,
  mk.pdf = TRUE
)

Arguments

cats

a genind object with spatial coordinates in the other slot

fric.raster

friction (resistance) raster, that specifies the landscape where the analysis should be computed on. If fric.raser is a stack a cost distances are calculated for each layer in the stack.

gen.distance

type of genetic distance that should be used. Depending on the genetic distance meassurement this is done on a subpopulation basis (D, Gst.Hedrick, Gst.Nei=Fst) or on an individual basis (Kosman, Smouse, propShared). propShared is the proportion of shared alleles between individuals.

NN

Number of neighbours used when calculating the cost distance (possible values 4,8 or 16). As the default is NULL a value has to be provided if pathtype is 'leastcost'. NN=8 is most commonly used as it avoids a directional bias, but be aware that linear structures may cause artefacts in the least-cost paths in the NN=8 case, therefore we strongly recommend to inspect the actual least-cost paths in the provided output.

pathtype

Type of cost distance to be calculated (based on function in the gdistance package. Available distances are 'leastcost', 'commute' or 'rSPDistance'. See functions in the gdistance package for futher explanations.

plotpath

switch if least cost paths should be plotted (works only if pathtype='leastcost'. Be aware this slows down the computation, but it is recommended to check least cost paths visually.

theta

value needed for rSPDistance function. see rSPDistance in package gdistance.

mk.resistance

switch to do the landscape genetic analysis based on resistance matrices, should be set to TRUE

mapdotcolor

see popgenreport

mapdotsize

see popgenreport

mapdotalpha

seepopgenreport

mapdottype

see popgenreport

mapzoom

see popgenreport

mk.custom

switch to add a customised part to the landgenreport

fname

see popgenreport

foldername

see popgenreport

path.pgr

see popgenreport

mk.Rcode

see popgenreport

mk.complete

see popgenreport

mk.pdf

see popgenreport

Details

Check the help pages of popgenreport how to include coordinates to a genind object. The coordinates need to be projected. Latlongs are not valid, because Euclidean distances are calcuated based on these coordinates. For an example how to convert latlongs into a projected format have a look at the vignette that comes with this package. The friction needs to be a raster and needs to be in the same projection as the genind object. Also the type of genetic distance to be used needs to be specified.

Value

Four distance matrices are returned. Pairwise Euclidean distances between subpopulations/individuals, cost distances, path lengths and genetic distances. Also following the approach of Wassermann et al. 2010 a series of partial mantel tests are performed. A multiple regression analysis based on Wang 2013 and Legendre 1994 is returned.The actual least-cost paths can be found under paths

Author(s)

Bernd Gruber ([e-mail])

References

Cushman, S., Wasserman, T., Landguth, E. and Shirk, A. (2013). Re-Evaluating Causal Modeling with Mantel Tests in Landscape Genetics. Diversity, 5(1), 51-72.

Landguth, E. L., Cushman, S. A., Schwartz, M. K., McKelvey, K. S., Murphy, M. and Luikart, G. (2010). Quantifying the lag time to detect barriers in landscape genetics. Molecular ecology, 4179-4191.

Wang,I 2013. Examining the full effects of landscape heterogeneity on spatial genetic variation: a multiple matrix regression approach for quantifying geographic and ecological isolation. Evolution: 67-12: 3403-3411.

Wasserman, T. N., Cushman, S. A., Schwartz, M. K. and Wallin, D. O. (2010). Spatial scaling and multi-model inference in landscape genetics: Martes americana in northern Idaho. Landscape Ecology, 25(10), 1601-1612.

See Also

popgenreport, wassermann, genleastcost, lgrMMRR

Examples

library(raster)
fric.raster <- readRDS(system.file("extdata","fric.raster.rdata", package="PopGenReport"))
lc<-landgenreport(cats=landgen, fric.raster=fric.raster,
gen.distance="D", NN=4, mk.resistance=TRUE, mk.pdf=FALSE)
names(lc$leastcost)