SPASIBA.inf: Inference, prediction for Continuous Spatial Assignment with...
In gilles-guillot/SPASIBA: Continuous Spatial Assignment

Description Usage Arguments Value Author(s)

Inference, prediction for Continuous Spatial Assignment with INLA

SPASIBA.inf(
  geno.ref,
  ploidy,
  coord.ref,
  sphere = FALSE,
  size.pop.ref,
  geno.unknown,
  quant.ref = NULL,
  quant.unknown = NULL,
  make.inf = FALSE,
  loc.infcov = NULL,
  kappa.geno.plug = NULL,
  tau.geno.plug = NULL,
  window = NULL,
  max.edge.mesh = NULL,
  offset.mesh = -0.1,
  marginal.quant = "none",
  make.pred = FALSE,
  nx.pred = 20,
  ny.pred = 20,
  make.assign = FALSE,
  proj.predref = NULL,
  freq.predref = NULL,
  verbose.inf = TRUE,
  verbose.pred = TRUE
)

`geno.ref:`	Matrix of allele counts of reference data. One row per population, one column per locus.
`ploidy:`	1/2
`coord.ref:`	coordinates of reference sampling sites. One row per sampling site, two columns (cartesian or lon-lat)
`sphere:`	Logical (`TRUE` or `FALSE`) indicating whether samples are living on a sphere or a flat domain
`size.pop.ref:`	Matrix with one row per population and one column per locus giving haploid population sample sizes of the various reference populations
`geno.unknown:`	Genotypes of individuals of unknwon geographic origin. One row per indivdual, one column per locus. This should contain allele counts of an arbitrary reference allele at each locus (hence 0,1, 2 or NA)
`quant.ref:`	Matrix of quantitative covariates for reference individuals. One row per individual. Currently not implemented.
`quant.unknown:`	Matrix of quantitative covariates for individuals of unknwon geographic origin. Currently not implemented.
`make.inf:`	Logical (`TRUE` or FALSE) indicating whether inference of covariance function (under the GMRF-SPDE model) should be carried out
`loc.infcov:`	Subset of loci to be used to carry out inference of covariance function (GMRF-SPDE model)
`kappa.geno.plug:`	Spatial scale parameter kappa of the hidden Gaussian random field (if known from inference carried out with the present function earlier)
`tau.geno.plug:`	Precision (inverse of variance) of the hidden Gaussian random field (if known from inference carried out earlier with the present function)
`window:`	A 2x2 matrix specifying the xmin, xmax, ymin, ymax of the rectangular window over which predicted allele frequencies maps will be computed. As default, the smallest rectangular window with edges parrallel to the axes will be used.
`max.edge.mesh:`	Max length of an edge in triangulation for GMRF-SPDE model. Change default value only if you are familiar with the triangulation function for the SPDE model in INLA.
`offset.mesh:`	Offset in mesh for GMRF-SPDE model. Change default value only if you are familiar with the triangulation function for the SPDE model in INLA.
`marginal.quant:`	Distribution of quantitative variables (norm or lnorm). Currently not implemented.
`make.pred:`	TRUE/FALSE indicating whether prediction of allele frequencies should be carried out
`nx.pred:`	Number of pixels horizontally in allele frequencies prediction
`ny.pred:`	Number of pixels vertically in allele frequencies prediction
`make.assign:`	TRUE/FALSE indicating whether assignment of individuals of unknown origin should be made
`proj.predref:`	Projection of grid used in prediction (typically as returned by a previous call to the present function when make.pred=TRUE)
`freq.predref:`	Allele frequencies (returned by this function when make.pred=TRUE)
`verbose.inf:`	TRUE if you want to get the `inla()` blurb
`verbose.pred:`	TRUE if you want to get the `inla()` blurb

A list with components below (some of them being only optionally returned). The most important in the list returned are ll and coord.unknown.est.

mesh.ref: An object of class inla.mesh containing info about the mesh used under the inference of the GMRF-SPDE parameters. See ?inla.mesh.2d for details.
cpu.used.infcov: CPU time used during the inference of the GMRF-SPDE parameters.
tau.geno.est: Precision (inverse of the variance) in the Matern field involved in the GMRF-SPDE model
kappa.geno.est: Scale parameter in the Matern field involved in the GMRF-SPDE model
make.inf: Logical (TRUE or FALSE) indicating whether the last call to the present function requested to estimate parameters of the GMRF-SPDE model.
coord.pred: Coordinates of the regular grid at which allele frequencies are predicted.
mesh.predref: Mesh (an object of class inla.mesh) used involved in the prediction of allele frequencies. See ?inla.mesh.2d for details.
proj.predref: An object typically returned by inla.mesh.project used to convert info between an irregular trangulation and a regular grid. See ?inla.mesh.projector for details.
u: The x component of a proj.predref object
v: The y component of a proj.predref object
cpu.used.predfreq.perloc: CPU time spent computing allele frequency maps
freq.predref: An array with dimensions (nx.pred,ny.pred,nloc) containing predicted frequency maps, with (nx.pred,ny.pred) dimension of the grid, nloc: number of loci.
ll: An array with dimensions (nx.pred,ny.pred,n.unknown) containing the log-likelihood of each individual to have origin at the various nodes of a regular grid. (nx.pred,ny.pred) dimension of the grid, n.unknown: number of individuals of unknown origin.
coord.unknown.est: The estimated geographic coordinates of the individuals.