HZ: Estimate parameters of the hybrid zone model by MCMC...

Description Usage Arguments Value Author(s) References

Description

Estimate parameters of a hybrid zone model by MCMC simulation. The function does not currently accept more than one genotype matrix.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
            HZ(coordinates,
               geno.dip.codom=NULL,
               geno.dip.dom=NULL,
               geno.hap=NULL,
               dist.IC=NULL,
               allele.freq=NULL,
               ncluster=NULL,
               cluster.indiv=NULL,
               path.mcmc.noadm=NULL,
               a.init=NULL,
               b.init=NULL,
               c.init=1,
               a.max=10,
               b.max=NULL,
               c.max=1,
               estimate.a=TRUE,
               estimate.b=TRUE,
               estimate.c=FALSE,
               common.param=TRUE,
               nit,
               thinning,
               path.mcmc.adm=NULL)  

Arguments

coordinates

Spatial coordinates of individuals. A matrix with 2

geno.dip.codom

Genotypes for diploid data with codominant markers. A matrix with one line per individual and two columns per locus.

geno.dip.dom

Genotypes for diploid data with dominant markers. A matrix with one line per individual and one column per locus. Presence/absence of a band should be coded as 0/1 (0 for absence / 1 for presence). Dominant and codominant markers can be analyzed jointly by passing variables to arguments geno.dip.codom and geno.dip.dom. Haploid data and diploid dominant data can not be analyzed jointly in the current version.

geno.hap

Genotypes of haploid data. A matrix with one line per individual and one column per locus. Dominant diploid data and haploid data can be analyzed jointly (e.g. to analyse microsatelite data or SNP data together with mtDNA. Haploid data and diploid dominant data can not be analyzed jointly in the current version.

dist.IC

A matrix with nindiv lines and ncluster column. The parameter ncluster being the number of clusters, most often 2 when there is a single hybrid zone. If dist.IC is missing, the user has to provide instead the path to a directory storing results from a no-admixture MCMC run.

allele.freq

An array with ncluster x nloc x nalmax If missing, it will be estimated from the output from the MCMC run to estimate clusters.

ncluster

Number of clusters. If missing, the user has to provide instead the path to a directory storing results from a no-admixture MCMC run

cluster.indiv

Cluster membership of individuals. A numeric vector with integer values (maximum values being the total number of clusters)

path.mcmc.noadm

Path to output files directory of the previous Geneland no-admixture run. It seems that the path should be given in the Unix style even under Windows (use \/ instead of \). This path *has to* end with a slash (\/) (e.g. path.mcmc="/home/me/Geneland-noadmixture/")

a.init

A numerical value to use as fixed or initial value for the a parameter

b.init

A numerical value to use as fixed or initial value for the b parameter

c.init

A numerical value to use as fixed or initial value for the c parameter

a.max

Maximum value allowed along MCMC simulation for parameter a (default is 1)

b.max

Maximum value allowed along MCMC simulation for parameter b (default is a small fraction of the study area diameter)

c.max

Maximum value allowed along MCMC simulation for parameter c (default is 1)

estimate.a

Logical. If TRUE, parameter a is estimated, if FALSE it is left at the initial value.

estimate.b

Logical. If TRUE, parameter b is estimated, if FALSE it is left at the initial value.

estimate.c

Logical. If TRUE, parameter c is estimated, if FALSE it is left at the initial value.

common.param

If TRUE, parameters a, b and c are common to all clusters. If FALSE, the program attemps to estimate cluster-specific values.

nit

Number of MCMC iteration.

thinning

Number of MCMC iterations between two writing steps (if thinning=1, all states are saved whereas if e.g. thinning=10 only each 10 iteration is saved)

path.mcmc.adm

Path to output files directory for the admixture model. It seems that the path should be given in the Unix style even under Windows (use \/ instead of \). This path *has to* end with a slash (\/) (e.g. path.mcmc="/home/me/Geneland-admixture/")

Value

No object is returned. All outputs are stored in ascii file located in the path.mcmc directory

Author(s)

G. Guillot

References

B. Guedj and G. Guillot, A Bayesian model for infering hybrid zones.


Geneland documentation built on April 14, 2017, 2:31 p.m.