estim {mvMORPH}

Ancestral states reconstructions and missing value imputation with phylogenetic/time-series models

Description

This function imputes the missing cases (NA values) according to a given phylogenetic model (object of class "mvmorph"); it can also do ancestral state reconstruction.

Usage

estim(tree, data, object, error=NULL, asr=FALSE)

Arguments

tree

Phylogenetic tree (an object of class "phylo" or "simmap") or a time-series.

data

Matrix or data frame with species in rows and continuous traits with missing cases (NA values) in columns (preferentially with names and in the same order than in the tree).

object

A fitted object from an mvMORPH model (class "mvmorph").

error

Matrix or data frame with species in rows and continuous traits sampling variance (squared standard errors) in columns.

asr

If asr=TRUE, the ancestral states are estimated instead of the missing cases.

Details

Missing observations for species in a phylogenetic tree are estimated according to a given evolutionary model (and parameters). Multivariate models are useful to recover the variance and covariance structure of the dataset to be imputed.

When asr=TRUE, the estimates, their variances and standard errors are those of the ancestral states at each node of the tree (this option is not available for the time-series). Note that if there are missing cases, they are first imputed before estimating the ancestral states.

Estimation of missing cases and ancestral states is performed using GLS (Generalized Least Squares) solution (See Cunningham et al. 1998).

Value

estimates

The imputed dataset

var

Variance of the estimates

se

Standard error of the estimates

NA_index

Position of the missing cases in the dataset

Author(s)

Julien Clavel

References

Clavel J., Merceron G., Escarguel G. 2014. Missing Data Estimation in Morphometrics: How Much is Too Much? Syst. Biol. 63:203-218.

Cunningham C.W., Omland K.E., Oakley T.H. 1998. Reconstructing ancestral character states: a critical reappraisal. Trends Ecol. Evol. 13:361-366.

See Also

mvMORPH mvOU mvEB mvBM mvSHIFT

Examples

## Simulated dataset
set.seed(14)
# Generating a random tree
tree<-pbtree(n=50)

# Setting the regime states of tip species
sta<-as.vector(c(rep("Forest",20),rep("Savannah",30))); names(sta)<-tree$tip.label

# Making the simmap tree with mapped states
tree<-make.simmap(tree,sta , model="ER", nsim=1)
col<-c("blue","orange"); names(col)<-c("Forest","Savannah")

# Plot of the phylogeny for illustration
plotSimmap(tree,col,fsize=0.6,node.numbers=FALSE,lwd=3, pts=FALSE)

# Simulate two correlated traits evolving along the phylogeny
traits<-mvSIM(tree,nsim=1, model="BMM", param=list(sigma=list(matrix(c(2,1,1,1.5),2,2),
         matrix(c(4,1,1,4),2,2)), names_traits=c("head.size","mouth.size")))

# Introduce some missing cases (NA values)
data<-traits
data[8,2]<-NA
data[25,1]<-NA

# Fit of model 1
fit<-mvBM(tree,data,model="BMM")

# Estimate the missing cases
imp<-estim(tree, data, fit)

# Check the imputed data
imp$estim[1:10,]

## We want the ancestral states values at each nodes:
nodelabels() # To see where the nodes are situated

imp2<-estim(tree, data, fit, asr=TRUE)

# Check the 10 firsts ancestral states
imp2$estim[1:10,]