R: Modeling Phylogenetic Signals using Eigenvector Maps

MPSEM-package {MPSEM}

R Documentation

Modeling Phylogenetic Signals using Eigenvector Maps

Description

Computational tools to represent phylogenetic signals using adapted eigenvector maps.

Details

Phylogenetic eignevector maps (PEM) is a method for using phylogeny to model features of organism, most notably quantitative traits. It consists in calculating sets of explanatory variables (eigenvectors) that are meant to represent different patters in trait values that are likely to have been inducted by evolution. These patterns are used to model the data (using a linear model for instance).

If one gets a ‘target’ species (i.e. a species for which the trait value is unknown), and providing that we know the phylogenetic relationships between that species and those of the model, the method allows to obtain the scores of that new species on the phylogenetic eigenfunctions underlying a PEM. These scores are used to make empirical predictions of trait values for the target species on the basis of those observed for the species of the model.

Functions PEM.build, PEM.updater, PEM.fitSimple, and PEM.forcedSimple allows one to build, update (i.e. recalculate with alternate weighting parameters) as well as to estimate or force arbitrary values for the weighting function parameters.

Functions getGraphLocations and Locations2PEMscores allows one to make predictions using method predict.PEM and a linear model. To obtain these linear model, user can use function lm or auxiliary functions lmforwardsequentialsidak or lmforwardsequentialAICc, which perform forward-stepwise variable addition on the basis of either familiwise type I error rate or the Akaike Information Criterion (AIC), respectively.

The package provides low-level utility function for performing operation on graphs (see graph-functions), calculate influence matrix (PEMInfluence), and simulate trait values (see trait-simulator).

A phylogenetic modeling tutorial using MPSEM is available as a package vignette (see example below).

The DESCRIPTION file:

Package:	MPSEM
Type:	Package
Version:	0.4-1
Date:	2022-01-10
Encoding:	UTF-8
Title:	Modeling Phylogenetic Signals using Eigenvector Maps
Author:	Guillaume Guenard, with contribution from Pierre Legendre
Maintainer:	Guillaume Guenard <guillaume.guenard@gmail.com>
Description:	Computational tools to represent phylogenetic signals using adapted eigenvector maps.
Depends:	R (>= 3.0.0), ape
Suggests:	knitr, caper, xtable
Imports:	MASS
License:	GPL-3
LazyLoad:	yes
NeedsCompilation:	yes
VignetteBuilder:	knitr
Repository:	CRAN
RoxygenNote:	7.1.2

Index of help topics:

MPSEM-package           Modeling Phylogenetic Signals using Eigenvector
                        Maps
PEM-class               Class and Methods for Phylogenetic Eigenvector
                        Maps (PEM)
PEM-functions           Phylogenetic Eigenvector Maps
graph-class             Class and Method for Directed Graphs
graph-functions         Graph Manipulation Functions
lm-utils                Linear Modelling Utility Functions
trait-simulator         Simulates the Evolution of a Quantitative
                        Trait.

Author(s)

Guillaume Guenard, with contribution from Pierre Legendre Maintainer: Guillaume Guenard <guillaume.guenard@gmail.com>

References

Guénard, G., Legendre, P., and Peres-Neto, P. 2013. Phylogenetic eigenvector maps (PEM): a framework to model and predict species traits. Meth. Ecol. Evol. 4: 1120–1131

Examples

## To view MPSEM tutorial
vignette("MPSEM", package="MPSEM")