PHENIX (version 1.3.2)

cor.par: Partial correlation

Description

This function calculates the partial correlation matrix between a set of traits and a third control variable.

Usage

cor.par(traits, c.trait, trait.names = FALSE, silent=FALSE)

Value

A matrix with the pairwise partial correlation between traits.

Arguments

traits

a dataframe or a matrix object containing traits as columns and individuals as rows

c.trait

a vector containing the third control variable to estimate partial correlations, for instance: organ size or individual size

trait.names

a logical. If 'FALSE' (default), trait names in the output are defined as "Trait 1", "Trait 2", ..., "Trait n". If 'TRUE', output trait names are the same than those in the input dataframe

silent

a logical. If 'FALSE' (default), return a warning message if any row contains missing data (and is removed from the original dataset before computation).

Author

R. Torices, A.J. Muñoz-Pajares

Details

This function is utilised internally by pintsc and pintsc.boot to estimate the phenotypic integration index from partial correlations. Rows containing missing values are removed.

Examples

# Dataset from Torices & Méndez (2014)
# This data set represents the dry mass (in grams) of inflorescence components of the sunflower
# species Tussilago farfara. The inflorescences were dissected in 'SCAPE', 'RECEPTACLE', 'MALEFL'
# (male flowers), 'OVAR' (reproductive part of female flowers), and 'RAYS' (the petaloid ray of
# female flower). Furthermore in the last column the total weight of the inflorescence is added
# 'TOTALSIZE'

data(tussilago)
print(tussilago)
cor.par(tussilago[,1:5], tussilago[,6], trait.names=TRUE) # NA automatically removed

traits<-na.exclude(tussilago) # NA manually removed
cor.par(traits[,1:5], traits[,6], trait.names=TRUE)