estimate.pgls: Estimates Number of Samples Needed to Change Significance for...

In bhjuarez/estimatoR.2016: Estimates Number of Samples Needed to Change Significance for Phylogenetic and Non-Phylogenetic Statistical Tests

Description

Tests for effect of known missing samples on significance of phylogenetic regression. It does so by iteratively simulating additional data and species on the phylogeny, until the significance of the regression changes.

Usage

estimate.pgls(dat, phy, est)

Arguments

dat	Data frame containing, data in first column and group assignment (using integers) in second column. Rownames corresponding to tip labels of tree.
phy	An object of type 'phy', a rooted phylogeny. Can be ultrametric or not. Tip labels corresponding to rownames of data table.
est	Number of missing taxa of the focal group (if known). This value will act as a maximum limit for iterations adding new species.

Value

Matrix with number of iterations (which corresponds to missing taxa to be sampled) in first column and new p-values for each iteration in second column.

Author(s)

Bryan H. Juarez, Orlando Schwery, Giulio Valentino Dalla Riva

See Also

estimate.lm, estimatoR-package

Examples

  fun <- dat[, 1] ~ as.vector(dat[, 2])  # define function
  dat <- cbind(rnorm(50, 50, 105), rep(1, 50))  # simulate data group 1
  dat <- rbind(dat, cbind(rnorm(50, 110, 90), rep(2, 50)))  # simulate data group 2
  colnames(dat) <- c("length", "group")  # give colnames
  rownames(dat) <- paste("t", seq(1:100), sep = "")  # name species to match tree tip labels later
  dat <- as.data.frame(dat)  # convert to data frame
  est <- 10  # set number of estimates (best set to number of known unsampled taxa)
  tree <- rtree(100, rooted = T, tip.label = paste("t", seq(1:100), sep = ""))  # simulate tree with matching tiplabels
  estimate.pgls(dat, tree, est) # test running

bhjuarez/estimatoR.2016 documentation built on May 12, 2019, 8:25 p.m.