Artificial Extinction

Source: R/ArtificialExtinction.R

Remove tokens that do not occur in a fossil "template" taxon from a living taxon, to simulate the process of fossilization in removing data from a phylogenetic dataset.

ArtificialExtinction(
  dataset,
  subject,
  template,
  replaceAmbiguous = "ambig",
  replaceCoded = "original",
  replaceAll = TRUE,
  sampleFrom = NULL
)

# S3 method for matrix
ArtificialExtinction(
  dataset,
  subject,
  template,
  replaceAmbiguous = "ambig",
  replaceCoded = "original",
  replaceAll = TRUE,
  sampleFrom = NULL
)

# S3 method for phyDat
ArtificialExtinction(
  dataset,
  subject,
  template,
  replaceAmbiguous = "ambig",
  replaceCoded = "original",
  replaceAll = TRUE,
  sampleFrom = NULL
)

ArtEx(
  dataset,
  subject,
  template,
  replaceAmbiguous = "ambig",
  replaceCoded = "original",
  replaceAll = TRUE,
  sampleFrom = NULL
)

Arguments

dataset

Phylogenetic dataset of class phyDat or matrix.

subject

Vector identifying subject taxa, by name or index.

template

Character or integer identifying taxon to use as a template.

replaceAmbiguous, replaceCoded

Character specifying whether tokens that are ambiguous (?) or coded (not ?) in the fossil template should be replaced with:

  • original: Their original value; i.e. no change;

  • ambiguous: The ambiguous token, ?;

  • binary: The tokens 0 or 1, with equal probability;

  • uniform: One of the tokens present in sampleFrom, with equal probability;

  • sample: One of the tokens present in sampleFrom, sampled according to their frequency.

replaceAll

Logical: if TRUE, replace all tokens in a subject; if FALSE, leave any ambiguous tokens (?) ambiguous.

sampleFrom

Vector identifying a subset of characters from which to sample replacement tokens. If NULL, replacement tokens will be sampled from the initial states of all taxa not used as a template (including the subjects).

Value

A dataset with the same class as dataset in which entries that are ambiguous in template are made ambiguous in subject.

Details

Further details are provided in Asher and Smith (2022) .

Note: this simple implementation does not account for character contingency, e.g. characters whose absence imposes inapplicable or absent tokens on dependent characters.

References

Asher R, Smith MR (2022). “Phylogenetic signal and bias in paleontology.” Systematic Biology, 71(4), 986--1008. doi:10.1093/sysbio/syab072 .

Examples

set.seed(1)
dataset <- matrix(c(sample(0:2, 4 * 8, TRUE),
                    "0", "0", rep("?", 6)), nrow = 5,
                    dimnames = list(c(LETTERS[1:4], "FOSSIL"),
                                    paste("char", 1:8)), byrow = TRUE)
artex <- ArtificialExtinction(dataset, c("A", "C"), "FOSSIL")