make_relprob_matrix_bi: Make a relative probability matrix for a single speciation...
In BioGeoBEARS: BioGeography with Bayesian (and Likelihood) Evolutionary Analysis in R Scripts

Description Usage Arguments Details Value Note Author(s) References See Also Examples

Given the identity of the states/geographic ranges on the left branch (Lstates), right branch (Rstates), and ancestral areas (ancareas_txt_tmp), construct the (text version) of the row of transition probabilities. This means that each nonzero cell gets a v for a vicariance event, a y for a sympatric speciation/range-copying event, a j for a founder-event/jump speciation event, and an s for a sympatric-subset event.

1
2
3

  make_relprob_matrix_bi(states_list = default_states_list(),
    split_ABC = FALSE, splitval = "",
    code_for_overlapping_subsets = NA, printwarn = 1)

`states_list`	A list of states, where each state consists of a list of areas. A default example list is provided.
`split_ABC`	`TRUE` or `FALSE` If `TRUE` then each state/range in the input geographic ranges (`states_list`) will be split on the argument contained in `split`.
`splitval`	The character to split on.
`code_for_overlapping_subsets`	Hypothetically, there is no reason that a vicariance event could happen, e.g. ABC–>AB, BC. This is disallowed in `LAGRANGE` BioGeoBEARS defaults, and, if one is going to employ the construct of discrete areas in the first place, overlaps should probably be avoided. But this parameter will allow experimentation. Here, `code_for_overlapping_subsets=NA` equals the default, and any other value means that overlapping vicariance events are included, with a number describing the number of areas in the overlap. Users could then manually convert this to a probability according to some function.
`printwarn`	If `printwarn>0` (`printwarn=1` by default), then print to screen a message describing the size of the cladogenesis matrix.

This function is utilized by apply in other functions (e.g. ) in an attempt to speed up calculation over rows. However, processing of text formulas via apply will never be fast enough for large matrices; see cladoRcpp for optimized functions.

This text-based matrix later gets evaluated by other functions to calculate the numerical probabilities. I.e., if j=0 and the other forms of speciation have weights equal to each other, this is the LAGRANGE cladogenesis model.

NOTE: This function is veeeeeeery slow, even for only 3 areas (i.e. 2^3=8 geographic ranges). It is mostly useful for illustration. See cladoRcpp for drastic improvements in calculating cladogenesis models.

probmat A matrix of strings, where each cell contains the parameters describing the conditional probability of that ancestor–>(Left descendant,Right descendant) range inheritance scenario.

Go BEARS!

Nicholas J. Matzke matzke@berkeley.edu

http://phylo.wikidot.com/matzke-2013-international-biogeography-society-poster https://code.google.com/p/lagrange/

Matzke_2012_IBS

ReeSmith2008

size_species_matrix, make_spmat_row

testval=1
probmat = make_relprob_matrix_bi(states_list=list("_",
c("A"), c("B"), c("C"), c("A","B"), c("B","C"), c("A","C"),
c("A","B","C")), split_ABC=FALSE, splitval="",
code_for_overlapping_subsets=NA, printwarn=1)
probmat