SIMMAP | sim histories

Simulating Character Histories

This topic is spread throughout the help documents of SIMMAP and therefore this page will attempt to bring it all together by providing a general overview, links to pages with the details of each step, and filling any gaps in the help documents on simulating histories. A brief guide to simulating a history is described in the Tutorial. The tutorial explains simulating single character histories and visualizing them so will not be dealt with here. That being said each of steps described below apply to both simulating a single history or many histories.

There are six basic steps to simulate character histories. Each will be described briefly later but it is recommended that you read the associated links first as they deal with these steps in greater detail.

Six basic steps:

Open data, tree, and parameter (molecular data only) files [Input Files]
Configure character coding [Character Coding]

Character coding defines how a character is dealt with during the analysis. For molecular data this defines the protein code - first, second, and third codon positions. For morphological data this defines whether characters are ordered or unordered.

Configure evolutionary model

[Molecular data: Substitution Models]
[Morphological data: Substitution Models, Morphology Priors]

Configure analysis [Tutorial, Predictive Distributions]
Run and monitor analysis [Tutorial]
Analyzing results [Output files, Molecular Statistics, Morphological Statistics, and Tree Histories]

Open Data, Tree, and Parameter Files

To open a data file which contains a Nexus representation of the character matrix and optionally a trees block select the Open Data... [O] in the File main menu. If successful the Data View and Tree View (if trees are included in the data file) windows will be displayed.

To open a tree file which contains a Nexus representation of a collection of trees in Newick format select the Open Trees... [T] in the File main menu. If successful the Tree View window will be displayed.

To open a parameter file (molecular data only) which contains a tab-delimited representation of model parameters select the Open Parameters... [R] in the File main menu. To view the parameters loaded into memory select Parameter View from the Window main menu.

Once one or more of these files has been successfully opened specific file information can be seen by requesting more file information from the Window main menu by selecting Show File Inspector [I], or clicking the More Info buttons in the Data View, Tree View, and Parameter View windows.

Configure Character Coding

Character coding is a definition of how each character is dealt with during a character history analysis.

When dealing with molecular data the user can opt to summarize the results of the character histories as single independent nucleotides or as a triplet, or codon. Please note that SIMMAP currently simulates character histories under the assumption that each site, a nucleotide, is evolving independently of other sites. This being said SIMMAP can analyze patterns of codon or amino acid evolution by summarizing histories across codon positions.

When dealing with morphological data the user can opt to have characters evolve in an "ordered" or "unordered" fashion. Ordering of character evolution is discussed in the help documents on Character Coding.

Configure Evolutionary Model

The details of the evolutionary model is the basis of how SIMMAP stochastically samples character histories. A wide variety of models are available for both molecular and morphological data (see Substitution Models and Morphology Priors, depending on the datatype). While these links discuss in detail how to configure the evolutionary model please consider the following:

SIMMAP has a set of default values for many of the options available to the user. These values are not an endorsement for their use. Please think critically about what values you are using during an analysis as the defaults may not be appropriate for YOUR data. This is most important when considering prior values on morphological character evolution (e.g., gamma rate, beta bias prior, and rescaling of the tree). Priors can be both useful and dangerous if miss-applied. You should check that the posterior probability, sampled from the prior, is not stacked up against one end (a boundary) of the prior distribution (i.e., you are sampling only the largest or smallest values and some probable values exist beyond this boundary). If this is happening it is likely that you are excluding probable values and the prior is too restrictive. Try a couple of reasonable parameters for the morphological priors.

Configure Analysis

Configuring a character history analysis may take some time and reading of the help documents to do what you want. While I absolutely recommend that you read the relevant help documents the following is a description of the options, their effects, and what you need to do.

Run and Monitor Analysis

Analyzing Results


	Help Documentation Current document


	Page Last Updated: 6 August 2008