Preparing files for the SOWH Test

In order to conduct an SOWH test, you need to have a Mesquite file (such as this example file, which you can download and open in Mesquite) containing the following elements:
  1. The data matrix to be analyzed.
  2. A constraint tree that embodies the general aspects of the hypothesis you wish to test; this tree needs to be in a tree window in the file. The example file has the constraint tree shown in a window called "Constraint tree"
  3. The optimal tree found from a search in which the constraint tree is enforced; that is, you need to have an optimal, constrained tree; this tree needs to be in a tree window in the file. The example file has the constraint tree shown in a window called "Hypothesis to test".
  4. Models of character evolution. If DNA data, then you should have a Composite DNA Simulation Model, consisting of various components, including a state frequency submodel, a site-to-site rate variation submodel, and a rate matrix model. The parameter values for these models should come from a tree analysis in a likelihood program (RAxML, GARLI, or PAUP), and should be the values inferred on the optimal constrained tree. In the example file, the Composite DNA Simulation Model is called "GTR+G+I", and its values were inferred from a PAUP analyses.
  5. Have the optimal constrained tree (the hypothesis tree you are attempting to test) visible in the frontmost window.