Lab 34 - Mesquite
Zool 575 Introduction to Biosystematics, (Sikes) Winter 2006
This lab will not be graded. We will be available during lab to answer questions (or emaill me at any time). You can complete this lab at your leisure and although Mesquite is on the computers in BI 182, it is a free download and you can install it on any Unix, PC, or Mac computer on which you have permission to do so. The program is available here: http://mesquiteproject.org/mesquite/mesquite.html
This lab is based on the tutorial prepared for Mesquite.
1. Navigate to the Mesquite_folder then to the examples folder inside
Open the introduction.nex file. This will launch Mesquite. You will see the different modules loading as it starts up.
You will be presented with a list of options to explore Mesquite's abilities.
Feel free, of course, to explore any and all of these options. However, for this lab you will work with a select few.
2. Select "Introducing the basic functioning of Mesquite"
Read and page through the windows.
When you get to the character matrix editor Mesquite will show you a sample data matrix. Note that one species is polymorphic regarding its tail color. Click into the cell that is polymorphic and you will see a footnote and other information below - why is this species polymorphic?
Click next when done. Stop at the tree window.
Note the small tabs at the top of the tree window ("graphics", "text", "parameters" etc.) click on "text" and you'll see the same tree in text format (easily copied & pasted). The small blue arrows in the upper left allow you to page through different stored trees.
Click and hold on a branch and then drag to another branch - this should move the branch and change the tree. This allows one to investigate different hypotheses of evolution. When you click next Mesquite will show you all the windows of those tabs & explains them.
Finish the Basic Functioning tour and return to the main menu.
3. Select the "Reconstructing Ancestral States" option
Page through this tutorial. It will open a datafile of morphological characters for beetles of the genus Bembidion - a speciality of David Maddison's. When you see a character traced on the tree (a few pages into the tour) take the opportunity to page through different traced characters (estimated using parsimony).
As you continue through this tutorial you will find explanations of how to change the parsimony model (eg from ordered to unordered [this didn't work with my copy]) and how to set up a step matrix to change the weight of changing from one state to another. This, for example, would help with conducting a sensitivity analysis to determine how much a character change needs to be weighted before the reconstruction changes.
A page will appear that explains differences between MacClade and Mesquite. One of these differences is that MacClade can calculate MPRs for a character but Mesquite cannot (yet).
After clicking through some examples of continuous characters you will arrive at a nice example of ancestral state reconstruction using likelihood.
Read these pages and click onwards to the page that shows an small slider to edit the rate of change for the Mk1 model. (This is really neat!) Read and then try the slider - as you change the rates the ancestral state probabilities change dynamically on the tree!!
The next page will show you the asymmetrical model which has two parameters - a different rate is allowed for gains versus losses (Does this seem more realistic when considering complex characters like wings?)
A few more pages into the tour and you will see a nice side-by-side comparison of an Mk1 to an AsymmMk reconstruction.
Continue on and you will come to a page showing the importance of branch lengths. This is also very informative. Take the ruler tool (should be already chosen by default in this tutorial) and use it to change the length of the longest red branch on the tree on the left. This will change the likelihood reconstruction on the tree on the right - note that the estimation of the ancestral state depends on the length of the branch. How short does the branch have to be to get the probability of state 1 above 75%?
Work through the tutorial for the ancestral state reconstruction. You will see more examples of what one can do with likelihood reconstruction. One interesting point explained is how one can decide among "close" reconstructions: Use the Trace menu to change the score reporting to log-likelihoods. For one state to be significantly preferred over another it should be 2 or more log likelihood units larger.