Wayne P. Maddison and David R. Maddison
This page contains some commonly asked questions about MacClade 3, or questions that illustrate particular techniques in using MacClade.
A: As far as we know. We have not experienced problems running MacClade with this system.
A: We don't think so. We are preparing version 4, which is designed to operate properly under Mac OS X.
A: When MacClade 3.05 or later starts up, it looks for the font MacCladeSmall. You need to install this font by either placing the MacClade Font File that came with MacClade 3.05 in the same folder as MacClade, or by installing it into your system. If you have System 7 or later, this can be accomplished by placing the Font File into you System Folder; if you have an earlier system, you should install the Font File using the Font/DA mover, as documented in your Macintosh documentation.
A: MacClade 3 runs on PowerMacs, but slowly. MacClade 3 is compiled for the older generation of Macintoshes (those with a 680x0 processor), not a PowerMac (which uses a PowerPC processor). A PowerMac has to run MacClade 3 under emulation (pretending to be a 680x0 machine), and so is slower than it would be if it were compiled for a PowerMac.
Version 4 of MacClade is available in 680x0 and PowerPC versions.
A: MacClade has been able since version 3.0 to print trees with branch lengths proportional to amount of reconstructed change (p. 349). Since version 3.02 the screen display in the tree window could also show branch lengths proportional to change This feature is available only when Trace All Changes is active, because it obtains the lengths from the calculations of Trace All Changes. MacClade cannot yet read in previously calculated or user-specified branch lengths. If you nonetheless want to use MacClade to draw a tree with a set of branch lengths that you specify, you can create a fictitious set of characters for your taxa, one binary character for each branch, and weight each character proportionally to the desired length of the corresponding branch. Then, activate Trace All Changes using the weighted option, making sure that those ficticious characters are the only included characters.
A: You can remember the original positions by naming them before characters are deleted. Do this by selecting the row of character names (or column, if the matrix is transposed) in the Data Editor. The character names are in the cells between the data themselves and the character numbers. To select the whole row, click on the cell for the first character's name, then while holding down Shift, touch on the cell for the last character's name. Then choose Fill from the Utilities menu (pp. 173 and 328 of the MacClade book). If you indicate, for instance, to fill the selected characters with "s" plus numbers beginning with 1, the characters will be named "s1", "s2", and so on. They will retain these names when some are deleted, thus recording their original positions. You might also want to make sure you set the codon positions (pp. 325-327 of book) before deleting the characters so the original codon positions are remembered as well.
A: Yes. If you have a DNA or RNA data file and you hold down the option key when you are in the Data Editor, you will see that the Match First menu item under Display changes to become the Color States menu item. Select it, and it will show A, C, G and T as different colors. (This rather primitive display has been significantly improved in MacClade 4.)
A: MacClade's Store Tree stores the tree alone; it does not store a snapshot of the tree including character tracing and so forth as it appeared when stored. Thus, when you ask for the tree again, the tree is restored, but the calculations and graphics done with it are those that are in effect when the tree is restored. The only tree that has fixings of states stored with it is the tree on screen at the time a data file is saved. The trees saved with Store Tree lose their fixings of states.
A: MacClade can be used to produce a table listing the length of the tree for each possible rooting of a clade (p. 221). If all of your characters are undirected (unordered, ordered, Dollo, symmetrical stepmatrix not violating the triangle inequality), then this feature can be used to tell you what you need to know. First, reroot the tree so the root is between the unstable clade and the rest of the tree. This will not affect the treelength because the characters are undirected. Then, expand what is now the clade of the rest of the tree using the magnifying glass tool so it occupies the tree window. At this point, it is a good idea to print the tree with branch numbers shown (p. 233). Then hold down the Option key and choose the All Rootings menu item in the Trees menu (as described on p. 221), then do a Treelengths chart (p. 285), saving lengths to a text file (p. 275). The text file resulting will show the lengths of trees named according to the number of the branch on which the clade was rerooted to make that tree. You will need to turn over the clade in your mind to place its root back in its original position to interpret these results, or you could print the tree in its original form and mark on its branches the numbers those branches had received in the rerooted version.
A: Use MacClade's Evolve Characters to simulate character evolution under the null model. Set the ancestral state to 0 and choose a fairly low probability of change. Evolve lots of characters on your tree, then do a Changes & Stasis chart to show reconstructed numbers of gains and losses. This will give you a notion of the average reconstructed imbalance between gains and losses. To find the probability of a character showing all gains given it showed five changes, you could find all the simulated characters showing 5 steps (e.g. in the Character Status window), exclude all other characters, and then do a Steps/etc. chart by character (top middle option) set for changes and applying the restriction of only 0 to 1 changes. Count what fraction of the characters that show 5 changes so restricted.