Introduction to MacClade

Portions of this page are from the manual that accompanies MacClade.
Copyright © 2002 David R. Maddison and Wayne P. Maddison

MacClade is used as a tool for phylogenetic analysis. But by its nature it also embodies a world view, a portrayal of a phylogenetic approach to studying diversity and evolution. It is relatively easy to see the diversity of living organisms, but it has proved more difficult to see that diversity in terms of its history; the slow development of a thoroughly phylogenetic perspective in biology attests to this challenge. Together the MacClade manual and program present methods for analyzing and exploring phylogenetic hypotheses, including hypotheses about character evolution. MacClade is one attempt to give our mind's eye phylogenetic lenses, to help us think about and see lineages and evolution.

The MacClade CD comes with versions of the program MacClade for both MacOS X and earlier versions of the MacOS (7.5 through 9.2.2). The CD also contains a complete, 502-page manual in PDF format (with linked movies and examples).

MacClade provides an interactive environment for exploring phylogeny. In MacClade's tree window, hypothesized phylogenetic trees or cladograms can be manipulated and character evolution visualized upon them. To manipulate the tree, tools are provided to move branches, reroot clades, create polytomies, and search automatically for more parsimonious trees. Character evolution is reconstructed on the tree and indicated by "painting" the branches. Alternative reconstructions of character evolution can be explored. Summaries of changes in all characters can be depicted on the tree. As trees are manipulated, MacClade updates statistics such as treelength and the results are depicted in graphics and charts.

Examining character evolution on a tree in MacClade's tree window

MacClade provides charts summarizing various aspects of character evolution on one or more trees, as well as charts comparing two or more trees. For example, the charts can show the number of trees of each length, the number of characters on the tree with different consistency indices, and so on. There are charts specifically designed for DNA/RNA sequence data, including one showing the number of changes on the tree at first, second, and third codon positions, and a chart of the relative frequencies of various transitions and transversions.

A chart showing the relative frequencies of reconstructed changes between bases on a tree,
with area of circle proportional to frequency of change

In MacClade's data editor, systematic and comparative data are entered and edited. The editor has numerous features for manipulating rows, columns, and blocks of data, and for recoding data.

Morphological data in MacClade's editor.

Many display features and tools in the editor were designed specifically for graphical manipulation and alignment of molecular sequences.

DNA sequence data in MacClade's editor.

Windows listing characters, taxa, trees, taxon sets, character sets, and so on provide a uniform interface for editing their properties, sorting, and selecting them.

If we had to admit a grand purpose to the features of MacClade, it might be "to help biologists explore the relationships between data and hypotheses in phylogenetic biology". We envisage MacClade's use by biologists of many backgrounds. For example, suppose:

1. A systematist is working on the phylogenetic relationships of snail species. She enters data for 100 morphological characters and 20 taxa in MacClade's data editor, saves the file, and reads it into Swofford's (2000) PAUP*. PAUP* is used to find parsimonious trees, but to her surprise, the resulting trees separate two species she had thought closely related. By moving back to MacClade, examining alternative trees by hand, and using the charting functions, she discovers that the unexpected result is due to the influence of two characters of the nervous system. This not only provokes her to examine those characters in more detail, but also to gather molecular data. The molecular data set, on 60 taxa and 1500 nucleotides, is too large to get exact solutions from PAUP*. PAUP*'s heuristic algorithms, combined with her ability to suggest and examine alternative trees in MacClade, convince her that the optimal trees for both data sets agree that the two species are not sister species. Although she had no intention of looking at fossil species, she realizes that this result suggests that their peculiar shared morphology might be primitive and old. By using the stratigraphic character type in MacClade, she discovers the minimum age of the ancestor in which the morphology was apparently derived was Paleocene, much older than the group of predators against which the morphology was thought to be a defense.
   
2. A population biologist has been studying the adaptive advantages of larval dispersal strategies. Although his past studies have focused on measuring risks and fecundities at two study sites in the Gulf of California, he realizes that his hypotheses could be tested by seeing if it accurately predicts a species' strategy according to its ecological position. After talking to a phylogenetic biologist, he reluctantly admits that his question is actually one of phylogenetic correlation between ecology and strategy. He finds a collaborator working on the phylogeny of the group, and together they use MacClade to map the evolution of larval dispersal strategies on the phylogenetic tree. He discovers that the correlation between strategy and ecology is not nearly as strong as he would have hoped, but he notices on MacClade's character tracing that the two groups on the phylogenetic tree with a special dispersal strategy also have the most species. His study shifts to an examination of the influence of dispersal strategy on speciation and extinction. After accumulating known phylogenies from numerous groups and exploring them with MacClade, a clear correlation emerges between speciation success and the special dispersal strategy. Itching to get back to his study sites, he discovers that his newly acquired familiarity with trees can be applied to his populations to examine patterns of dispersal and gene flow using reconstructed gene trees. He uses MacClade's random tree and random character generation to formulate null models to test his population hypotheses.
   
3. A molecular biologist studying cell surface proteins in a herbivorous insect suspects that a particular domain might be involved in binding secondary compounds made by the host plants. A direct chemical approach to demonstrate binding proves difficult, but the genes coding for this protein have been sequenced in several dozen related species. Using MacClade to help reconstruct the phylogeny of the species and to chart how evolutionary changes of amino acids are distributed along the length of the protein, she finds that this domain is indeed the most rapidly evolving. Furthermore, by comparing the phylogenies of the insect and the plant host using Page's COMPONENT (1993), she makes a convincing case that one of the insect groups arose with a shift to a new plant host that had powerful and diverse secondary compounds. Not only did MacClade reconstruct many amino acid changes along that branch of the insect phylogenetic tree, but it also showed that changes in other amino acids were concentrated in the clade of insects living on noxious hosts. With such evidence that the domain was evolving in response to host secondary compounds, the molecular biologist faced with renewed vigor the direct chemical approach to examining the binding.
   
4. An evolutionary biologist teaching a basic evolution course wants to introduce students to phylogenetic reconstruction. In a live demonstration MacClade is used first to help students picture phylogenetic trees in their minds' eyes. Then, the concept of fit between a tree and a character is shown via MacClade's character tracing. Students are given an example data set and let loose. Soon they are moving branches around with increasing rapidity, treating the search for parsimonious trees like a video game, competing against one another for shorter trees. In the process they discover that grouping by derived similarity will gain them shorter trees. One of them notices that one character disagrees with most of the remaining characters on whatever tree they can find. Indeed, on the parsimonious trees this character shows many cases of convergence. Another character is found that is similarly convergent, and students discuss what evolutionary processes might lead to the two characters having correlated evolutionary changes.

These scenarios illustrate the varied uses of a phylogenetic perspective, and of MacClade in making it accessible.