Program Description and Implementation
FastMorphologyGFC is a program written in Python to transform an initial rough data matrix of Specimens X Characters into a ready to use Nexus Paup-file in fractions of a second. Any type of character can be incorporated, including multistate polymorphic, although continuous characters (measurements) should be transformed before hand (see below). The program uses the Generalized Frequency Coding Technique (GFC; Smith Gutberlet, 2001) to transform character states into a parsimony-friendly coding. The result of this coding is that all characters have the same weight (different than 1), independent of how many states they have. Regular binary characters (0/1) that are monomorphic at each OTU will behave in the same way as in a traditional application. Binary characters with the two states present (together) on at least one OTU will be handled as in the frequency bins method (sensu Wiens, 1995). Polymorphic multistate characters will be handled using the GFC coding technique, dividing characters into subcharacters and assigning specific codes and weights to each. Both unordered and ordered characters can be analyzed (see Smith & Gutberlet, 2001) since FastMorphologyGFC will create sets for cumulative and none cumulativefrequencies, respectively for each type. Both types of corresponding weight-sets are created in the Nexus file, under the USW and ESW weighting schemes (see Smith & Gutberlet, 2001). FastMorphologyGFC will assign a weight of 32767 to each character, so that if you want to interpret number of steps in the typical way you should divide your resulting tree-length by this number. We recommend citing the program as:
Chang, V., and E. N. Smith. 2001. FastMorphologyGFC Version
Smith & Gutberlet (2001) discuss the advantage of using GFC over other techniques to include polymorphic multistate characters in parsimony analysis. They also discuss philosophical implications and practical comparisons between using USW (unequal subcharacter weighting) versus ESW (equal subcharacter weighting) weighting in GFC parsimony analysis. As stated in the paper I favor using USW, because useful synapomorphies can be obtained from rare character states, these should be more informative and weighted more.
Another program, CodeThis! (Gutberlet et al, 2001), was created to obtain GFC codes and weights. Character states have to be already assigned (grouped) to each OTU. CodeThis! runs in Microsoft Excel, and works each character at a time producing codes and weights. The program is useful for visualizing each character alone. For large data sets getting data in the format to be used for CodeThis! can take a long time, as does making the Nexus file of concatenated codes and weights. This formatting time is what encouraged the development of FastMorphologyGFC. For more information about CodeThis! and GFC please visit the website at: http://home.earthlink.net/~mgutberl/codethis/codethis.htm.
Please note that the file is in zip format. Windows XP has a built in extractor or a third party utility such as winzip or winrar can be used.