The folks at the Woods Hole Workshop on Molecular Evolution have posted a new version of the PAUP* GUI for Mac OSX. This link replaces one to an expired version of this GUI that was posted last year. If last year's version is any indication, it seems likely that this new version will expire at some point a few months from now. The new version of the GUI remains only partly functional and very buggy. Nevertheless, you may find it useful for some basic fuctions if you haven't yet learned how to use the fully-functional command line version of PAUP*. The link and the GUI executable itself are unsecured, though a line appears when you run the application saying "This version prepared for exclusive use fo authorized course participants." I'll start feeling guilty about using it as soon as I get the PAUP* manual I bought from Sinauer years ago.
Saccularina sp.
1 week ago
8 comments:
God, I just wish Sinauer would upgrade PAUP to work on OSX.
I don't think Sinauer is the problem. In any case, I'm beyond my mourning for PAUP*. I think most of us have passed through the three stages of PAUP* withdrawal: (1) maintaining an old computer with OS9 solely to run PAUP*, (2) learning command line functions to replace the GUI menus, and, finally, (3) realizing that other programs are capable of building better phylogenetic trees faster than PAUP*.
Oh definitely, PAUP* for tree-searches is definitely dead - we're all Bayesians now! but it's the various swiss army knife functions in PAUP* that I miss, like generating random trees, calculating various stats like the CI/RI, making consensus trees, etc. Most of the functions can be replaced by other programs (TreeStat, TreeAnnotator, FigTree, etc) but there's still a few missing.
Amen, Simon. And am I an idiot, or is PAUP still the easiest way to take a partitioned dataset and spit out separate files for each partition? (The initial clause of this question is rhetorical.) For those of you using Geneious, how does it's PAUP* interface compare with this one?
For those of you who aren't Geneious users, the function John is referring to is a menu driven front-end for PAUP that opens as a plug-in within Geneious. To get it to work you need to have a (purchased) copy of the PAUP* executable on your computer. The Geneious PAUP interface may be a convenient alternative for some users, but won't save more sophisticated users the trouble of learning the command line functions. You can use the menus offered by Geneious to change basic options like method of analysis (P, ML, NJ), tree search strategy (Heuristic or Branch and Bound), and the method for finding the starting tree (stepwise or neighborjoining). You cannot, however, use this menu to change some other important parameters like the number of sequence addition replicates or the type of swapping being conducted. You can also run custom PAUP blocks that contain all the options your heart desires through the Geneious front-end, but you might as well just use PAUP natively if you're going to go to the trouble of creating a PAUP block. Nevertheless, one feature that may make the Geneious PAUP interface worth using is that it has bundled Modeltest as well, making it possible to select a model and run analyses in a single step (note - I have yet to use this function in Geneious).
I haven't used PAUP* in years and I agree with Glor, there are much better options. Perhaps the most interesting thing about PAUP* now though is what it tells us about developing important scientific software. There was a time when PAUP was the best computer program in biology. It must have been an enormous effort to rewrite, patch, fix, let alone develop this software. I don't know details but from my knowledge of government funding agencies I seriously doubt Dave Swofford had regular funding to hire programmers and postdocs specifically to assist with PAUP. The commercial model for PAUP* really doesn't look like it worked either (although I paid for it). What would have happened if PAUP* had been made open source in the mid 1990s? I think lots of different people would have contributed to its development and ultimately ensured its persistence. Shame. Yet if we look at the best phylogenetic software available today are any of them open source? Hardly any. I really hope they have better luck.
Umm...Hardly any of the best phylogenetic programs available today are open source?
Was that a typo?
Not sure about most of you, but virtually all of the phylogenetic apps that I use are open source.
Must've been a typo...
"virtually all of the phylogenetic apps that I use are open source"
One problem Brian is that one user's favourite programs may be Open Source (OS) but what about the software in general? It could be that the list below is also biased, I chose them by just remembering programs I've used or bookmarked over the last few years. There are many others but it doesn't look like "virtually all" to me. OS does not mean free, nor that you can download and compile the source for yourself. I don't count licenses that prevent you from modifying the code for yourself. There are lots of types of OS licence but GNU/GPL is common. "copyright" means not OS.
MEGA, copyright
ARB, copyright
*Mesquite, GNU open source
RAxML, hard to tell license from website or recent publications (but RAXMLIII paper 2005 did say it was "Open Source")
GARLI, creative commons license, no modification allowed
PhyML, copyright
Treefinder, copyright
*MrBayes GNU/GPL open source
TNT, copyright
bioNJ, copyright
*RapidNJ, GNU open source
QuickTree, license unclear in paper and web, source code available
*HYPHY, GPL license open source
PAML, copyright
SplitsTree4, copyright
MultiPhyl, ??? I can't find exactly, but it doesn't looks open source
Geneious, copyright
Post a Comment