Introduction to Some Computer Programs Used in Phylogenetics
Michael D. Crisp
Division of Botany and Zoology
Australian National University
Canberra ACT 2601
A number of programs are available for phylogenetic analysis. Brief descriptions are given below.
A World Wide Web site by David and Wayne Maddison (http://phylogeny.arizona.edu/tree/home.pages/concepts.html) gives a concise introduction to phylogenetic reconstruction. More importantly, this page has a hypertext link to a page of information on 20 different phylogeny computer programs (http://phylogeny.arizona.edu/tree/programs/programs.html), with a brief description of each, and links to pages from which these programs may be downloaded, or where more information about them can be obtained.
Alternatively, Joe Felsenstein maintains a list of 100 phylogeny packages (http://evolution.genetics.washington.edu/phylip/software.html), with a brief description of each, and links to pages from which these programs may be downloaded, or where more information about them can be obtained.
Other lists of phylogenetics software include those maintained by:-
- International Willi Hennig Society
(http://www.vims.edu/~mes/hennig/software.html) - Taxonomy and Systematics at the University of Glasgow
(http://taxonomy.zoology.gla.ac.uk/software.html) - University of California Museum of Paleontology
(http://www.ucmp.berkeley.edu/subway/phylo/phylosoft.html)
PAUP
The most heavily-cited program is David Swofford's PAUP version 3.0 for the Apple Macintosh range of computers (the latest version is 3.1.1). At the time of writing, PAUP is unavailable, but Swofford is developing a new, enhanced version (4.0), which will be distributed commercially by Sinauer Associates (http://www.sinauer.com/Titles/frswofford.htm). This package will contain new algorithms for maximum likelihood analysis and neighbor joining, and statistical tests such as bootstrapping and permutation tail probability tests. Most of the add-ons are specifically for analysing sequence data. The standard version of PAUP, with a 'point and click' mode of operation, has been developed only for Macintosh and PowerMac (native mode). However, Swofford has also developed a 'portable' version which will run on virtually any system, e.g. Unix and DOS. It runs only from a command prompt, but has all the features of its Mac counterpart. However, in my tests of the pre-release version, it ran much slower than either Hennig86 on a PC or PAUP on a Mac. No Windows version is being developed. The Smithsonian Laboratory of Molecular Systematics gopher site (gopher://nmnhgoph.si.edu:70/11/.lms/.lmsftp) provides information about PAUP and related phylogenetic inference packages (Freqpars and MacClade).
MacClade
MacClade is a program for Macintosh by Wayne P. Maddison and David R. Maddison. It is designed for analyzing character evolution on trees which have been derived by other means (e.g. PAUP), and includes various statistics. Using MacClade is a great way of learning 'tree-thinking', and for trying 'what if' phylogenetic scenarios and looking at their consequences in terms of character evolution. It has a graphical spreadsheet editor for entering and editing data matrices, a tree window for viewing and manipulating trees, charting facilities, and so on. Thus it is also an ideal program for prettying-up your trees for publication. A more complete description is given in MacClade's WWW site (http://phylogeny.arizona.edu/macclade/macclade.html). The current version is 3.06, published by Sinauer Associates (http://www.sinauer.com/system.shtml#Maddison). It comes with a book (with a general introduction to phylogenetic reconstruction and character evolution), so the price varies according to the options chosen.
Phylip
Joseph Felsenstein's Phylip package, currently version 3.5, includes 31 programs to do parsimony (e.g. DNAPARS, DNAPENNY, PROTPARS), transformed evolutionary distances (DNADIST), additive trees (NEIGHBOR, FITCH), and maximum likelihood (DNAML) analyses for a variety of data types, including molecular sequences (DNA and protein), gene frequencies, continuous characters, discrete (0/1) characters, and restriction sites. It also includes programs to calculate statistics, plot phylogenies, and construct consensus trees. These can be run from the Analysis/Phylip submenu of ANGIS, but are also obtainable, free of charge, as C source code (which can be compiled on almost anything else) and as executable files for MS-DOS, Windows and Macintosh computers, from either the anonymous ftp site (128.95.12.41) or the WWW site (http://evolution.genetics.washington.edu/phylip.html). Phylip is greedy for memory: it needs at least 3Mb free. It is the most widely-distributed phylogeny package, presumably at least partly to do with the low cost! But, also, it offers great flexibility and excellent documentation. The Phylip documentation includes a comprehensive guide to other software packages.
Hennig86
Hennig86 was written by James S. ('Steve') Farris for MS-DOS computers. It is still distributed in its original version (1.5) dating from 1988. Nevertheless, it is a very good parsimony program, the best available for PCs. To obtain it contact one of Farris' distributors, such as Ian Kitching (ijk@nhm.ac.uk), Arnold Kluge (akluge@umich.ed), or Robert Raven at the Queensland Museum. It currently costs US$55. It is a fast but unfriendly and relatively inflexible program. The command names are counter-intuitive, e.g. 'yama' for quit (God of Destruction), 'hennig' to search for a tree, etc. Also, everything is numbered from zero. However, it is worth using because it is very efficient on big heuristic searches, and often finds more, shorter trees more quickly than PAUP from the same data set. Although the commands are cryptic, when you have learned them the program is relatively easy to use. The original manual is brief and obscure, so an 'alternative' simplified manual is available (http://www.vims.edu/~mes/mes/henhelp.html).
RnA
J.S. Farris' RnA (also known as 'RA') is a heuristic parsimony program for MS-DOS computers. It costs $US40, from one of Farris' distributors (see above). RnA is quite different from Hennig86: it is user-unfriendly, it allows very little (actually no) choice, and it is very very fast. The data-format is unconventional (taxon names in brackets, which means there is no need to specify the number of characters and taxa). It is not a powerful tree-searcher, which means that it often calculates the statistics on wrong trees (it is essentially a statistics calculator). This would have been okay if you could input the tree topology into it, but you can't. So it is useful if the tree-topology is obvious, otherwise not.
SeqApp
SeqApp is a biological sequence editor and analysis program by Don Gilbert (SeqApp@bio.indiana.edu) for Macintosh computers. Features include: multiple sequence alignment editor; single sequence editor window; read and write several sequence file formats (e.g. Genbank, PAUP, and many others); easy hand-alignment features, including coloured bases and sliding; automatic multiple sequence alignment through Clustal (external application bundled with SeqApp); automatic gel fragment alignment to contigs through CAP (external application); consensus, reverse, complement, degap operations; restriction maps; dot plots; translate DNA to/from protein using various codon tables; internet send mail, read mail; internet gopher information retrieval, including Genbank; internet sequence analysis services; user-definable links to external analysis programs; and more.
SeqApp is freely available, with instructions for obtaining it being available at the WWW site (http://iubio.bio.indiana.edu/IUBio-Software+Data/molbio/seqapp/). Don Gilbert has ceased developing SeqApp (the latest release is 1.9a), and instead he is working on a new program, SeqPup, which will have more features. You will find instructions for obtaining a copy of the latter at the same site (http://ftp.bio.indiana.edu/IUBio-Software+Data/molbio/seqpup/java/seqpup-doc.html).
