PHYLIP

Phylogeny Inference Package

PHYLIP Logo

Version 3.66

July, 2006

by Joseph Felsenstein

Department of Genome Sciences and Department of Biology
University of Washington
Box 357730
Seattle, WA 98195-7730
USA

E-mail address: `joe (at) gs.washington.edu`

Contents of This Document
A Brief Description of the Programs
Copyright Notice for PHYLIP
The Documentation Files and How to Read Them
What The Programs Do
Running the Programs
      A word about input files
      Running the programs on a Windows machine
      Running the programs on a Macintosh with Mac OS X
      Running the programs on a Macintosh with Mac OS 8 or 9
      Running the programs on a Unix or Linux system
      Running the programs in MSDOS
      Running the programs in background or under control of a command file
Preparing Input Files
      Input and output files
      Data file format
The Menu
The Output File
The Tree File
The Options and How To Invoke Them
      Common options in the menu
        The U (User tree) option
        The G (Global) option
        The J (Jumble) option
        The O (Outgroup) option
        The T (Threshold) option
        The M (Multiple data sets) option
        The W (Weights) option
        The option to write out the trees into a tree file
        The (0) terminal type option
The Algorithm for Constructing Trees
      Local rearrangements
      Global rearrangements
      Multiple jumbles
      Saving multiple tied trees
      Strategy for finding the best tree
A Warning on Interpreting Results
Relative Speed of Different Programs and Machines
      Relative speed of the different programs
      Speed with different numbers of species
      Relative speed of different machines
General Comments on Adapting the Package to Different Computer Systems
Compiling the programs
      Unix and Linux
      On Windows systems
           Compiling with Cygnus Gnu C++
           Compiling with Microsoft Visual C++
           Compiling with Borland C++
           Compiling with Metrowerks Codewarrior for Windows
      Macintosh
           Compiling with GCC on Mac OS X with our Makefile
           Compiling with GCC on Mac OS X with X Windows
           Compiling with Metrowerks Codewarrior
           Creating the Metrowerks project file.
      VMS VAX systems
      Parallel computers
      Other computer systems
Frequently Asked Questions
      How to make it do various things
      Background information needed:
      Questions about distribution and citation:
      Questions about documentation
      Additional Frequently Asked Questions, or: "Why didn't it occur to you to ...
      (Fortunately) obsolete questions
New Features in This Version
Coming Attractions, Future Plans
Endorsements
      From the pages of Cladistics
      ... in the pages of other journals:
      ... and in the comments made by users when they register:
References for the Documentation Files
Credits
Other Phylogeny Programs Available Elsewhere
      PAUP*
      MacClade
      MrBayes
      MEGA
      PAML
      TREE-PUZZLE
      DAMBE
      Hennig86
      RnA
      NONA
      TNT
How You Can Help Me
In Case of Trouble

PHYLIP, the Phylogeny Inference Package, is a package of programs for inferring phylogenies (evolutionary trees). It has been distributed since 1980, and has over 20,000 registered users, making it the most widely distributed package of phylogeny programs. It is available free, from its web site:

PHYLIP is available as source code in C, and also as executables for some common computer systems. It can infer phylogenies by parsimony, compatibility, distance matrix methods, and likelihood. It can also compute consensus trees, compute distances between trees, draw trees, resample data sets by bootstrapping or jackknifing, edit trees, and compute distance matrices. It can handle data that are nucleotide sequences, protein sequences, gene frequencies, restriction sites, restriction fragments, distances, discrete characters, and continuous characters.

Copyright Notice for PHYLIP

The following copyright notice is intended to cover all source code, all documentation, and all executable programs of the PHYLIP package.

© Copyright 1980-2006. University of Washington. All rights reserved. Permission is granted to reproduce, perform, and modify these programs and documentation files. Permission is granted to distribute or provide access to these programs provided that this copyright notice is not removed, the programs are not integrated with or called by any product or service that generates revenue, and that your distribution of these documentation files and programs are free. Any modified versions of these materials that are distributed or accessible shall indicate that they are based on these program. Institutions of higher education are granted permission to distribute this material to their students and staff for a fee to recover distribution costs. Permission requests for any other distribution of this program should be directed to license (at) u.washington.edu .

PHYLIP comes with an extensive set of documentation files. These include the main documentation file (this one), which you should read fairly completely. In addition there are files for groups of programs, including ones for the molecular sequence programs, the distance matrix programs, the gene frequency and continuous characters programs, the discrete characters programs, and the tree drawing programs. Finally, each program has its own documentation file. References for the documentation files are all gathered together in this main documentation file. A good strategy is to:

There is an excellent guide to using PHYLIP 3.6 also available. It was written by Jarno Tuimala of the Center for Scientific Computing in Espoo, Finland and is available as a PDF here.

Here is a short description of each of the programs. For more detailed discussion you should definitely read the documentation file for the individual program and the documentation file for the group of programs it is in. In this list the name of each program is a link which will take you to the documentation file for that program. Note that there is no program in the PHYLIP package called PHYLIP.

This section assumes that you have obtained PHYLIP as compiled executables (for Windows, Mac OS X, Mac OS 9 or Linux), or have obtained the source code and compiled it yourself (for Linux, Unix, or OpenVMS). For machines for which compiled executables are available, there will usually be no need for you to have a compiler or compile the programs yourself. This section describes how to run the programs. Later in this document we will discuss how to download and install PHYLIP (in case you are reading this without yet having done that). Normally you will only read your copy of this document after downloading and installing PHYLIP.

A word about input files.

For all of these types of machines, it is important to have the input files for the programs (typically data files) prepared in advance. They can be prepared in any editor, but it is important that they be saved in Text Only ("flat ASCII") format, not in the format that word processors such as Microsoft Word want to write. It is up to you to read the PHYLIP documentation files which describe the files formats that are needed. There is a partial description in the next section of this document. The input files can also be obtained by running a program that produces output files in PHYLIP format (some of these programs do, and so do programs by others such as sequence alignment programs such as ClustalW and sequence format conversion programs such as Readseq). There is not any input file editor available in any program in PHYLIP (you should not simply start running one of the programs and then expect to click a mouse somewhere to start creating a data file).

When they start running, the programs look first for input files with particular names (such as infile, treefile, intree, or fontfile). Exactly which file names they look for varies a bit from program to program, and you should read the documentation file for the particular program to find out. If you have files with those names the programs will use them and not ask you for the file name. If they do not find files of those names, the programs will say that they cannot find a file of that name, and ask you to type in the file name. For example, if DnaML looks for the file infile and does not find one of that name, it prints the message:

This does not mean that an error has occurred. All you need to do is to type in the name of the file.

The program looks for the input files in the same folder that the program is in (a folder is the same thing as a "directory"). In Windows, Mac OS X, Linux, or Unix, if you are asked for the file name you can type in the path to the file, as part of the name (thus, if the file is in the folder containing the current folder, you can type in a file name such as ../myfile.dna). If you do not know what a "folder" is, or what "above" means, then you are a member of the new generation who just clicks the mouse and assumes that a list of file names will magically appear. (Typically members of this generation have no idea where the files are on their system, and accumulate enormous amounts of unnecessary clutter in their file systems.) In this case you should ask someone to explain folders to you.

Running the programs on a Windows machine.

Double-click on the icon for the program. A window should open with a menu in it. Further dialog with the program occurs by typing on the keyboard in response to what you see in the window. The programs can be interrupted either by typing Control-C (which means to press down on the Ctrl key while typing the letter C), or by using the mouse to open the File menu in the upper-left corner of the program's window area and then select Quit. Other than this, most PHYLIP programs make no use of the mouse. The tree-drawing programs Drawtree and Drawgram do allow use of the mouse to select some options.

The programs open a window for their menus. This window may be too small for your tastes. They can be resized by tugging on the lower-right corner of the window. In addition, the font may be too small. On most versions of Windows, you can click on the small icon symbol at the upper-left corner of the window, and choose the Properties menu choice there. One of its options is to change the font and size of the print. I prefer large font sizes such as 16x12. Some versions of Windows wuch as Windows XP also ask you if you want to apply this choice to all such windows, not just to the current one.

The programs can also be run in a Command window under Windows, in much the same way as they were under the MSDOS operating system, which is what the Command window emulates. Type the name of the program in lower-case letters (such as dnaml). To interrupt the program while it is running, type Control-C (which means to press down on the Ctrl key while typing the letter C).

Running the programs on a Macintosh with Mac OS X

We have provided a Mac OS X version of the executables. The programs can be run by clicking on their icons. It is also possible to run them using a Terminal window by typing the correct name, but this is harder. The executables open a Terminal window in which the menus appear and in which responses can be typed. The programs can be interrupted by typing a control-C (press down the "control" key in the lower-left corner of the keyboard and then type "c"), or by closing the Terminal window by using the red button in the upper-left corner of the window.

On Mac OS X systems the Mac OS 8 or 9 executables can also be run using the Classic environment included with it which allows executables for earlier versions of Mac OS to run.

If your Mac OS X system has the X windows windowing system installed, you could also make excutables that will work with that. See below, under "Compiling with GCC on Mac OS X with X Windows" for instructions on how to do this.

One problem we have often encountered using Mac OS X is that it is possible for data files to have the wrong kind of characters at the ends of their lines. They may have carriage-return (ASCII/ISO 13 or control-M) characters at the ends of their lines when they should instead have the Unix newline character (ASCII/ISO 10 or control-J) there. This can happen with files transferred from other operating systems or files produced in some word processors. It results in segmentation-fault or memory errors. If you encounter these, check this possibility carefully.

Running the programs on a Macintosh with Mac OS 8 or 9

Double-click on the icon for the program. A window should open. Further dialog with the program occurs by typing on the keyboard in response to what you see in the window. The programs can be interrupted by using the mouse to open the File menu in the upper-left corner of the program's window area and then select Quit. Alternatively, you can use the Command-Q key combination.

When you use Quit, the program will ask you whether you want to save a file whose name is the program name (often followed by .out -- for example, if you are using Dnaml it will ask you if you want to save file Dnaml.out. This file is simply a record of everything that displayed on the program window, and you usually will not want to save it. Pressing the Enter key or selecting the Do Not Save button with the mouse will keep this from being saved.

If you encounter memory limitations on a Mac OS 8 or 9 Macintosh, and determine that this is not due to a problem with the format of the input file, as it often will be, you may be able to solve it by raising the limits of the stack and heap sizes of the program. To do this click on the program and then select Get Info from the Finder File menu. This will open a window which can be made to show the memory limits of the program. These can be changed by selecting them and typing in larger numbers. This may relieve nagging memory problems. If it does not, consult your local documentation and suspect problems with your input file format.

Running the programs on a Unix or Linux system.

Type the name of the program in lower-case letters (such as dnaml). To interrupt the program while it is running, type Control-C (which means to press down on the Ctrl key while typing the letter C).

On some systems you may need to type ./ before the program name, so that in the above case it would be ./dnaml. This is mostly needed if the users PATH does not include their current directory, something which is often done as a security precaution.

Running the programs in background or under control of a command file

In running the programs, you may sometimes want to put them in background so you can proceed with other work. On systems with a windowing environment they can be put in their own window, and commands like the Unix and Linux nice command used to make them have lower priority so that they do not interfere with interactive applications in other windows. This part of the discussion will assume either a Windows system or a Unix or Linux system. I will note when the commands work on one of these systems but not the other. Mac OS X is actually Unix (surprise! surprise!) and you can run PHYLIP programs in background on any Mac OS X system by simply following the instructions for Unix, using a terminal window to do so if necessary. Running jobs in background on Mac OS 8 or 9 systems is an arcane art into whose mysteries I have not been initiated (or perhaps no one has been initiated).

If there is no windowing environment, on a Unix or Linux system you will want to use an ampersand (&) after the command file name when invoking it to put the job in the background. You will have to put all the responses to the interactive menu of the program into a file and tell the background job to take its input from that file.

On Windows systems there is no & or nice command but input and output redirection and command files work fine in a Commmand window. A command file can either be invoked by clicking on its icon or by typing its name from a Command window. The a file of commands must have a name ending in .bat, such as foofile.bat. You can run the batch file from a Command window by typing its name (such as foofile) without the .bat.

For example: suppose you want to run Dnapars in a background, taking its input data from a file called sequences.dat, putting its interactive output to file called screenout, and using a file called input as the place to store the interactive input. The file input need only contain two lines:

which is what you would have typed to run the program interactively, in response to the program's request for an input file name if it did not find a file named infile, in in response the the menu.

To run the program in background, in Unix or Linux you would simply give the command:

These run the program with input responses coming from input and interactive output being put into file screenout. The usual output file and tree file will also be created by this run (keep that in mind as if you run any other PHYLIP program from the same directory while this one is running in background you may overwrite the output file from one program with that from the other!).

If you wanted to give the program lower priority, so that it would not interfere with other work, and you have Berkeley Unix type job control facilities in your Unix or Linux (and you usually do), you can use the nice command:

which lowers the priority of the run. To also time the run and put the timing at the end of screenout, you can do this:

On Unix or Linux systems you may also want to explore putting the interactive output into the null file /dev/null so as to not be bothered with it (but then you cannot look at it to see why something went wrong). If you have problems with creating output files that are too large, you may want to explore carefully the turning off of options in the programs you run.

If you are doing several runs in one, as for example when you do a bootstrap analysis using Seqboot, Dnapars (say), and Consense, you can use an editor to create a "command file" with these commands:

seqboot < input1 > screenout
mv outfile infile
dnapars < input2 >> screenout
mv outtree intree
consense < input3 >> screenout

This is the Unix or Linux version -- in the Windows version, the renaming of files and the appending of output to the file screenout is handled differently.

On Unix or Linux the command file might be named something like foofile, and on Windows systems might be named foofile.bat.

On Unix or Linux the command file must be given execute permission by using the command chmod +x foofile followed by the command rehash. The job that foofile describes can be run in background on Unix or Linux by giving the command

On Windows systems it can be run by clicking on the icon of the command file. Its icon will have a little gear symbol.

Note that you must also have the interactive input commands for Seqboot (including the random number seed), Dnapars, and Consense in the separate files input1, input2, and input3. Note also that when PHYLIP programs attempt to open a new output file (such as outfile, outtree, or plotfile, if they see a file of that name already in existence they will ask you if you want to overwrite it, and offer alternatives including writing to another file, appending information to that file, or quitting the program without writing to the file. This means that in writing batch files it is important to know whether there will be a prompt of this sort. You must know in advance whether the file will exist. You may want to put in your batch file a command that tests for the existence of a pre-existing output file and if so, removes it. You might even want to put in a command that creates a file of that name, so that you can be sure it is there! Either way, you will then know whether to put into your file of keyboard responses the proper response to the inquiry about overwriting that output file.

The input files for PHYLIP programs must be prepared separately - there is no data editor within PHYLIP. You can use a word processor (or text editor) to prepare them yourself, or you can use a program that produces a PHYLIP-format output. Sequence alignment programs such as ClustalW commonly have an option to produce PHYLIP files as output, and some other phylogeny programs, such as MacClade and TreeView, are capable of producing a PHYLIP-format file.

The format of the input files is discussed below, and you should also read the other PHYLIP documentation relevant to the particular type of data that you are using, and the particular programs you want to run, as there will be more details there.

It is very important that the input files be in "Text Only" or "flat ASCII" format. This means that they contain only printable ASCII/ISO characters, and not any unprintable characters. Many word processors such as Microsoft Word save their files in a format that contains unprintable characters, unless you tell them not to. For Microsoft Word you can select Save As from its File menu, and choose Text Only as the file format. This can also be done in WordPad utility in Windows . Other word processors will have equivalent options. Text editors such as the vi and emacs editors on Unix and Linux, Windows Notepad, the SimpleText editor in Mac OS, or the pico editor that comes with the pine mailer program, produce their files in Text Only format and should not cause any trouble.

Input and output files

For most of the PHYLIP programs, information comes from a series of input files, and ends up in a series of output files:

                   -------------------
                  |                   |
infile ---------> |                   |
                  |                   |
intree ---------> |                   | -----------> outfile
                  |                   |
weights --------> |      program      | -----------> outtree
                  |                   |
categories -----> |                   | -----------> plotfile
                  |                   |
fontfile -------> |                   |
                  |                   |
                   -------------------

The programs interact with the user by presenting a menu. Aside from the user's choices from the menu, they read all other input from files. These files have default names. The program will try to find a file of that name - if it does not, it will ask the user to supply the name of that file. Input data such as DNA sequences comes from a file whose default name is infile. If the user supplies a tree, this is in a file whose default name is intree. Values of weights for the characters are in weights, and the tree plotting program need some digitized fonts which are supplied in fontfile (all these are default names).

For example, if DnaML looks for the file infile and does not find one of that name, it prints the message:

Data file format

I have tried to adhere to a rather stereotyped input and output format. For the parsimony, compatibility and maximum likelihood programs, excluding the distance matrix methods, the simplest version of the input data file looks something like this:

   6   13
Archaeopt CGATGCTTAC CGC
HesperorniCGTTACTCGT TGT
BaluchitheTAATGTTAAT TGT
B. virginiTAATGTTCGT TGT
BrontosaurCAAAACCCAT CAT
B.subtilisGGCAGCCAAT CAC

The first line of the input file contains the number of species and the number of characters (in this case sites). These are in free format, separated by blanks. The information for each species follows, starting with a ten-character species name (which can include blanks and some punctuation marks), and continuing with the characters for that species. The name should be on the same line as the first character of the data for that species. (I will use the term "species" for the tips of the trees, recognizing that in some cases these will actually be populations or individual gene sequences).

The name should be ten characters in length, filled out to the full ten characters by blanks if shorter. Any printable ASCII/ISO character is allowed in the name, except for parentheses ("(" and ")"), square brackets ("[" and "]"), colon (":"), semicolon (";") and comma (","). If you forget to extend the names to ten characters in length by blanks, the program will get out of synchronization with the contents of the data file, and an error message will result.

Note that Tab characters count as only one character in the species names. Their inclusion can cause trouble. The name will appear to you to be filled out to a full ten characters, but it may be shorter than that to the program. If you make the data file in a word processor such as Word, you would be well-advised to make sure the names contain no Tab characters. You can check for their presence by advancing through the names with your cursor keys, and looking for sudden jumps of two or more characters. It is best to fill the names out with blanks, not Tabs.

In the discrete-character programs, DNA sequence programs and protein sequence programs the characters are each a single letter or digit, sometimes separated by blanks. In the continuous-characters programs they are real numbers with decimal points, separated by blanks:

The conventions about continuing the data beyond one line per species are different between the molecular sequence programs and the others. The molecular sequence programs can take the data in "aligned" or "interleaved" format, in which we first have some lines giving the first part of each of the sequences, then some lines giving the next part of each, and so on. Thus the sequences might look like this:

    6   39
Archaeopt CGATGCTTAC CGCCGATGCT
HesperorniCGTTACTCGT TGTCGTTACT
BaluchitheTAATGTTAAT TGTTAATGTT
B. virginiTAATGTTCGT TGTTAATGTT
BrontosaurCAAAACCCAT CATCAAAACC
B.subtilisGGCAGCCAAT CACGGCAGCC

TACCGCCGAT GCTTACCGC
CGTTGTCGTT ACTCGTTGT
AATTGTTAAT GTTAATTGT
CGTTGTTAAT GTTCGTTGT
CATCATCAAA ACCCATCAT
AATCACGGCA GCCAATCAC

Note that in these sequences we have a blank every ten sites to make them easier to read: any such blanks are allowed. The blank line which separates the two groups of lines (the ones containing sites 1-20 and ones containing sites 21-39) may or may not be present, but if it is, it should be a line of zero length and not contain any extra blank characters (this is because of a limitation of the current versions of the programs). It is important that the number of sites in each group be the same for all species (i.e., it will not be possible to run the programs successfully if the first species line contains 20 bases, but the first line for the second species contains 21 bases).

Alternatively, an option can be selected in the menu to take the data in "sequential" format, with all of the data for the first species, then all of the characters for the next species, and so on. This is also the way that the discrete characters programs and the gene frequencies and quantitative characters programs want to read the data. They do not allow the interleaved format.

In the sequential format, the character data can run on to a new line at any time (except in the middle of a species name or, in the case of continuous character and distance matrix programs where you cannot go to a new line in the middle of a real number). Thus it is legal to have:

though note that the full ten characters of the species name must then be present: in the above case there must be a blank after the "t". In all cases it is possible to put internal blanks between any of the character values, so that

Note that you can convert molecular sequence data between the interleaved and the sequential data formats by using the Rewrite option of the J menu item in Seqboot.

If you make an error in the format of the input file, the programs can sometimes detect that they have been fed an illegal character or illegal numerical value and issue an error message such as BAD CHARACTER STATE:, often printing out the bad value, and sometimes the number of the species and character in which it occurred. The program will then stop shortly after. One of the things which can lead to a bad value is the omission of something earlier in the file, or the insertion of something superfluous, which cause the reading of the file to get out of synchronization. The program then starts reading things it didn't expect, and concludes that they are in error. So if you see this error message, you may also want to look for the earlier problem that may have led to the program becoming confused about what it is reading.

Some options are described below, but you should also read the documentation for the groups of the programs and for the individual programs.

The Menu

The menu is straightforward. It typically looks like this (this one is for Dnapars):

DNA parsimony algorithm, version 3.6

Setting for this run:
  U                 Search for best tree?  Yes
  S                        Search option?  More thorough search
  V              Number of trees to save?  10000
  J   Randomize input order of sequences?  No. Use input order
  O                        Outgroup root?  No, use as outgroup species  1
  T              Use Threshold parsimony?  No, use ordinary parsimony
  N           Use Transversion parsimony?  No, count all steps
  W                       Sites weighted?  No
  M           Analyze multiple data sets?  No
  I          Input sequences interleaved?  Yes
  0   Terminal type (IBM PC, ANSI, none)?  ANSI
  1    Print out the data at start of run  No
  2  Print indications of progress of run  Yes
  3                        Print out tree  Yes
  4          Print out steps in each site  No
  5  Print sequences at all nodes of tree  No
  6       Write out trees onto tree file?  Yes

  Y to accept these or type the letter for one to change

If you want to accept the default settings (they are shown in the above case) you can simply type Y followed by pressing on the Enter key. If you want to change any of the options, you should type the letter shown to the left of its entry in the menu. For example, to set a threshold type T. Lower-case letters will also work. For many of the options the program will ask for supplementary information, such as the value of the threshold.

Note the Terminal type entry, which you will find on all menus. It allows you to specify which type of terminal your screen is. The options are an IBM PC screen, an ANSI standard terminal, or none. Choosing zero (0) toggles among these three options in cyclical order, changing each time the 0 option is chosen. If one of them is right for your terminal the screen will be cleared before the menu is displayed. If none works, the none option should probably be chosen. The programs should start with a terminal option appropriate for your computer, but if they do not, you can change the terminal type manually. This is particularly important in program Retree where a tree is displayed on the screen - if the terminal type is set to the wrong value, the tree can look very strange.

The other numbered options control which information the program will display on your screen or on the output files. The option to Print indications of progress of run will show information such as the names of the species as they are successively added to the tree, and the progress of rearrangements. You will usually want to see these as reassurance that the program is running and to help you estimate how long it will take. But if you are running the program "in background" as can be done on multitasking and multiuser systems, and do not have the program running in its own window, you may want to turn this option off so that it does not disturb your use of the computer while the program is running. Note also menu option 3, "Print out tree". This can be useful when you are running many data sets, and will be using the resulting trees from the output tree file. It may be helpful to turn off the printing out of the trees in that case, particularly if those files would be too big.

The Output File

Most of the programs write their output onto a file called (usually) outfile, and a representation of the trees found onto a file called outtree.

The exact contents of the output file vary from program to program and also depend on which menu options you have selected. For many programs, if you select all possible output information, the output will consist of (1) the name of the program and its version number, (2) some of the input information printed out, and (3) a series of phylogenies, some with associated information indicating how much change there was in each character or on each part of the tree. A typical rooted tree looks like this:

                                     +-------------------Gibbon
        +----------------------------2
        !                            !      +------------------Orang
        !                            +------4
        !                                   !  +---------Gorilla
  +-----3                                   +--6
  !     !                                      !    +---------Chimp
  !     !                                      +----5
--1     !                                           +-----Human
  !     !
  !     +-----------------------------------------------Mouse
  !
  +------------------------------------------------Bovine

The interpretation of the tree is fairly straightforward: it "grows" from left to right. The numbers at the forks are arbitrary and are used (if present) merely to identify the forks. For many of the programs the tree produced is unrooted. Rooted and unrooted trees are printed in nearly the same form, but the unrooted ones are accompanied by the warning message:

to indicate that this is an unrooted tree and to warn against taking the position of its root too seriously. Mathematicians still call an unrooted tree a tree, though some systematists unfortunately use the term "network" for an unrooted tree. This conflicts with standard mathematical usage, which reserves the name "network" for a completely different kind of graph). The root of this tree could be anywhere, say on the line leading immediately to Mouse. As an exercise, see if you can tell whether the following tree is or is not a different one from the above:

             +-----------------------------------------------Mouse
             !
   +---------4                                   +------------------Orang
   !         !                            +------3
   !         !                            !      !       +---------Chimp
---6         +----------------------------1      !  +----2
   !                                      !      +--5    +-----Human
   !                                      !         !
   !                                      !         +---------Gorilla
   !                                      !
   !                                      +-------------------Gibbon
   !
   +-------------------------------------------Bovine

   remember: this is an unrooted tree!

(it is not different). It is important also to realize that the lengths of the segments of the printed tree may not be significant: some may actually represent branches of zero length, in the sense that there is no evidence that those branches are nonzero in length. Some of the diagrams of trees attempt to print branches approximately proportional to estimated branch lengths, while in others the lengths are purely conventional and are presented just to make the topology visible. You will have to look closely at the documentation that accompanies each program to see what it presents and what is known about the lengths of the branches on the tree. The above tree attempts to represent branch lengths approximately in the diagram. But even in those cases, some of the smaller branches are likely to be artificially lengthened to make the tree topology clearer. Here is what a tree from Dnapars looks like, when no attempt is made to make the lengths of branches in the diagram proportional to estimated branch lengths:

                 +--Human
              +--5
           +--4  +--Chimp
           !  !
        +--3  +-----Gorilla
        !  !
     +--2  +--------Orang
     !  !
  +--1  +-----------Gibbon
  !  !
--6  +--------------Mouse
  !
  +-----------------Bovine

  remember: this is an unrooted tree!

When a tree has branch lengths, it will be accompanied by a table showing for each branch the numbers (or names) of the nodes at each end of the branch, and the length of that branch. For the first tree shown above, the corresponding table is:

 Between        And            Length      Approx. Confidence Limits
 -------        ---            ------      ------- ---------- ------

    1          Bovine            0.90216     (  0.50346,     1.30086) **
    1          Mouse             0.79240     (  0.42191,     1.16297) **
    1             2              0.48553     (  0.16602,     0.80496) **
    2             3              0.12113     (     zero,     0.24676) *
    3             4              0.04895     (     zero,     0.12668)
    4             5              0.07459     (  0.00735,     0.14180) **
    5          Human             0.10563     (  0.04234,     0.16889) **
    5          Chimp             0.17158     (  0.09765,     0.24553) **
    4          Gorilla           0.15266     (  0.07468,     0.23069) **
    3          Orang             0.30368     (  0.18735,     0.41999) **
    2          Gibbon            0.33636     (  0.19264,     0.48009) **

      *  = significantly positive, P < 0.05
      ** = significantly positive, P < 0.01

Ignoring the asterisks and the approximate confidence limits, which will be described in the documentation file for Dnaml, we can see that the table gives a more precise idea of what the lengths of all the branches are. Similar tables exist in distance matrix and likelihood programs, as well as in the parsimony programs Dnapars and Pars.

Some of the parsimony programs in the package can print out a table of the number of steps that different characters (or sites) require on the tree. This table may not be obvious at first. A typical example looks like this:

 steps in each site:
         0   1   2   3   4   5   6   7   8   9
     *-----------------------------------------
    0!       2   2   2   2   1   1   2   2   1
   10!   1   2   3   1   1   1   1   1   1   2
   20!   1   2   2   1   2   2   1   1   1   2
   30!   1   2   1   1   1   2   1   3   1   1
   40!   1

The numbers across the top and down the side indicate which site is being referred to. Thus site 23 is column "3" of row "20" and has 1 step in this case.

There are many other kinds of information that can appear in the output file, They vary from program to program, and we leave their description to the documentation files for the specific programs.

The Tree File

In output from most programs, a representation of the tree is also written into the tree file outtree. The tree is specified by nested pairs of parentheses, enclosing names and separated by commas. We will describe how this works below. If there are any blanks in the names, these must be replaced by the underscore character "_". Trailing blanks in the name may be omitted. The pattern of the parentheses indicates the pattern of the tree by having each pair of parentheses enclose all the members of a monophyletic group. The tree file could look like this:

In this tree the first fork separates the lineage leading to Mouse and Bovine from the lineage leading to the rest. Within the latter group there is a fork separating Gibbon from the rest, and so on. The entire tree is enclosed in an outermost pair of parentheses. The tree ends with a semicolon. In some programs such as Dnaml, Fitch, and Contml, the tree will be unrooted. An unrooted tree should have its bottommost fork have a three-way split, with three groups separated by two commas:

Here the three groups at the bottom node are A, (B,C,D), and (E,F). The single three-way split corresponds to one of the interior nodes of the unrooted tree (it can be any interior node of the tree). The remaining forks are encountered as you move out from that first node. In newer programs, some are able to tolerate these other forks being multifurcations (multi-way splits). You should check the documentation files for the particular programs you are using to see in which of these forms you can expect the user tree to be in. Note that many of the programs that actually estimate an unrooted tree (such as Dnapars) produce trees in the treefile in rooted form! This is done for reasons of arbitrary internal bookkeeping. The placement of the root is arbitrary. We are working toward having all programs be able to read all trees, whether rooted or unrooted, multifurcating or bifurcating, and having them do the right thing with them. But this is a long-term goal and it is not yet achieved.

For programs that infer branch lengths, these are given in the trees in the tree file as real numbers following a colon, and placed immediately after the group descended from that branch. Here is a typical tree with branch lengths:

((cat:47.14069,(weasel:18.87953,((dog:25.46154,(raccoon:19.19959, bear:6.80041):0.84600):3.87382,(sea_lion:11.99700, seal:12.00300):7.52973):2.09461):20.59201):25.0,monkey:75.85931);

Note that the tree may continue to a new line at any time except in the middle of a name or the middle of a branch length, although in trees written to the tree file this will only be done after a comma.

These representations of trees are a subset of the standard adopted on 24 June 1986 at the annual meetings of the Society for the Study of Evolution by an informal committee (its final session in Newick's lobster restaurant - hence its name, the Newick standard) consisting of Wayne Maddison (author of MacClade), David Swofford (PAUP), F. James Rohlf (NTSYS-PC), Chris Meacham (COMPROB and the original PHYLIP tree drawing programs), James Archie, William H.E. Day, and me. This standard is a generalization of PHYLIP's format, itself based on a well-known representation of trees in terms of parenthesis patterns which is due to the famous mathematician Arthur Cayley, and which has been around for over a century. The standard is now employed by most phylogeny computer programs but unfortunately has yet to be decribed in a formal published description. Other descriptions by me and by Gary Olsen can be accessed using the Web at:

The Options and How To Invoke Them

Most of the programs allow various options that alter the amount of information the program is provided or what is done with the information. Options are selected in the menu.

Common options in the menu

A number of the options from the menu, the U (User tree), G (Global), J (Jumble), O (Outgroup), W (Weights), T (Threshold), M (multiple data sets), and the tree output options, are used so widely that it is best to discuss them in this document.

The U (User tree) option. This option toggles between the default setting, which allows the program to search for the best tree, and the User tree setting, which reads a tree or trees ("user trees") from the input tree file and evaluates them. The input tree file's default name is intree. In many cases the programs will also tolerate having the trees be preceded by a line giving the number of trees:

((Alligator,Bear),((Cow,(Dog,Elephant)),Ferret));
((Alligator,Bear),(((Cow,Dog),Elephant),Ferret));
((Alligator,Bear),((Cow,Dog),(Elephant,Ferret)));

An initial line with the number of trees was formerly required, but this now can be omitted. Some programs require rooted trees, some unrooted trees, and some can handle multifurcating trees. You should read the documentation for the particular program to find out which it requires. Program Retree can be used to convert trees among these forms (on saving a tree from Retree, you are asked whether you want it to be rooted or unrooted).

In using the user tree option, check the pattern of parentheses carefully. The programs do not always detect whether the tree makes sense, and if it does not there will probably be a crash (hopefully, but not inevitably, with an error message indicating the nature of the problem). Trees written out by programs are typically in the proper form.

The G (Global) option. In the programs which construct trees (except for Neighbor, the "...penny" programs and Clique, and of course the "...move" programs where you construct the trees yourself), after all species have been added to the tree a rearrangements phase ensues. In most of these programs the rearrangements are automatically global, which in this case means that subtrees will be removed from the tree and put back on in all possible ways so as to have a better chance of finding a better tree. Since this can be time consuming (it roughly triples the time taken for a run) it is left as an option in some of the programs, specifically Contml, Fitch, and Dnaml. In these programs the G menu option toggles between the default of local rearrangement and global rearrangement. The rearrangements are explained more below.

The J (Jumble) option. In most of the tree construction programs (except for the "...penny" programs and Clique), the exact details of the search of different trees depend on the order of input of species. In these programs J option enables you to tell the program to use a random number generator to choose the input order of species. This option is toggled on and off by selecting option J in the menu. The program will then prompt you for a "seed" for the random number generator. The seed should be an integer between 1 and 2³²-3 (which is 4,294,967,293), and should be of form 4n+1, which means that it must give a remainder of 1 when divided by 4. This can be judged by looking at the last two digits of the number (for example, in the upper limit given above, the last two digits are 93, which is of form 4n+1. Each different seed leads to a different sequence of addition of species. By simply changing the random number seed and re-running the programs one can look for other, and better trees. If the seed entered is not odd, the program will not proceed, but will prompt for another seed.

The Jumble option also causes the program to ask you how many times you want to restart the process. If you answer 10, the program will try ten different orders of species in constructing the trees, and the results printed out will reflect this entire search process (that is, the best trees found among all 10 runs will be printed out, not the best trees from each individual run).

Some people have asked what are good values of the random number seed. The random number seed is used to start a process of choosing "random" (actually pseudorandom) numbers, which behave as if they were unpredictably randomly chosen between 0 and 2³²-1 (which is 4,294,967,295). You could put in the number 133 and find that the next random number was 221,381,825. As they are effectively unpredictable, there is no such thing as a choice that is better than any other, provided that the numbers are of the form 4n+1. However if you re-use a random number seed, the sequence of random numbers that result will be the same as before, resulting in exactly the same series of choices, which may not be what you want.

The O (Outgroup) option. This specifies which species is to have the root of the tree be on the line leading to it. For example, if the outgroup is a species "Mouse" then the root of the tree will be placed in the middle of the branch which is connected to this species, with Mouse branching off on one side of the root and the lineage leading to the rest of the tree on the other. This option is toggled on and off by choosing O in the menu (the alphabetic character O, not the digit 0). When it is on, the program will then prompt for the number of the outgroup (the species being taken in the numerical order that they occur in the input file). Responding by typing 6 and then an Enter character indicates that the sixth species in the data (the 6th in the first set of data if there are multiple data sets) is taken as the outgroup. Outgroup-rooting will not be attempted if the data have already established a root for the tree from some other consideration, and may not be if it is a user-defined tree, despite your invoking the option. Thus programs such as Dollop that produce only rooted trees do not allow the Outgroup option. It is also not available in Kitsch, Dnamlk, or Clique. When it is used, the tree as printed out is still listed as being an unrooted tree, though the outgroup is connected to the bottommost node so that it is easy to visually convert the tree into rooted form.

The T (Threshold) option. This sets a threshold forn the parsimony programs such that if the number of steps counted in a character is higher than the threshold, it will be taken to be the threshold value rather than the actual number of steps. The default is a threshold so high that it will never be surpassed (in which case the steps whill simply be counted). The T menu option toggles on and off asking the user to supply a threshold. The use of thresholds to obtain methods intermediate between parsimony and compatibility methods is described in my 1981b paper. When the T option is in force, the program will prompt for the numerical threshold value. This will be a positive real number greater than 1. In programs Mix, Move, Penny, Protpars, Dnapars, Dnamove, and Dnapenny, do not use threshold values less than or equal to 1.0, as they have no meaning and lead to a tree which depends only on considerations such as the input order of species and not at all on the character state data! In programs Dollop, Dolmove, and Dolpenny the threshold should never be 0.0 or less, for the same reason. The T option is an important and underutilized one: it is, for example, the only way in this package (except for program Dnacomp) to do a compatibility analysis when there are missing data. It is a method of de-weighting characters that evolve rapidly. I wish more people were aware of its properties.

The M (Multiple data sets) option. In menu programs there is an M menu option which allows one to toggle on the multiple data sets option. The program will ask you how many data sets it should expect. The data sets have the same format as the first data set. Here is a (very small) input file with two five-species data sets:

      5    6
Alpha     CCACCA
Beta      CCAAAA
Gamma     CAACCA
Delta     AACAAC
Epsilon   AACCCA
5    6
Alpha     CACACA
Beta      CCAACC
Gamma     CAACAC
Delta     GCCTGG
Epsilon   TGCAAT

The main use of this option will be to allow all of the methods in these programs to be bootstrapped. Using the program Seqboot one can take any DNA, protein, restriction sites, gene frequency or binary character data set and make multiple data sets by bootstrapping. Trees can be produced for all of these using the M option. They will be written on the tree output file if that option is left in force. Then the program Consense can be used with that tree file as its input file. The result is a majority rule consensus tree which can be used to make confidence intervals. The present version of the package allows, with the use of Seqboot and Consense and the M option, bootstrapping of many of the methods in the package.

Programs Dnaml, Dnapars and Pars can also take multiple weights instead of multiple data sets. They can then do bootstrapping by reading in one data set, together with a file of weights that show how the characters (or sites) are reweighted in each bootstrap sample. Thus a site that is omitted in a bootstrap sample has effectively been given weight 0, while a site that has been duplicated has effectively been given weight 2. Seqboot has a menu selection to produce the file of weights information automatically, instead of producing a file of multiple data sets. It can be renamed and used as the input weights file.

The W (Weights) option. This signals the program that, in addition to the data set, you want to read in a series of weights that tell how many times each character is to be counted. If the weight for a character is zero (0) then that character is in effect to be omitted when the tree is evaluated. If it is (1) the character is to be counted once. Some programs allow weights greater than 1 as well. These have the effect that the character is counted as if it were present that many times, so that a weight of 4 means that the character is counted 4 times. The values 0-9 give weights 0 through 9, and the values A-Z give weights 10 through 35. By use of the weights we can give overwhelming weight to some characters, and drop others from the analysis. In the molecular sequence programs only two values of the weights, 0 or 1 are allowed.

The weights are used to analyze subsets of the characters, and also can be used for resampling of the data as in bootstrap and jackknife resampling. For those programs that allow weights to be greater than 1, they can also be used to emphasize information from some characters more strongly than others. Of course, you must have some rationale for doing this.

The weights are to be provided in an input file whose default name is weights. The weights in it are a simple string of digits. Blanks in the weightfile are skipped over and ignored, and the weights can continue to a new line. In programs such as Seqboot that can also output a file of weights, the input weights have a default file name of inweights, and the output file name has a default file name of outweights.

Weights can be used to analyze different subsets of characters (by weighting the rest as zero). Alternatively, in the discrete characters programs they can be used to force a certain group to appear on the phylogeny (in effect confining consideration to only phylogenies containing that group). This is done by adding an imaginary character that has 1's for the members of the group, and 0's for all the other species. That imaginary character is then given the highest weight possible: the result will be that any phylogeny that does not contain that group will be penalized by such a heavy amount that it will not (except in the most unusual circumstances) be considered. Of course, the new character brings extra steps to the tree, but the number of these can be calculated in advance and subtracted out of the total when reporting the results. This use of weights is an important one, and one sadly ignored by many users who could profit from it. In the case of molecular sequences we cannot use weights this way, so that to force a given group to appear we have to add a large extra segment of sites to the molecule, with (say) A's for that group and C's for every other species.

The option to write out the trees into a tree file. This specifies that you want the program to write out the tree not only on its usual output, but also onto a file in nested-parenthesis notation (as described above). This option is sufficiently useful that it is turned on by default in all programs that allow it. You can optionally turn it off if you wish, by typing the appropriate number from the menu (it varies from program to program). This option is useful for creating tree files that can be directly read into the programs, including the consensus tree and tree distance programs, and the tree plotting programs.

The (0) terminal type option . (This is the digit 0, not the alphabetic character O). The program will default to one particular assumption about your terminal (ANSI in the case of Linux, Unix, or Mac OS X, none in the case of Mac OS 9, and IBM PC in the case of Windows). You can alternatively select it to be either an IBM PC, or nothing. This affects the ability of the programs to clear the screen when they display their menus, and the graphics characters used to display trees in the programs Dnamove, Move, Dolmove, and Retree. In the case of Windows, the screen will clear properly with either the IBM PC or the ANSI settings, but the graphics characters needed by Move, Dnamove, Dolmove, or Retree will display correctly only with the IBM PC setting.

All of the programs except Factor, Dnadist, Gendist, Dnainvar, Seqboot, Contrast, Retree, and the plotting and consensus tree programs act to construct an estimate of a phylogeny. Move, Dolmove, and Dnamove let you construct it yourself by hand. All of the rest but Neighbor, the "...penny" programs and Clique make use of a common approach involving additions and rearrangements. They are trying to minimize or maximize some quantity over the space of all possible evolutionary trees. Each program contains a part that, given the topology of the tree, evaluates the quantity that is being minimized or maximized. The straightforward approach would be to evaluate all possible tree topologies one after another and pick the one which, according to the criterion being used, is best. This would not be possible for more than a small number of species, since the number of possible tree topologies is enormous. A review of the literature on the counting of evolutionary trees will be found one of my papers (Felsenstein, 1978a) and in my book (Felsenstein, 2004, chapter 3).

Since we cannot search all topologies, these programs are not guaranteed to always find the best tree, although they seem to do quite well in practice. The strategy they employ is as follows: the species are taken in the order in which they appear in the input file. The first two (in some programs the first three) are taken and a tree constructed containing only those. There is only one possible topology for this tree. Then the next species is taken, and we consider where it might be added to the tree. If the initial tree is (say) a rooted tree with two species and we want the resulting three-species tree to be a bifurcating tree, there are only three places where we could add the third species. Each of these is tried, and each time the resulting tree is evaluated according to the criterion. The best one is chosen to be the basis for further operations. Now we consider adding the fourth species, again at each of the five possible places that would result in a bifurcating tree. Again, the best of these is accepted. This is usually known as the Sequential Addition strategy.

Local rearrangements

The process continues in this manner, with one important exception. After each species is added, and before the next is added, a number of rearrangements of the tree are tried, in an effort to improve it. The algorithms move through the tree, making all possible local rearrangements of the tree. A local rearrangement involves an internal segment of the tree in the following manner. Each internal segment of the tree is of this form (where T1, T2, and T3 are subtrees - parts of the tree that can contain further forks and tips):

the segment we are discussing being indicated by the asterisks. A local rearrangement consists of switching the subtrees T1 and T3 or T2 and T3, so as to obtain one of the following:

Each time a local rearrangement is successful in finding a better tree, the new arrangement is accepted. The phase of local rearrangements does not end until the program can traverse the entire tree, attempting local rearrangements, without finding any that improve the tree.

This strategy of adding species and making local rearrangements will look at about (n-1)x(2n-3) different topologies, though if rearrangements are frequently successful the number may be larger. I have been describing the strategy when rooted trees are being considered. For unrooted trees there is a precisely similar strategy, though the first tree constructed may be a three-species tree and the rearrangements may not start until after the addition of the fifth species.

These local rearrangements have come to be called Nearest Neighbor Interchanges (NNIs) in the phylogeny literature.

Though we are not guaranteed to have found the best tree topology, we are guaranteed that no nearby topology (i. e. none accessible by a single local rearrangement) is better. In this sense we have reached a local optimum of our criterion. Note that the whole process is dependent on the order in which the species are present in the input file. We can try to find a different and better solution by reordering the species in the input file and running the program again (or, more easily, by using the J option). If none of these attempts finds a better solution, then we have some indication that we may have found the best topology, though we can never be certain of this.

Note also that a new topology is never accepted unless it is better than the previous one, so that the rearrangement process can never fall into an endless loop. This is also the way ties in our criterion are resolved, namely by sticking with the tree found first. However, the tree construction programs other than Clique, Contml, Fitch, and Dnaml do keep a record of all trees found that are tied with the best one found. This gives you some immediate idea of which parts of the tree can be altered without affecting the quality of the result.

Global rearrangements

A feature of most of the programs, such as Protpars, Dnapars, Dnacomp, Dnaml, Dnamlk, Restml, Kitsch, Fitch, Contml, Mix, and Dollop, is "global" optimization of the tree. In four of these (Contml, Fitch, Dnaml and Dnamlk) this is an option, G. In the others it automatically applies. When it is present there is an additional stage to the search for the best tree. Each possible subtree is removed from the tree from the tree and added back in all possible places. This process continues until all subtrees can be removed and added again without any improvement in the tree. The purpose of this extra rearrangement is to make it less likely that one or more a species gets "stuck" in a suboptimal region of the space of all possible trees. The use of global optimization results in approximately a tripling (3 x ) of the run-time, which is why I have left it as an option in some of the slower programs.

What PHYLIP calls "global" rearrangements are more properly called SPR (subtree pruning and regrafting) by Swofford et. al. (1996) as distinct from the NNI (nearest neighbor interchange) rearrangements that PHYLIP also uses, and the TBR (tree bisection and reconnection) rearrangements that it does not use. My book (Felsenstein, 2004, chapter 4) contains a review of work on these and other rearrangements and search methods.

The programs doing global optimization print out a dot "." after each group is removed and re-added to the tree, to give the user some sign that the rearrangements are proceeding. A new line of dots is started whenever a new round of global rearrangements is started following an improvement in the tree. On the line before the dots are printed there is printed a bar of the form "!---------------!" to show how many dots to expect. The dots will not be printed out at a uniform rate, but the later dots, which represent removal of larger groups from the tree and trying them consequently in fewer places, will print out more quickly. With some compilers each row of dots may not be printed out until it is complete.

It should be noted that Penny, Dolpenny, Dnapenny and Clique use a more sophisticated strategy of "depth-first search" with a "branch and bound" search method that guarantees that all of the best trees will be found. In the case of Penny, Dolpenny and Dnapenny there can be a considerable sacrifice of computer time if the number of species is greater than about ten: it is a matter for you to consider whether it is worth it for you to guarantee finding all the most parsimonious trees, and that depends on how much free computer time you have! Clique finds all largest cliques, and does so without undue burning of computer time. Although all of these problems that have been investigated fall into the category of "NP-hard" problems that in effect do not have a rapid solution, the cases that cause this trouble for the largest-cliques algorithm in Clique apparently are not biologically realistic and do not occur in actual data.

Multiple jumbles

As just mentioned, for most of these programs the search depends on the order in which the species are entered into the tree. Using the J (Jumble) option you can supply a random number seed which will allow the program to put the species in in a random order. Jumbling can be done multiple times. For example, if you tell the program to do it 10 times, it will go through the tree-building process 10 times, each with a different random order of adding species. It will keep a record of the trees tied for best over the whole process. In other words, it does not just record the best trees from each of the 10 runs, but records the best ones overall. Of course this is slow, taking 10 times longer than a single run. But it does give us a much greater chance of finding all of the most parsimonious trees. In the terminology of Maddison (1991) it can find different "islands" of trees. The present algorithms do not guarantee us to find all trees in a given "island" from a single run, so multiple runs also help explore those "islands" that are found.

Saving multiple tied trees

For the parsimony and compatibility programs, one can have a perfect tie between two or more trees. In these programs these trees are all saved. For the newer parsimony programs such as Dnapars and Pars, global rearrangement is carried out on all of these tied trees. This can be turned off in the menu.

For trees with criteria which are real numbers, such as the distance matrix programs Fitch and Kitsch, and the likelihood programs Dnaml, Dnamlk, Contml, and Restml, it is difficult to get an exact tie between trees. Consequently these programs save only the single best tree (even though the others may be only a tiny bit worse).

Strategy for finding the best tree

In practice, it is advisable to use the Jumble option to evaluate many different orderings of the input species. It is advisable to use the Jumble option and specify that it be done many times (as many as different orderings of the input species). (This is usually not necessary when bootstrapping, though the programs will then default to doing it once to avoid artifacts caused by the order in which species are added to the tree.)

People who want a magic "black box" program whose results they do not have to question (or think about) often are upset that these programs give results that are dependent on the order in which the species are entered in the data. To me this property is an advantage, for it permits you to try different searches for better trees, simply by varying the input order of species. If you do not use the multiple Jumble option, but do multiple individual runs instead, you can easily decide which to pay most attention to - the one or ones that are best according to the criterion employed (for example, with parsimony, the one out of the runs that results in the tree with the fewest changes).

In practice, in a single run, it usually seems best to put species that are likely to be sources of confusion in the topology last, as by the time they are added the arrangement of the earlier species will have stabilized into a good configuration, and then the last few species will by fitted into that topology. There will be less chance this way of a poor initial topology that would affect all subsequent parts of the search. However, a variety of arrangements of the input order of species should be tried, as can be done if the J option is used, and no species should be kept in a fixed place in the order of input. Note that the results of the "...penny" programs and Clique are not sensitive to the input order of species, and Neighbor is only slightly sensistive to it, so that multiple Jumbling is not possible with those programs. Note also that with global search, which is standard in many programs and in others is an option, each group (including each individual species) will be removed and re-added in all possible positions, so that a species causing confusion will have more chance of moving to a new location than it would without global rearrangement.

Probably the most important thing to keep in mind while running any of the parsimony or compatibility programs is not to overinterpret the result. Many users treat the set of most parsimonious trees as if it were a confidence interval. If a group appears in all of the most parsimonious trees then they treat it as well established. Unfortunately the confidence interval on phylogenies appears to be much larger than the set of all most parsimonious trees (Felsenstein, 1985b). Likewise, variation of result among different methods will not be a good indicator of the size of the confidence interval. Consider a simple data set in which, out of 100 binary characters, 51 recommend the unrooted tree ((A,B),(C,D)) and 49 the tree ((A,D),(B,C)). Many different methods will all give the same result on such a data set: they will estimate the tree as ((A,B),(C,D)). Nevertheless it is clear that the 51:49 margin by which this tree is favored is not statistically significantly different from 50:50. So consistency among different methods is a poor guide to statistical significance.

Relative speed of the different programs

C compilers differ in efficiency of the code they generate, and some deal with some features of the language better than with others. Thus a program which is unusually fast on one computer may be unusually slow on another. Nevertheless, as a rough guide to relative execution speeds, I have tested the programs on three data sets, each of which has 10 species and 40 characters. The first is an imaginary one in which all characters are compatible - ("The Willi Hennig Memorial Data Set" as J. S. Farris once called ones like it). The second is the binary recoded form of the fossil horses data set of Camin and Sokal (1965). The third data set has data that is completely random: 10 species and 20 characters that have a 50% chance that each character state is 0 or 1 (or A or G). The data sets thus range from a completely compatible one in which there is no homoplasy (paralellism or convergence), through the horses data set, which requires 29 steps where the possible minimum number would be 20, to the random data set, which requires 49 steps. We can thus see how this increasing messiness of the data affects running times. The three data sets have all had 20 sites of A's added to the end of each sequence, so as to prevent likelihood or distance matrix programs from having infinite branch lengths (the test data sets used for timing previous versions of PHYLIP were the same except that they lacked these 20 extra sites).

    10   40
A         CACACACAAAAAAAAAAACAAAAAAAAAAAAAAAAAAAAA
B         CACACAACAAAAAAAAAACAAAAAAAAAAAAAAAAAAAAA
C         CACAACAAAAAAAAAAAACAAAAAAAAAAAAAAAAAAAAA
D         CAACAAAACAAAAAAAAACAAAAAAAAAAAAAAAAAAAAA
E         CAACAAAAACAAAAAAAACAAAAAAAAAAAAAAAAAAAAA
F         ACAAAAAAAACACACAAAACAAAAAAAAAAAAAAAAAAAA
G         ACAAAAAAAACACAACAAACAAAAAAAAAAAAAAAAAAAA
H         ACAAAAAAAACAACAAAAACAAAAAAAAAAAAAAAAAAAA
I         ACAAAAAAAAACAAAACAACAAAAAAAAAAAAAAAAAAAA
J         ACAAAAAAAAACAAAAACACAAAAAAAAAAAAAAAAAAAA

    10   40
MesohippusAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
HypohippusAAACCCCCCCAAAAAAAAACAAAAAAAAAAAAAAAAAAAA
ArchaeohipCAAAAAAAAAAAAAAAACACAAAAAAAAAAAAAAAAAAAA
ParahippusCAAACAACAACAAAAAAAACAAAAAAAAAAAAAAAAAAAA
MerychippuCCAACCACCACCCCACACCCAAAAAAAAAAAAAAAAAAAA
M. secunduCCAACCACCACCCACACCCCAAAAAAAAAAAAAAAAAAAA
Nannipus  CCAACCACAACCCCACACCCAAAAAAAAAAAAAAAAAAAA
NeohippariCCAACCCCCCCCCCACACCCAAAAAAAAAAAAAAAAAAAA
Calippus  CCAACCACAACCCACACCCCAAAAAAAAAAAAAAAAAAAA
PliohippusCCCACCCCCCCCCACACCCCAAAAAAAAAAAAAAAAAAAA

    10   40
A         CACACAACCAAACAAACCACAAAAAAAAAAAAAAAAAAAA
B         AAACCACACACACAAACCCAAAAAAAAAAAAAAAAAAAAA
C         ACAAAACCAAACCACCCACAAAAAAAAAAAAAAAAAAAAA
D         AAAAACACAACACACCAAACAAAAAAAAAAAAAAAAAAAA
E         AAACAACCACACACAACCAAAAAAAAAAAAAAAAAAAAAA
F         CCCAAACACCCCCAAAAAACAAAAAAAAAAAAAAAAAAAA
G         ACACCCCCACACCCACCAACAAAAAAAAAAAAAAAAAAAA
H         AAAACAACAACCACCCCACCAAAAAAAAAAAAAAAAAAAA
I         ACACAACAACACAAACAACCAAAAAAAAAAAAAAAAAAAA
J         CCAAAAACACCCAACCCAACAAAAAAAAAAAAAAAAAAAA

Here are the timings of many of the version 3.6 programs on these three data sets as run after being compiled by Gnu C (version 3.2) and run on an AMD Athlon XP 2200+ computer under Linux.

	Hennigian Data	Horses Data	Random Data
Protpars	0.00500	0.00670	0.01289
Dnapars	0.01050	0.00940	0.00980
Dnapenny	0.01400	0.00860	1.71100
Dnacomp	0.00240	0.00250	0.00590
Dnaml	0.17749	0.23970	0.21350
Dnamlk	0.21740	0.19450	0.24400
Proml	1.3527	3.2085	2.0055
Promlk	3.3567	8.6078	4.4886
Dnainvar	0.00020	0.00020	0.00020
Dnadist	0.00140	0.00080	0.00150
Protdist	0.09220	0.09210	0.09310
Restml	0.14560	0.28810	0.21540
Restdist	0.00110	0.00090	0.00080
Fitch	0.00760	0.01280	0.00880
Kitsch	0.00180	0.00260	0.00280
Neighbor	0.00020	0.00050	0.00050
Contml	0.01310	0.01500	0.01780
Gendist	0.00070	0.00070	0.00070
Pars	0.00780	0.00610	0.02930
Mix	0.00360	0.00410	0.00610
Penny	0.00190	0.00470	0.8060
Dollop	0.00480	0.00450	0.00820
Dolpenny	0.00200	0.01060	1.1270
Clique	0.00100	0.00070	0.00130

In all cases the programs were run under the default options with optimized compiler switches (-03 -fomit-frame-pointer), except as specified here. The data sets used for the discrete characters programs have 0's and 1's instead of A's and C's. For Contml the A's and C's were made into 0.0's and 1.0's and considered as 40 2-allele loci. For the distance programs 10 x 10 distance matrices were computed from the three data sets. For the restriction sites programs A and C were changed into + and -. It does not make much sense to benchmark Move, Dolmove, or Dnamove, although when there are many characters and many species the response time after each alteration of the tree should be proportional to the product of the number of species and the number of characters. For Dnaml, Dnamlk, and Dnadist the frequencies of the four bases were set to be equal rather than determined empirically as is the default. For Restml the number of enzymes was set to 1.

In most cases, the benchmark was made more accurate by analyzing 100 data sets using the M (Multiple data sets) option and dividing the resulting time by 100. Times were determined as user times using the Linux time command. Several patterns will be apparent from this. The algorithms (Mix, Dollop, Contml, Fitch, Kitsch, Protpars, Dnapars, Dnacomp, and Dnaml, Dnamlk, Restml) that use the above-described addition strategy have run times that do not depend strongly on the messiness of the data. The only exception to this is that if a data set such as the Random data requires extra rounds of global rearrangements it takes longer. The programs differ greatly in run time: the protein likelihood programs Proml and Promlk were very slow, and the other likelihood programs Restml, Dnaml and Contml are slower than the rest of the programs. The protein sequence parsimony program, which has to do a considerable amount of bookkeeping to keep track of which amino acids can mutate to each other, is also relatively slow.

Another class of algorithms includes Penny, Dolpenny, Dnapenny and Clique. These are branch-and-bound methods: in principle they should have execution times that rise exponentially with the number of species and/or characters, and they might be much more sensitive to messy data. This is apparent with Penny, Dolpenny, and Dnapenny, which go from being reasonably fast with clean data to very slow with messy data. Dolpenny is particularly slow on messy data - this is because this algorithm cannot make use of some of the lower-bound calculations that are possible with Dnapenny and Penny. Clique is very fast on all data sets. Although in theory it should bog down if the number of cliques in the data is very large, that does not happen with random data, which in fact has few cliques and those small ones. Apparently the "worst-case" data sets that cause exponential run time are much rarer for Clique than for the other branch-and-bound methods.

Neighbor is quite fast compared to Fitch and Kitsch, and should make it possible to run much larger cases, although the results are expected to be a bit rougher than with those programs.

Speed with different numbers of species

How will the speed depend on the number of species and the number of characters? For the sequential-addition algorithms, the speed should be proportional to somewhere between the cube of the number of species and the square of the number of species, and to the number of characters. Thus a case that has, instead of 10 species and 20 characters, 20 species and 50 characters would take (in the cubic case) 2 x 2 x 2 x 2.5 = 20 times as long. This implies that cases with more than 20 species will be slow, and cases with more than 40 species very slow. This places a premium on working on small subproblems rather than just dumping a whole large data set into the programs.

An exception to these rules will be some of the DNA programs that use an aliasing device to save execution time. In these programs execution time will not necessarily increase proportional to the number of sites, as sites that show the same pattern of nucleotides will be detected as identical and the calculations for them will be done only once, which does not lead to more execution time. This is particularly likely to happen with few species and many sites, or with data sets that have small amounts of evolutionary divergence.

For programs Fitch and Kitsch, the distance matrix is square, so that when we double the number of species we also double the number of "characters", so that running times will go up as the fourth power of the number of species rather than the third power. Thus a 20-species case with Fitch is expected to run sixteen times more slowly than a 10-species case.

For programs like Penny and Clique the run times will rise faster than the cube of the number of species (in fact, they can rise faster than any power since these algorithms are not guaranteed to work in polynomial time). In practice, Penny will frequently bog down above 11 species, while Clique easily deals with larger numbers.

For Neighbor the speed should vary only as the cube of the number of species, so a case twice as large will take only eight times as long. This will make it an attractive alternative to Fitch and Kitsch for large data sets.

Suggestion: If you are unsure of how long a program will take, try it first on a few species, then work your way up until you get a feel for the speed and for what size programs you can afford to run.

Execution time is not the most important criterion for a program, particularly as computer time gets much cheaper than your time or a programmer's time. With workstations on which background jobs can be run all night, execution speed is not overwhelmingly relevant. Some of us have been conditioned by an earlier era of computing to consider execution speed paramount. But ease of use, ease of adaptation to your computer system, and ease of modification are much more important in practice, and in these respects I think these programs are adequate. Only if you are engaged in 1960's style mainframe computing, or if you have very large amounts of data is minimization of execution time paramount. If you spent six months getting your data, it may not be overwhelmingly important whether your run takes 10 seconds or 10 hours.

Nevertheless it would have been nice to have made the programs faster. The present speeds are a compromise between speed and effectiveness: by making them slower and trying more rearrangements in the trees, or by enumerating all possible trees, I could have made the programs more likely to find the best tree. By trying fewer rearrangements I could have speeded them up, but at the cost of finding worse trees. I could also have speeded them up by writing critical sections in assembly language, but this would have sacrificed ease of distribution to new computer systems. There are also some options included in these programs that make it harder to adopt some of the economies of bookkeeping that make other programs faster. However to some extent I have simply made the decision not to spend time trying to speed up program bookkeeping when there were new likelihood and statistical methods to be developed.

Relative speed of different machines

It is interesting to compare different machines using Dnapars as the standard task. One can rate a machine on the Dnapars benchmark by summing the times for all three of the data sets. Here are relative total timings over all three data sets (done with various versions of Dnapars) for some machines, taking an AMD Athlon 1.2 GHz computer running Linux with gcc as the standard. Benchmarks from versions 3.4 and 3.5 of the program are also included (respectively the Pascal and C versions whose timings are in parentheses). They are compared only with each other and are scaled to the rest of the timings using the joint runs on the 386SX and the Pentium MMX 266. This use of separate standards is necessary not because of different languages but because different versions of the package are being compared. Thus, the "Time" is the ratio of the Total to that for the Pentium, adjusted by the scalings of machines using 3.4 and 3.5 when appropriate. The Relative Speed is the reciprocal of the Time.

Machine	Operating System	Compiler	Total	Time	Relative Speed
Toshiba T1100+	MSDOS	Turbo Pascal 3.01A	(269)	10542	0.00009486
Apple Mac Plus	Mac OS	Lightspeed Pascal 2	(175.84)	6891	0.00014511
Toshiba T1100+	MSDOS	Turbo Pascal 5.0	(162)	6349	0.00015750
Macintosh Classic	Mac OS	Think Pascal 3	(160)	6271	0.00015947
Macintosh Classic	Mac OS	Think C	(43.0)	4771	0.0002096
IBM PS2/60	MSDOS	Turbo Pascal 5.0	(58.76)	2303	0.0004343
80286 (12 Mhz)	MSDOS	Turbo Pascal 5.0	(47.09)	1845.4	0.0005419
Apple Mac IIcx	Mac OS	Think Pascal 3	(42)	1645.5	0.0006077
Apple Mac SE/30	Mac OS	Think Pascal 3	(42)	1645.6	0.0006077
Apple Mac IIcx	Mac OS	Lightspeed Pascal 2	(39.84)	1561.6	0.0006404
Apple Mac IIcx	Mac OS	Lightspeed Pascal 2#	(39.69)	1555.0	0.00006431
Zenith Z386 (16MHz)	MSDOS	Turbo Pascal 5.0	(38.27)	1539.0	0.0006498
Macintosh SE/30	Mac OS	Think C	(13.6)	1508.4	0.0006630
386SX (16 MHz)	MSDOS	Turbo Pascal 6.0	(34)	1333.6	0.0007498
386SX (16 MHz)	MSDOS	Microsoft Quick C	(12.01)	1333.6	0.0007499
Sequent-S81	DYNIX	Silicon Valley Pascal	(13.0)	509.0	0.0019646
VAX 11/785	Unix	Berkeley Pascal	(11.9)	466.3	0.002144
80486-33	MSDOS	Turbo Pascal 6.0	(11.46)	449.0	0.02227
Sun 3/60	SunOS	Sun C	(3.93)	435.7	0.002295
NeXT Cube (68030)<