Tutorial: Creating neighbor-joining trees with jackknife
analysis in Paup.

NOTE: I HAVE ONLY TESTED THESE ON A FEW ALIGNMENTS - PLEASE
LET ME KNOW IF THERE ARE ANY PROBLEMS! THANKS FOR HEPLING ME
WORK THE BUGS OUT OF THESE PROGRAMS.

PAUP is a fast and powerful phylogenetics program by David
Swofford (http://paup.csit.fsu.edu/). We have access to a
command-line version of this program in our nun accounts.
GCG also has a driver program for PAUP called PAUPSEARCH
(see the GCG manual).

PAUP can do a lot of things, but we usually just want to
make the same one or two types of tree. To make things
easier, I've written a program that converts fasta format
sequence alignments to NEXUS format (the format PAUP
requires) and appends some PAUP commands to create a PAUP
execution script. Take a look at the execution scripts once
you've created them to get an idea of what the PAUP commands
look like.

Here are instructions for creating trees in PAUP. Because
there's a bug in PAUP that prevents node labels on bootstrap
trees from being written properly, I've written the program
to perform a jackknife analysis instead - this is another
data resampling method that gives results similar to
bootstrapping. paupboot.py is set to make neighbor-joining
(nj) trees according to the GTR (General Time-Reversible)
model of evolution with a gamma distribution of rates, aka
REV-gamma.

I have indicated UNIX commands with a $ prompt; for example,
$ ls means type 'ls' at the prompt (it doesn't mean to type
'$ ls'). Output from commands is indicated with an '-->'.
All commands should be typed on a single line.

You will find the following programs useful (*see below).
Instructions are available by typing -h after the program
name. Command line syntax is very similar to GCG (see the
examples).

paupboot.py         writes PAUP execution scripts
fastaview.py        view FASTA format alignments as
                    interleaved and numbered 
trim.py             remove sub-regions of an alignment 
treeparser.py       modify existing trees.

Examples use the alignment below (you can copy-paste it into
a new file called test.fasta) like this:

$ cat > test.fasta [paste the alignment at the prompt, hit
return then Ctrl+D]

or like this

$ pico test.fasta [paste the alignment at the prompt, hit
Ctrl+X and follow the instructions]

>s1
ATGAGAGTGAGGGAGATCAAGAGGAATTATCAGCTCCTATGGAGATGGGGCATCATGCTC
CTTGGGATATTAATGATCTGTAAT---GAACAATTATGG 
>s2 
ATGAGAGTGAAGGGGATCAGGAGGAATTGTCAGCACTGGTGGAAATGGGGCATCATGCTC
CTTGGGATATTAATGATCTGTAATGCTGAACAATTGTGG 
>s3
ATGAGAGTGAAGGGGATCAGGAGGAATTGTCAGCGCTGGTGGAAATGGGGCATCATGCTC
CTTGGGATATTAATGATCTGTAATGCTGAACAATTGTGG 
>s4
ATGAGAGTGAAGGGGATCAGGAGGAATTATCAGCACTGGTGGAAATGGGGCATCATGCTC
CTTGGGATATTAATGATCTGTAATGCTGAACAATTGTGG 
>s5
ATGAAAGTGAAGGAGACCAAGAGGAATTGGCAGCGCTTGTGGAGATGGGGCATCATGCTC
CTTGGGATGTTGATGATCTGTAGTGCAGAAAAATTGTGG


1. Align your sequences and put them into FASTA format. Make
sure your sequence names don't contain non-alphanumeric
characters other than '_' (underscore). (This is a good
general rule to follow when naming sequences). Remove
poorly-aligned regions with lots of gaps with trim.py.

(for example, $ trim.py buh.fasta -r=1-100 -out=buh_trimmed.fasta 
will write the first 100 characters of each sequence in 
buh.fasta to buh_trimmed.fasta).

************************** 
2. Create the PAUP execution script:

The following command will create an execution script for a
neighbor-joining tree. 

$ paupboot.py test.fasta -m=nj 
--> Writing test.njscript

Or you can do a jackknife analysis (100 replicates): 

$ paupboot.py test.fasta -m=jack 
--> Writing test.jackscript

This second analysis will label nodes according to the
number of jackknife replicates that contained the same set
of sequences branching from that node. This is a measure of
our confidence in those branches.

************************** 
3. Run PAUP 
You may have to add PAUP to your account by typing 
$ ipm paup

Once you have done this, execute the paup scripts. The
jackknife analysis can take a while for large alignments.

$ paup test.njscript 
...

--> 1 tree saved to file "~/020607_testpaup/test_nj.pauptrees"

$ paup test.jackscript 
... 
--> 1 tree saved to file "~/020607_testpaup/test_jack.pauptrees"

************************** 
4. (Optional) Combine the results from these two analyses: 

PAUP has no mechanism for showing
the results of a jackknife analysis and branch lengths on
the same tree. You can combine the analyses using
treeparser.py. This command will transfer the node labels
from test_jack.pauptrees to test_nj.pauptrees, and will
write a new file.

$ treeparser.py -intree=test_nj.pauptrees
-nodelab=test_jack.pauptrees -out=test_combo.pauptrees 
--> Writing test_combo.pauptrees ...

************************** 
5. View and save the trees.

Treeview will open the output from PAUP and treeparser
directly. Treeview is available for both OS 9.x and OSX
(maybe also for PC?). Treeview for 9.x has more options. To
show the node labels, select "Tree" --> "Show Internal Node
Labels" in the program menu. Trees exported from Treeview
for 9.x can be opened in Illustrator and modified.

6. Recap: All of the above can be distilled into the
following lines: 
$ paupboot.py test.fasta -m=nj 
$ paupboot.py test.fasta -m=jack 
$ paup test.njscript 
$ paup test.jackscript 
(the second two lines can be skipped if you
add -e to the first two commands) 
$ treeparser.py -intree=test_nj.pauptrees -nodelab=test_jack.pauptrees

That's it! 
************************** 

* you can get access
to the programs I have written with this command: 
$ /afs/isis.unc.edu/home/n/g/nghoffma/public/add_nh_profile

This command adds my directories to the search path in your
account; You must log out and log back in for the changes to
take effect.

Don't do this if you have modified the file
~/public/.profile.personal to customize your environment:
instead, look at
/afs/isis.unc.edu/home/n/g/nghoffma/public/.profile.personal
to see what to add to your path.

A note on these programs: output from one program can be
piped to another. For example, to trim buh.fasta, translate
the output, and view the translation in an interleaved
format, use this command:

$ trim.py buh.fasta -r=1-100 | translate.py | fastaview.py