Path: utzoo!attcan!uunet!shelby!rutgers!mcnc!uvaarpa!murdoch!biochsn.acc.Virginia.EDU!wrp
From: wrp@biochsn.acc.Virginia.EDU (William R. Pearson)
Newsgroups: bionet.software
Subject: Re: Evolutionary tree software
Message-ID: <1990Oct24.134126.16376@murdoch.acc.Virginia.EDU>
Date: 24 Oct 90 13:41:26 GMT
References: <9010232157.AA20584@genbank.bio.net>
Sender: news@murdoch.acc.Virginia.EDU
Organization: University of Virginia
Lines: 33

In article <9010232157.AA20584@genbank.bio.net> DANJ@jhuhyg.bitnet (Dan) writes:

]I'm looking for a program which will determine the "evolutionary tree" of a num
]ber of genes.  This was done by Hanks, Quinn and Hunter(1988).Science,241,442-
]452.  They cited: Fitch and Margoliash(1967),Science 155,279.
]             and  Feng and Doolittle(1987), J. Mol. Evol. 25, 351.
](Hanks et al. used it in a review of protein kinases)
]Does anybody out there have or know where I could find this software?

	A very powerful package of programs for building and
displaying evolutionary trees is available from Joe Felsenstein
(joe@genetics.washington.edu) via anonymous FTP from the same host:
genetics.washington.edu:/pub.  These programs are written in Pascal,
and provide both distance (the Fitch and Margoliash method you cite)
and parsimony (more highly regarded by many evolutionary
practitioners) methods for building trees.

	For protein sequences, parsimony methods require aligned
sequences.  The Feng and Doolittle programs provide one way to do the
multiple alignment (you can get them from the people in Molecular
Biology and Genetics at Johns Hopkins, I do not think that Doolittle
uses the net very much) but an improved method for multiple alignment,
which also provides a tree, is available from Jotun Hein (Hein,
Methods in Enzymology, (1990) v 183 pp 626-645). Hein seems to be in
Japan at the moment; you can try reaching him at:
jhein@ddbj.nig.ac.jp, or at hein@cc.umontreal.ca (this latter address
has a long delay associated with it).

	Multiple alignments are computationally demanding, do not try
them on a computer that charges you for cpu usage, or your advisor
will be very upset.

Bill Pearson