UVa FASTA Downloads

Unix/MacOSX source code for the FASTA programs is available from http://faculty.virginia.edu/wrpearson/fasta.

The latest versions of the program are in http://faculty.virginia.edu/wrpearson/fasta/CURRENT/ and http://faculty.virginia.edu/wrpearson/fasta/CURRENT/windows.

Overview

To run the FASTA programs on your own computers, you will need to (1) download and install the programs, and (2) download some databases to search.

Current versions - The current version of the FASTA programs is version 35, which includes fasta35, ssearch35, fastx/y35, tfastx/y35, prss35, prfx35, lalign35 etc. All of the fasta3 programs can be downloaded in a single file, either as Unix/MacOSX source code or as a Windows ZIP archive. (The fasta3.tar.gz and fasta3/windows/fasta3.zip files are the same as the files in the CURRENT directory.

Version 35 of the FASTA programs includes several major improvements: statistical estimates are much more accurate; threaded versions of the program are more efficient; and version 35 now includes lalign35, a program that finds multiple local alignments using Webb Miller's sim algorithm (and implementation of the Waterman and Eggert approach, PubMed) In addition, FASTAv35 includes ggsearch35 and glsearch35, which perform optimal global-global and global-local alignments and searches.

Older versions - In the past, a complete installation of the FASTA programs required downloading both the fasta3 and fasta2 programs, because some programs, notably lalign and align, were available only in fasta2. Now that lalign35 and ggsearch35 are available in the fasta3 package, there is much less reason to download fasta2. Nevertheless, the most recent version of the fasta2 programs can be downloaded as Unix/MacOSX source code or as a Windows ZIP archive

Downloading and Installing the FASTA3 programs

1. Uncompress and "un-tar" the fasta3.tar.gz file. At the Unix/Linux/MacOSX command line type: gunzip -c fasta3.tar.gz | tar xvf -

2. The fasta3 distribution has Makefile's for most common architectures, including linux, Sun Solaris, MacOSX (both PowerPC and Intel), freebsd, IBM/Aix, etc. To see a list of the possible architectures, type (at the command line):
   ls -l Makefile.*

   Once you have identified a Makefile for your machine, e.g. linux, type:
   make -f Makefile.linux fasta35
   To compile just the fasta program or:
   make -f Makefile.linux_sse2 all
   To compile all the programs.

   You may encounter some warning messages, which can be ignored.

   By default, Makefile.os_x provides an Altivec accelerated version of Smith-Waterman, and Makefiles with sse2 include an Intel SSE2 accelerated Smith-Waterman. These Smith-Waterman versions are typically more than 10X faster than unaccelerated versions, and can provide very fast sequence and profile Smith-Waterman searches.

   In addition, makefiles that include sql in their name can be used to make versions of the programs that can read SQL databases directly.

3. Once the programs are compiled, you can test whether fasta works by typing:
   fasta35 -q mgstm1.aa prot_test.lseg
   which runs a test query sequence against a small test library.
4. Once the programs are compiled, you may want to copy them to a more visible location, e.g. /usr/local/bin or /seqprg/bin. The easiest way to copy all the programs is to edit the Makefile you used and specify the executable directory in the line that starts: XDIR=/seqprg/bin. You can then type make -f Makefile.linux install to install the programs, and make -f Makefile.linux clean-up to remove unnecessary files from the installation directory.

Downloading Sequence databases