NAME
HMMER
- software suite for making and using hidden Markov models of biological sequences
SYNOPSIS
hmmt - create an HMM from unaligned training sequences
hmmb - build an HMM from a multiple alignment
hmma - construct a multiple alignment of some sequences using an HMM
hmms - search a sequence database for complete (global) matches to an HMM
hmmls
- search a sequence database for local matches of an HMM, to subsequences in longer database sequences (i.e., ignores overhangs on the end of the database sequences)
hmmsw
- search a sequence database for partial matches, of part of an HMM to subsequences in the database sequences: i.e., Smith/Waterman style searching; reports only the single best hit per database sequence; ideal for protein database searches
hmmfs
- search a sequence database for multiple nonoverlapping Smith/Waterman-style matches; linear memory requirement; ideal for nucleic acid database searches
hmme - emit likely sequences from a model
hmm-convert
- convert compact binary HMM files to machineindependent, portable ASCII format
DESCRIPTION
These programs build and use fully probabilistic hidden Markov models (HMMs) of the primary consensus sequence of a family of proteins or nucleic acids. HMMs are similar to "profiles" (Gribskov, McLachlan, and Eisenberg, PNAS 84:4355-4358 1987) but are more firmly grounded theoretically, and seem to be more powerful empirically. HMMs are useful both for multiple sequence alignment and for sensitive searching of sequence databases. HMMs for profile-like biological sequence analysis were introduced by David Haussler and colleagues; see Krogh et al., J. Mol. Biol. 235:1501-1531 1994.
For all the programs, unaligned sequence files can be in FASTA, Genbank, EMBL, or Swissprot format. The programs automatically detect what format the file is in. Aligned sequence files can be in GCG MSF format or SELEX format. SELEX format is a simple format of one line per sequence, containing the name first, followed by the aligned sequence. MSF and SELEX alignment files can also be used where unaligned format files are required; the sequences will be read in and their gaps removed.
If you forget the command-line syntax or available options of any of the programs, you can type just the name of the program with no other arguments and get a short help message. If you also include option {t -h}, you get version info as well (the software version number is helpful if you report bugs or other problems to me).
The programs work on RNA, DNA, and protein sequence. They automatically detect what your sequences are. However, the behavior of the programs when a nucleic acid model is used to analyze protein sequences, or vice versa, is undefined, and certain other situations may arise (such as searching the complementary strand of a database) that are nonsensical in one context. Be forewarned.
Some of the programs may have experimental options listed for their usage. These options are separated from the rest and annotated as experimental. Supported options are in small letters, experimental options are in caps. Use experimental options at your own risk; they're just my own mechanism for trying out new ideas, and I don't always support them adequately.
For more information, see man pages for the individual programs, or the User's Guide which is part of this distribution.
SEE ALSO
Individual man pages: hmmb(l), hmme(l), hmmfs(l), hmmls(l), hmms(l), hmmsw(l), hmmt(l), hmm-convert(l)
User guide and tutorial: Userguide.ps
BUGS
Please report bugs to Sean Eddy at eddy@genetics.wustl.edu. Reported bugs are generally fixed within hours or days and new code is immediately made available via ftp and the Web.
NOTES
This software and documentation is Copyright (C) 1992-1995,
Sean R. Eddy. It is freely distributable under terms of the GNU General Public License. See COPYING, in the source code distribution, for more details, or contact me.
Sean Eddy
Dept. of Genetics, Washington Univ. School of Medicine 660 S. Euclid Box 8232
St Louis, MO 63110 USA
Phone: 1-314-362-7666
FAX : 1-314-362-2985
Email: eddy@genetics.wustl.edu