PFAM Database

Copyright (C) 1996-1997 The Pfam Consortium
Pfam is freely distributed under the terms of the GNU Library General Public License.

PFAM is a protein families database of alignments and HMMs. It is a large collection of multiple sequence alignments and hidden Markov models covering many common protein domains.

Pfam's alignments are of complete domains, not of motif signatures. Pfam's development is tightly tied to the development of hidden Markov model (HMM) profile methods of modeling protein multiple alignments. Our goal is to enable rapid automatic annotation of common protein domains.

Pfam multiple alignments come in two forms. "Seed" alignments are of representative, nonredundant sets of sequences. Seeds are checked reasonably carefully in a manual alignment editor. "Full" alignments are HMM-generated automatic alignments of every homologous domain in the current release of Swissprot that is detected by an HMM built from the seed alignment. This strategy lets us automatically update Pfam with ease; our seed alignments are (mostly) stable from release to release.

Pfam was originally a collaboration between Erik Sonnhammer (Sanger Centre , Hinxton UK) and Sean Eddy (Dept. of Genetics, Washington University, St. Louis USA). Pfam is now a consortium that includes researchers at the Sanger Centre in Hinxton UK, the WashU Dept. of Genetics in St. Louis USA, and Erik Sonnhammer, now at the NCBI.

Installed on the Opteron cluster.

Opteron version:

  • The database files are in the directory /biomed/db/pfam/current/.

Reference:

  • E.L.L. Sonnhammer, S.R. Eddy, R. Durbin, "Pfam: A Comprehensive Database of Protein Families Based on Seed Alignments" Proteins (28:405-420, 1997).

See also: