ANNOUNCING PFAM RELEASE 3.4
Pfam is a collection of protein domain family alignments which were
constructed semi-automatically using profile hidden Markov models.
Pfam families contain functional annotation and cross-references to
other databases. Query sequences can be searched against the Pfam
library of profile hidden Markov models at the web sites below.
Pfam 3.4 contains 1407 families. 57% of proteins in SWISSPROT 35
and TrEMBL 5 have at least one match to a Pfam family.
The changes between Pfam release 3.4 and Pfam release 3.3 are
18 new families have been added including the collagen triple
helical repeat, which is found in 390 proteins. The zn-protease
familiy has been removed because it only covered the active site
of a large superfamily and was not able to discriminate well
between members and non-members of the family. The zn-protease
family is being replaced by a number of subfamilies.
For interactive access and searching see URLs
The release is also available in flat file by anonymous ftp:
The Pfam HMM library is compatible with HMMER2 software, available
The Pfam consortium
<birney at sanger.ac.uk>