The methods of the previous chapter were applied to cluster all the
protein sequences in the SWISSPROT database.  The resulting
classification splits the protein space into well defined groups of
proteins, which are closely correlated with natural biological
families and superfamilies.  Nearly all of the clusters we found seem
biologically meaningful, some corresponding to well known families,
and many others representing less studied families. Some clusters
consist exclusively of unknown proteins or hypothetical proteins.  Our
analysis concerns complete proteins, and it is not limited only to
those subsequences which are identified as functionally or
structurally important motifs and domains.  Indeed, not all the
emerging clusters are correlated with a specific domain, although some
clusters are characterized by a domain that is common to many or all
member proteins.

Needless to say, this overwhelming body of information cannot be
properly surveyed here.  An interactive web site that contains the
results of our analysis has been constructed \\
(http://www.protomap.cs.huji.ac.il), where users can get acquainted
with this ``map'' of the protein space.  The following sections give
the results of the overall assessment of our classification in
comparison with several well known domain-based databases, and offer a
glimpse of our classification.