Biology Reference
In-Depth Information
Nevertheless, recent publications from both academia and industry
convincingly demonstrate that careful application of structure-based
virtual screening in combination with follow-up experimental verifica-
tion can indeed lead to the discovery of new compounds active against
diverse enzymes, receptors, and clinically relevant drug targets, such as
NF-
111
the histone arginine methyl-
transferase,
112
a cytochrome P450,
113
a fish estrogen receptor,
114
or the
CK2 protein kinase
66
; and are able to complement assay-based high-
throughput screening approaches.
115
Further examples of successful
structure-based virtual screening have recently been summarized in
Cavasotto and Orry.
116
κ
B,
110
the nuclear receptor PPAR
γ,
6. Protein Model Portal
While the PDB currently holds approximately 51 000 experimentally
derived coordinate entries representing 18 000 different proteins, several
millions of comparative protein models have been generated for the
protein sequences contained in the UniProtKB database using these
experimentally elucidated structures as templates.
13,17,18,20,117
Databases of
annotated comparative models such as SWISS-MODEL Repository
17, 20
and others
117-119
allow cross-referencing with other nonstructure-centric
resources, and make comparative models accessible to nonexperts.
We have developed the Protein Model Portal (http://www.protein-
modelportal.org) as part of the PSI Structural Genomics Knowledgebase
to provide a single portal to all structure information available for a given
protein within various databases, thereby implementing the first step of
the community workshop recommendation
120
on archiving structural
models of biological macromolecules. The Protein Model Portal cur-
rently allows querying models from six structural genomics centers,
MODBASE, and SWISS-MODEL Repository, as well as PDB experi-
mental structures, through a single search interface. All models
are mapped to a common unique sequence reference system based
on UniProt/UniParc,
121
which allows one to cross-query all resources
simultaneously and to dynamically annotate the model target sequen-
ces with functional
121
and domain annotation.
122
The current release
