The main entry point into the STRING website is the
protein search box on its start page. It supports queries
for multiple proteins, can be restricted to certain organisms
or clades of organisms, and uses a weighted scheme
to rank annotation text matches and identifier matches.
Users can also arrive via a number of external websites (29–
32) that maintain cross-links with STRING, including the
partner resources Search Tool for Interactions of Chemicals
(STITCH; 33) and eggNOG (34)––the latter both
share protein sequences, annotations and name-spaces with
STRING.Athirdway to enter STRING is via logging on to
the My Data section; this allows users to upload gene-lists,
create identifier mappings, view their browsing history and
provide additional ‘payload’ data to be displayed alongside
the interactions.
Once a protein or set of proteins is identified, users proceed
to the network view (Figure 1). From there, it is possible
to inspect the interaction evidence, to re-adjust the
score-cutoffs and network size limits and to view detailed
information about the interacting proteins. Upon switching
to the ‘advanced’ mode (via the tool panel below the
network), users can also cluster and rearrange the network
and test for statistical enrichments in the network. The latter
feature has been enhanced for the current version 10.0
of STRING: enrichment detection now also covers human
disease associations and tissue annotations, which might
be statistically enriched in a given network. For this feature,
STRING connects with the partner databases TISSUES
(http://tissues.jensenlab.org) and DISEASES (http:
//diseases.jensenlab.org), which also share sequence and
name spaces with STRING, and which annotate proteins
to tissues or to disease entities based on a combination of
automated text-mining and knowledge imports.