Advanced search provides
the capability of combining multiple searches of specific types of data
in a logical AND or OR. The result is a list of structures that comply
with ALL or ANY search criteria, respectively. More complex logic will
be provided in a later release of the RCSB PDB. The easiest way to
understand how to use advanced search is to work through an example:
This will produce a long
list of possible query types organized by category. If you are not sure
of what one to pick select one and a brief description will appear.
Once a query type is selected you can also select the question mark to
get more help on that query. If it was not what you were seeking,
select a different query type.
In the above example a query type of UniProt Accession Number(s) is selected and a brief prompt is provided.
Click the button that says
Result Count. This will provide the number of PDB structures that match the query.
In this example 183 structures are found to contain
this UniProt identifier P69905. Note that this query is matching a
sequence, but it is structures that are returned. That is, the
structure may contain not only this single sequence but other different
sequences as well. All structures will be listed that contain the
selected sequence, irrespective of what else they contain.
The result count provides a
sense of how focused the query has become and whether it needs to be
refined further. At this point the user may select
Submit Query to list the 183 structures that match the search criteria or further refine the query.
By selecting the + a search criteria can be added.
Here we have added a search (query type deposition
date) for all structures deposited between January 1, 2008 and November
1, 2009. As previously the result count indicates the number of
structure that match the single criteria. In this example 10008
structures were added in that period.
At this point additional search criteria could be added if desired. Alternatively the
button will return all structures deposited in the selected time period
AND which contain the designated protein sequence from UniProt.
While not relevant here since structures are selected
that match a specific sequence, in other circumstances you may reduce
the number of sequence redundant structures returned by using the
Remove Similar Sequences option.
In the above example the 10008 structures deposited
between January 1, 2008 and November 1, 2009 are filtered such that if
any one of the polypeptide chains in a given structure matches any
polypeptide chain in another structure at 90% or greater sequence
identity, only one of those structures will be displayed. For this
query this has the effect of reducing the number of structures
resulting from the query from 10008 to 5523. A variety of sequence
identities can be selected as cutoffs.
The default is a AND condition between individual components of the query which is defined by the
Match all condition shown below.
The alternative is
Match Any which is a logical OR operator.
Several types of searches that are related with Chemical Components like "Chemical Name" or "SMILES", provide
the option to retrieve Ligand (short term for chemical components) results instead of the default "Structure" results.
By selecting "Ligand" from the "Results" drop box for a Chemical Name search, for the word "adenosine" for example,
one may retrieve the Ligands that match the search criteria - instead of the PDB Structures that contain such Ligands.
It is possible to use this option even in composite searches as long as one of the sub-queries relates to chemical components.