Biopolymer - Structure Database Searching

Biopolymer

7 Structure Database Searching

Introduction

The Str_DB pulldown is in the Biopolymer module in Insight II. Insight II provides graphical interface to a protein database search program, template_db_search. The protein fragments which satisfy the search query may be read into Insight and displayed.

The database information is derived for a set of known protein structures from the Brookhaven Protein Databank and stored in the file $BIOSYM/data/biopolymer/database.dat. The same list of proteins used in loop searches are used to create this database. The database $BIOSYM/gifts/biopolymer/structure.db consists of all the pdb files release by PDB before 8/6/97.

The stored structural information is mostly residue-based and includes some information on atoms and secondary structure. Searches of this database can address many common protein structure questions.

Tutorials

The Biopolymer Pilot tutorial lesson 8 describes the creation of a custom database of pepsin-like and retroviral aspartic proteases and subsequently uses this database in a search exercise.

To start Pilot, click the mortarboard icon on the Insight II toolbar. When the Pilot interface appears, click Select and then choose the Biopolymer tutorial from the list.

Methodology and Implementation

Searching the structural database

The protein structural information has been extracted from a set of proteins in the current Brookhaven Protein Databank and stored in a single file which by default is called $BIOSYM/data/biopolymer/database.dat.

For each protein the stored information consists of:

Name of the Brookhaven PDB file
Text description of protein name, family, crystallographic form and refinement method
Number of segments, residues, atoms and solvent atoms
Ligands residue name, description and number of atoms

For each residue the stored information consists of:

Segment name
Residue name and type
Secondary structure type ¹
C atom coordinates
Side chain "center" coordinate
Pseudo torsion between four successive C atoms in a chain
Main chain and angles
Side chain torsion angles

A program external to Insight, called template_db_search, is used to search the structural database file. This program can be run stand-alone using a control file created within Insight or edited manually.

Defining a structure database query

There are four types of information that can be used to define a structure database query:

Proteins to search

The search can be limited to those proteins with a resolution better than some specified value or using a keyword such as "`hemoglobin" in their description. However, by default all proteins are searched.

Templates

A template is a fragment of consecutive residues. Each residue may be defined by residue type (or some wildcard group of residue types such as "hydrophobic"), by secondary structure type, by main chain or side chain torsions or by intra-template C-C distance. Each residue in the template may be defined as much or as little as required.

Constraints

A constraint is a limiting relationship between two templates. It specifies that the two templates must be a certain distance apart or a certain number of residues apart in the sequence. For example:

A limit of maximum and minimum allowed distance between C atoms and/or side chain "center"
A limit on the maximum and minimum number of residues in sequence between residues in different templates

If two or more templates are defined, they must be linked by constraints, otherwise they are completely independent and should be treated as separate database queries.

A database search usually finds a number of hit fragments that satisfy the query.

Search parameters

The user can specify how tightly the search criteria must be adhered to by changing the search tolerance. For example, if you specify a required main chain or side chain torsion, then, by default, all structures within 30° of the required value will be retrieved. After you enter the required information using the Define_Template, IntraTmplt_Cnstrt, InterTmplt_Cnstrt and Run_Search commands, the control file for the search program is generated, and the search runs as a background job under Insight II.

The insight II interface to the structure database search program produces an input file called run_name.ddb that contains the details of the search query, name of database file to search, maximum number of protein hits, etc. In the following example a local database (called ./asp_proteinase.db) is searched for a maximum of 50 hits. The two templates are denoted asp1 and asp2 respectively and both consist of tripeptides ASP-any-GLY with no side chain or main chain conformation restrictions. The query requires hits to have an ASP CA to ASP CA distance between 5.5 and 7.5 Å and that the second template is after the first template (i.e., between 0 and 500 residues) in the protein sequence. Such a query finds both pepsin-like and retroviral aspartic proteases.

A graphical representation of some of the details of this query is shown in Figure 10 below.

Figure 10 . A graphical representation of an example query.

An example of the input file that represents this query is shown below:


NHIT 50

DTOL 1.0 

ATOL 30.0 30.0 

CTOL 50.0 

STOL 30.0 

DBAS ./asp_proteinase.db

WILD ./wildcard.dat

RSLN 10.0

TMPL 3 asp1

RESD ASP * * * *

RESD ANY * * * *

RESD GLY * * * *

TMPL 3 asp2

RESD ASP * * * *

RESD ANY * * * *

RESD GLY * * * *

CONS CACA asp1:1 asp2:1 5.5 7.5 0

CONS IRNG asp1:1 asp2:1 0 500 0

The default database file used for searching is found in $BIOSYM/data/biopolymer/database.dat.

Browsing database search results

The output of the Run_Search command is a list of hits stored in an Insight II table file called run_name.tab. In addition, a summary of the calculation, including query details and the hits found is included in a file run_name.log in the local directory. The Read_Search_Result command allows the list of protein structural hits to be loaded into an Insight II table by setting the Load_Type parameter to Load_Hit_List. When the Load_Type parameter is set to Load_Hit_Protein, then one or more protein structure hits may be loaded into Insight II and superimposed using the atom subsets that are then automatically defined.

For example, if the table of structure database hits contains the following rows


Table 1



Hit     Protein asp1:ASP  asp2:ANY  asp1:GLY  asp2:ASP asp2:ANY asp2:GLY

........

hit7    5pep    32:ASP    33:THR    34:GLY    215:ASP  216:THR  217:GLY

.....

hit9    4hvp    A25:ASP   A26:THR   A27:GLY   B25:ASP  B26:THR  B27:GLY\

.....

then loading hits 5 and 7 will produce subsets named HIT7$PEP and HIT9$HVP respectively, and the subsets can simply be used in the Transform/Superimpose command in the viewer module to superimpose the two hits.

The two hits in the above table represent porcine pepsin and HIV protease, respectively.


Template 1:      H-H-H-H-H-H-N-N-N-N-N-N

Template 2:                  N-N-N-N-N-N-H-H-H-H-H-H



Fragment found: H-H-H-H-H-H-N-N-N-N-N-N-H-H-H-H-H-H

First residues in the two templates separated by 12 residues:


Template 1:      H-H-H-H-H-H-N-N-N-N-N-N

Template 2:                              N-N-N-N-N-N-H-H-H-H-H-H



Fragment found: H-H-H-H-H-H-N-N-N-N-N-N-N-N-N-N-N-N-H-H-H-H-H-H

In the case which the first residues area separated by 6 residues, some residues in the hit structure are in both templates.

Database search details

When the Run_Search command is executed, Insight II will launch a background job by calling the script:


$BIOSYM/bin/biopolymer/Run_Search

The structure database search is then performed by the executable


$BIOSYM/$BIOSYM_PLATFORM/biosym_exe/template_db_search

The job can be made to run locally or remotely by setting the Background_Job/Setup_Bkgd_Job/Background_Job parameter to Run_Search and setting the Host parameter to either Local or to the hostname of a machine on the local area network. The progress of the search calculation can be monitored using the Background_Job/Completion_Status command and the job aborted using the Background_Job/Kill_Bkgd_Job command. For the supplied database.dat file a simply query of the database can be performed in a few minutes.

The Run_Search command will default to a maximum of 50 hits. In addition to setting the torsion angle constraint tolerances the search can be restricted to crystal structures of a given resolution by the setting the Xtal_Resolution parameter to High, Medium or Low. The definitions of these terms are as follows:

High < 2.0 Angstroms.
Medium < 5.0 Angstroms.
Low < 10.0 Angstroms.

The Run_Search/Job_Name parameter is used as a root name, and the input and output file names are all derived from this root name. The structure database search calculation can be run from the command line as follows:


Run_Search Job_Name

and the same set of files produced by running from within Insight II will be produced.

Custom database creation

The Str_DB/Create_DB command may be used to create a customized version of the database file searched by the Run_Search command. The the PDBList_File parameter accepts the name of a file containing a list of names of PDB files. This file has a simple format, each line must specify the full directory and path name of the PDB file to be used. When a Job_Name parameter has been specified, the database creation runs as a background job. A script


$BIOSYM/bin/biopolymer/Run_Crebase

is called from Insight II and this in turn calls the executable


$BIOSYM/$BIOSYM_PLATFORM/biosym_exe/crebase

which creates the run_name.db file used in the search. By default, the Create_DB command will create a new database file, but if the Append_DB boolean is turned on, then the new data will be added to an existing run_name.db file.

Overview of commands

What follows is a brief summary of the function of each command in the Str_DB pulldown.

Define_Template

A template is a fragment of consecutive residues. The Define_Template command allows you to either create a template by sequentially adding residues to the template or to modify an existing template. Each residue in the template may be defined by residue type (or a wildcard group of residue types such as "hydrophobic"), by secondary structure type, or by main chain or side chain torsions. Each residue in the template may be defined as much or as little as required.

There are two ways to add a residue to a template. If you load a protein structure into Insight, you can click the residue you want to add to the template. The parameters for this residue will be filled automatically. You can also add a residue by typing in values or selecting from the value aid.

IntraTmplt_Cnstrnt

The Intratmplt_Cnstrnt command defines the relationship of two residues in a template in terms of the distance between two Catoms, a C atom and side chain center, or two side chain centers.

InterTmplt_Cnstrnt

The Intertmplt_Cnstrnt command defines a constraint between two residues in different templates in terms of the residue distance or residue separation range in a protein sequence.

The distance between two residues can be between two Catoms, a C atom and side chain center, or two side chain centers.

Delete_Query

The Delete_Query command allows the user to delete individual templates or constraints, or to delete all elements of a query.

List_Query

The List_Query command lists defined queries to the text port or to a file. The format of the list is the same as that in the command file jobname.ddb written out by the Run_Search command. Please refer to the manual for details of the format.

Run_Search

The Run_Search command writes out the control file for searching the database and starts the search as a background job.

Read_Search_Results

The Read_Search_Results command allows you to load either a table of hit lists or a list of hit pdb files with a subset of hit residues.

Create_DB

The Create_DB command allows you to create a database file from a list of pdb file names.

¹ Kabsch and Sanders method is used to determine secondary structure type.

Last updated September 30, 1997 at 11:31AM PDT.