Features

• Atomic interactions represented as Structural Interaction Fingerprints (SIFts)
• 23 different interaction types including 9 ring-ring interaction geometries and 3 different atom-ring interaction types
• Identification of modified and mutated residues in PDB structures
• Hierarchical & exhaustive RECAP fragmentation of PDB ligands
• Calculation of Ultrafast Shape Recognition (USR) moments for all PDB ligands and up to 200 conformer per chemical component
• Structural alignments of binding sites based on SCOP 1.75
• Sequence-to-Structure mapping (UniProt to PDB)
• Calculation of various surface area descriptors such as solvent-exluded surface area or gyration
• Identification of natural substrates, products and cofactors through KEGG
• Identification of FDA approved drugs
• Application Programming Interface (API) written in Python
• PyMOL interface
• Cross references to external databases

Documentation

• Interaction types in Credo
• The Credo database schema
• The Credo API tutorial
• Using the Credo API with PyMOL for visualisation.

Screenshots

A gallery of screenshots can be viewed here. The structural interactions for ligand J07 in PDB structure 2P33 can be downloaded as PyMOL scene at the bottom of the page.

Downloading the database

The Credo database table dumps and database schema are available for download here. The schema.sql file is part of the mercurial repository.

Supported Relational database management systems (RDBMS)

The Credo database was created with MySQL 5.1.31 using the MyISAM engine,i.e. foreign keys are not used in the database and the SQLAlchemy mappers. Hence, it should be straightforward to use the table dumps to recreate Credo with a different RDBMS such as PostgreSQL or Oracle.

Required Python modules

• Numpy
• Mercurial to checkout the source code for the Credo API.
• SQLAlchemy version 0.5+ to use the API
• MyChem 0.7.0 is used for cheminformatics functionality in the full API.
• IPython is recommended as interactive Python shell.
• PyMOL to use the visualisation functions.

Creating the schema and populating the database

First, you will have to download the schema and the database tables dumps. The database can be recreated from the schema by

51&q=CREATE&lr=lang_en">CREATE 51&q=DATABASE&lr=lang_en">DATABASE CREDO;
51&q=USE&lr=lang_en">USE CREDO;
SOURCE 51&q=SCHEMA&lr=lang_en">schema.sql;

The data can be inserted by either from the mysql command line

51&q=USE&lr=lang_en">USE CREDO;
51&q=LOAD&lr=lang_en">LOAD 51&q=DATA&lr=lang_en">DATA 51&q=LOCAL&lr=lang_en">LOCAL 51&q=INFILE&lr=lang_en">INFILE <file> 51&q=INTO&lr=lang_en">INTO 51&q=TABLE&lr=lang_en">TABLE <table> 51&q=FIELDS&lr=lang_en">FIELDS 51&q=TERMINATED%20BY&lr=lang_en">TERMINATED BY '@\t' 51&q=LINES&lr=lang_en">LINES  51&q=TERMINATED%20BY&lr=lang_en">TERMINATED BY '#\n'; 

or by using mysqlimport:

mysqlimport -u [USER] -h [HOST] -p CREDO [FILE]

Checking out the source

The Mercurial repository resides at http://bitbucket.org/aschreyer/credoscript.

Check out (clone) the current development trunk (Edge Credo API) with:

$ hg clone http://bitbucket.org/aschreyer/credoscript/ credoscript

The Credo API can be imported as follows (assuming the Credo.py file is in your $PYTHONPATH):

from credoscript import *

Software used to create the CREDO database

• OpenEye toolkits
• RDKit: Cheminformatics and Machine Learning Software
• Open Babel: The Open Source Chemistry Toolbox

CREDO license

The Credo Database and Application Programming Interface by Adrian Schreyer is licensed under a Creative Commons Attribution-Share Alike 2.0 UK: England & Wales License.