PDBlib v2.0 Available

Phil Bourne system at CUHHCA.HHMI.COLUMBIA.EDU
Wed Jul 27 17:09:33 EST 1994


                                PDBlib v2.0

PDBlib is a C++ class library for representing a macromolecular structure
at the level of detail parsable from a PDB file, although the 
implementation is designed to be independent of the PDB and thus data
can be loaded from a variety of sources. 

The design of PDBlib v1.0 and a couple of applications that use the library
are described in Chang, Shindyalov, Pu, and Bourne, CABIOS 10:5, 1-12.
Version 1.0 has been available since late 1993.

PDBlib 2.0 has been tested on SunOS with AT&T CC 3.0 & Objectstore CC 2.0 
and SGI with AT&T CC 3.0. Source code and PDBview, a rendering tool
that uses the library, are available via anonymous ftp from
cuhhca.hhmi.columbia.edu as the compressed tar file 
/pub/programs/PDB/PDBlib/PDBLib.2.0.tar.Z

The development of this library was supported by NSF grant IRI-9116798.

New features in PDBlib V2.0 not described in the CABIOS paper are:

1. Template classes.

Most of the classes in PDBlib release 1.0 that behave similarly have been 
re-implemented as template classes.

These template classes include:
	
	List classes:
	Iterator classes:
	Vector & Matrix classes:
	 
2. Copy constructor.

Copy Constructors have been added to all the macromolecular classes in the 
library. Users can apply a copy construction to a macromolecular object to
generate a new object with identical information to the original one. These 
operations have been used in our object-oriented database project to 
create persistent macromolecular objects. By optionally supplying 
a rotation matrix and a transformation vector these copy constructors are 
also used to generate crystallographic symmetry (see below). 

3. Symmetry.

Space group and unit cell information are included.
A SpaceGroup object which contains rotation and transformation information and
number of symmetric units can be retrieved by invoking the method 
MySpaceGroup() at a Compound (protein) level. Similarly, symmetry related
molecules can be generated by invoking MyCrystalSymmetricUnit().

4. Data members have been moved to private or protected sections.

Most of the data members that were declared as public in PDBLib 1.0 have been 
moved to private or protected sections. Inlined functions are provided to access
these data efficiently.


5. Rooted from single class

In PDBLib 2.0 all macromolecular classes are rooted from a single class. This 
modification will be useful when one wants to develop an application that is
mutually applicable to all macromolecular classes.


6. PdbInfo become part_of a Compound rather than is_a.

PDBInfo (information such as header, remark...) have been changed to as
an object that is "part_of" a Compound object. This the first step toward
integration of CIF information.

7. Memory leaks.

PDBLib 2.0 has been purified. All observable memory leaks were corrected.


9. Improved pdb file import function.

The PdbFile.Import() function has been improved to be more robust. Over 98% of
the PDB files (Jan. 1994 release) can be successfully converted into a memory
resident object representation. 



More information about the Biomatrx mailing list