Version 2.0.4¶

OEChem 2.0.4¶

OEReadMolecule will now correctly handle right-justified atomic symbols in MDL molfiles. These files are being generated by the RCSB PDB custom SDF file writer.
OEPerceiveCIPStereo, the Cahn-Ingold-Prelog stereochemistry perception algorithm, has been rewritten to solve the following issues:
- Performance issues for large molecules have been resolved.
- Incorrect chirality perception for certain isotopic molecules, detailed in the blog post R or S? Let’s Vote, has been corrected.
- Incorrect chirality perception for stereocenters with both a hydrogen and either a deuterium or tritium attached has been fixed.
- Incorrect chirality perception for stereocenters with both a hydrogen and a *atom (atomic number 0) has been fixed.

oemolistream.Setgz with a false argument called after a stream has finished reading molecules will no longer cause OEReadMolecule to crash on subsequent calls. It will now return false to indicate that the stream has finished.
OEMolDatabase.Open will now fail and throw a warning message if there is no disk space left for uncompressing compressed input files. In earlier releases, the operation would appear to succeed and lead to corrupted molecules past a certain point. Furthermore, the directory that OEMolDatabase uses for temporary files can now be controlled through the TEMP or TMP environment variables on Windows or the TMPDIR environment variable on Linux and OSX.
OEMol will now work with the V3000 file format much more reliably. This was accomplished by making sure that multi-conformer molecules can handle groups appropriately. The OEGroupBase class is a container of atom and bond pointers inside an OEMCMolBase. The OEGroupBase class is currently used to store MDL-enhanced stereo information. Each conformer of the multi-conformer molecule has its own separate container. Now methods such as OEMolBase.GetGroups will provide access to the groups of the active conformation.

OEFormat_OEB format reading will no longer cause a rare crash when dealing with large amounts of SD data, PDB data, or bond integer types.
OEMatchBase now checks parameters to the overloaded versions of the OESubsetMol function to ensure that the given match corresponds to source and destination molecules.
OEMolBase.NewGroup method signatures now take const std::vector of atom and/or bond pointers. Previously, these arguments were not const even though the vectors were not altered.
OEReadMolecule will now perform chirality perception for 0D MDL molfiles to ensure that OEAtomBase.IsChiral and OEBondBase.IsChiral are correct.
OEGetAtomicNum is now tolerant of leading spaces for the string argument passed in.

OEBinaryIOHandlerBase had accidentally been removed from Java in the 2014.Oct release. This broke the older idiom for preserving rotor-offset compression in OEB files using OEInitDefaultHandler. The preferred API is to now use OEPreserveRotCompress directly on the molecule streams, avoiding the need to access the OEBinaryIOHandlerBase internals entirely.
The OEUnaryRoleSetPredicateFunc function was renamed to OEUnaryRoleSetBoolFunc for the sake of consistency.
OERoleSet predicates can now be created and customized in the same way as atom and bond predicates. OERoleSet predicates are an important part of customizing the new OEBio TK Splitting Macromolecular Complexes feature.
OERoles class has been added as a concrete implementation of the OERoleSet abstract base class.

An API has been added for classifying and separating the various components of a macromolecular complex by their functional roles. See the Splitting Macromolecular Complexes section for more information.
- OESplitMolComplex function
- OEGetMolComplexComponents function
- OECountMolComplexSites function
- OESplitMolComplexOptions class
- OEMolComplexCategorizer class
- OEResidueCategoryData class
- OEMolComplexFilter class
OEGetResidues function has been added to allow iteration over all unique OEResidue objects in an OEMolBase.

Protein Preparation chapter has been added to document the APIs for splitting macromolecular complexes into their components.

OEThrow.SetLevel and OEThrow.GetLevel race conditions should be largely alleviated now. OEErrorHandler.SetLevel and OEErrorHandler.GetLevel will now affect the error message level for the current thread only; all new threads launch with the default error level of OEErrorLevel_Default (Info or above).

Warning

Multi-threaded OpenEye Toolkit programs may need to be updated to allow the desired output to be sent to standard error. However, the previous behavior was race-condition prone and difficult to get correct.

OERandom constructor will no longer take an unsigned int as input. Previously, the unsigned int would be implicitly cast to bool and would control whether to time seed the constructor. However, most users actually intend to use an unsigned int as a random number seed meant for the OERandom.Seed method. This change should catch this common mistake at compile time.

OEMath Functions chapter now contains more detailed descriptions of the functions.
Documentation has been added for the following classes:

OEFileTempPath will now use the directory specified by the TMPDIR environment variable on Linux and OSX. This allows the user to control where OpenEye Toolkits will create temporary files. This is especially useful since OEMolDatabase can use a lot of temporary storage when dealing with compressed data files.

**Example of the New 2D Ring Templates**¶
Pure Carbon Skeleton Ring Templates	Ring Templates with Specified CIS/Trans Bonds