Szybki TK 1.8.6¶

Major bug fixes¶

A memory issue that caused periodic crash of JVM in a situation when torsion optimization was performed and Coulomb terms were removed from the MMFF94 or MMFF94S force field in Java-wrapped code has been fixed.
Possible memory corruption in the IEFF force field has been fixed.
Previously, making the protein partially flexible and selecting a run type of either OERunType::SinglePoint or OERunType::TorsionsOpt had been causing crashes. These bugs have both been fixed.
The order of the residue names generated in the logging output for flexible polar hydrogens for the Protein-Ligand IEFF-MMFF calculation will now be consistent between runs. Previously, the order of the residue names could change in subsequent runs.
A discontinuity in the flat-bottom harmonic potential has been fixed. The potential function generated by the method OESzybki::SetHarmonicConstraints had a discontinuity in the primary function and its first derivative that caused the optimizer to over-constrain optimizations with harmonic constraints with flat-bottom. The class OEHarmonicPotential has been reimplemented.

Language in the warning regarding the input of atoms with unspecified chirality is improved. It now reads: Input molecule contains chiral atoms with unspecified chirality. This warning may be generated when using the FreeForm part of the Szybki TK.
OEEstimateConfFreeEnergies now accepts an input ensemble containing a molecule with 20 or more rotatable bonds. Previously, encountering such a molecule resulted in a warning and then that molecule was skipped.

Documentation for constants describing optimization in torsional coordinates and translational-rotational coordinates (OERunType::TorsionsOpt and OERunType::SolidBodyOpt has been expanded. Specifically, short descriptions of the terms torsional coordinates, and solid-body optimization have been added.
Missing documentation for the OEOptType::NEWTON constant, which sets Newton type optimization, has been added.