
.. _section_szybki_releasenotes_183:

Szybki TK 1.8.3
================================================================================

New features
--------------------------------------------------------------------------------

.. oe:language:: c++

   * The C++ **Szybki TK** library, ``oeszybki``, now links against the
     ``oeieff`` library. Please update build scripts accordingly to
     account for the new dependency.

* New forcefield for ligand-ligand and protein-ligand interactions has been added. 
  It is a combination of MMFF94S (or MMFF94) with intermolecular potential called
  IEFF developed at OpenEye. In this combined force field, MMFF94S is used
  to describe the intramolecular interactions while IEFF which has 3 components
  (vdW, Coulomb and Fermi repulsion) handles intermolecular interactions.
  Coulomb terms in IEFF require precalculated and assigned atomic multipoles
  to ligand and protein molecules. This task can be achieved with two OpenEye codes 
  called MOLQM and PROTQM. IEFF has been described in the following publication::

     N. Hamaguchi, L. Fusti-Molnar and S. Wlodek
     Force-field and quantum-mechanical binding study of selected SAMPL3 host-guest complexes
     J. Comput. Aided Mol. Des. Vol. 26, pp. 577-582, 2012

  Please contact support@eyesopen.com if you are interested in testing
  this new forcefield.

* New overloaded method :oe:meth:`OESz::OESzybki::SetTorsionConstraint` has been
  added to the public API. It allows selectively constraining a specific 
  torsion in the optimized molecule.

* A set of flexible residues in ligand optimization with partially flexible protein
  is no longer restricted to the specified distance from the ligand. This goal has
  been achieved by the expansion of the :oe:const:`OESz::OEProtFlex` namespace for 
  new constants which might be passed to the 
  ``OESz::OESzybkiOptions::SetProteinFlexibilityType`` method, and the addition
  of the new method ``OESz::OESzybkiOptions::AddFlexibleResidue`` which can
  be used multiple times to make a set of residues flexible.

* Added new virtual method ``OEOpt::OEFunc1::IsRedundant``. It allows
  to check if two formally numerically different sets of coordinates indeed
  represent two different points in the coordinate space. This method has
  no trivial implementation in :oe:meth:`OEMolPotential::OEQuatAdaptor::IsRedundant`.    

* MMFF calculations now allow for a "DUMMY" atom type,
  :oe:const:`OEChem::OEMMFFType::DUMMY` to specify that an atom should
  be avoided during an MMFF energy calculation.

* Added a DUMMY atom type to the AMBER force field to specify that an
  atom should be avoided during an AMBER energy calculation.

Major bug fixes
--------------------------------------------------------------------------------

* Fixed a rare crash when using the AMBER force field on some platforms.

* A combination of Newton optimization with the use of the following two methods:
  ``OESz::OESzybkiOptions::SetProteinElectrostaticModel`` and
  ``OESz::OESzybkiOptions::SetExactVdWProteinLigand`` with parameters 
  :oe:const:`OESz::OEProteinElectrostatics::ExactCoulomb` and ``true`` respectively
  caused that Coulomb potential was missing from the intermolecular interaction.
  That bug has been fixed.

* :oe:meth:`OESz::OESzybkiResults::GetLnQvib` and
  :oe:meth:`OESz::OESingleConfResult::GetLnQvibrational` will no
  longer return negative infinity on Windows for large ligands. 

Minor bug fixes
--------------------------------------------------------------------------------

* Fixed the name of the saved electrostatic grid file when the name of the molecule
  is used as part of the file name, by replacing all non alpha-numeric characters with 
  an underscore character.

* An attempt to use Newton optimization with a potential for which second derivatives
  are not implemented is terminated with a proper error message.

* When AM1BCC charges are selected in Sheffield or PB solvation models for
  multiply charged species, SCF procedure in AM1 might not converge within 
  20 cycles (this has been observed for methotrexate only so far). To make 
  sure that the electron density is converged, and calculated atomic 
  charges stable, we allow up to 1000 SCF iterations for multiply charged
  molecules.

* Return value of the method :oe:meth:`OESz::OESzybkiResults::GetTotalEnergy`
  for protein-ligand system with partially flexible protein does not contain
  intrinsic MMFF energy of the protein. 

* An attempt to use protein flexibility type which is not a member of the
  :oe:const:`OESz::OEProtFlex` namespace in the method 
  ``OESz::OESzybkiOptions::SetProteinFlexibilityType` issues an
  error message ``Unrecognized protein flexibility``.

* An error message generated upon the attempt of using partially flexible protein
  in solid body optimization type is generated instead of warning. 

* A more informative error message is generated upon the attempt of free ligand
  entropy estimation in the situation when protein environment is set.

* An error message is generated when trying to use Newton optimization for
  protein-ligand systems with protein electrostatic models different than
  :oe:const:`OESz::OEProteinElectrostatics::NoElectrostatics` or
  :oe:const:`OESz::OEProteinElectrostatics::ExactCoulomb`

* Flexible polar hydrogens are printed when residue polar hydrogens are asked to be
  optimized along with the ligand.
  
Documentation fixes
--------------------------------------------------------------------------------

* Documentation for the ``OEMolMMFF::OEMMFFSubsetVdw`` class has been removed. 







