.. _highlights_2022_Spring:

******************************
Release Highlights 2022.1
******************************


Enhanced Stereo Information Support
====================================

The 2022.1 release adds support for enhanced stereochemistry designations in molecules, 
defining relationships between stereocenters. Enhanced stereochemistry
information on any molecule is honored by :oe:meth:`OEFlipper<OEConfGen::OEFlipper>`
during stereo enumeration when asked to do so by setting appropriate options. Enhanced 
stereochemistry support has also been extended to both the :ref:`OMEGA <chapter_omega>` and
the :ref:`chapter_flipper` applications.

Support for :term:`CXSMILES` as a valid molecular input format via
:oe:const:`OEChem::OEFormat::CXSMILES` has been added.  :term:`CXSMILES` is specifically 
intended to support the import of `Enamine <https://enamine.net>`_ download structure files 
containing enhanced stereochemistry designations. While many other features are possibly 
provided in the :term:`CXSMILES` appendix information, only the enhanced stereochemistry group 
information is supported at this time.   Please contact support@eyesopen.com with specific needs 
for additional :term:`CXSMILES` features to allow prioritization in future releases.

**OMEGA** and **Flipper** applications can now accept molecule files in :term:`CXSMILES` format.

.. figure:: images/2022-Spr/enhanced_stereo.png
   :width: 200pt
   :align: center

   **Molecule with enhanced stereo designation. The molecule has 16 stereoisomers when
   stereocenters are treated independently and only 4 stereoisomers when the
   enhanced stereo is enabled.**


Sheffield Solvation Model for Proteins
=======================================

The simple two parameter Sheffield Solvation Model, as introduced by Grant et al. (\ [Grant-2007]_ \)
for calculating the electrostatic component of molecular solvation energy, is a semi-empirical model
that performs well for small molecules and requires few computational resources. This model does
not work well for macromolecules like proteins and complexes, where many of the atoms are hidden or
only partially exposed to the solvent.

A modified three parameter Sheffield Solvation Model is parameterized and implemented in this release,
where solvent exposure is treated by correcting the effective atomic radii with the introduction of
a third additional parameter. The new model is available as part of the 
:oe:cls:`OESheff::OESheffield` API in the **OEFF toolkit**. The model is also available for use
through the :oe:cls:`OESz::OEFixedProteinLigandOptimizer` and the 
:oe:cls:`OESz::OEFlexProteinLigandOptimizer` APIs in the **SZYBKI toolkit**, and the :ref:`OptLigandInDU`
and :ref:`OptimizeDU` applications in **SZYBKI**.

Electrostatic solvation energies for protein conformers are calculated using both the new Sheffield
Solvation Model and the Poisson–Boltzmann (PB) model, and the difference in solvation energies between 
conformers of the same protein is compared between the two models. A set of 232 NMR proteins and peptide 
structures were used for this study that produced 41491 solvation energy differences between pairs of 
conformations. The results are compared in the image below.

.. figure:: images/2022-Spr/protein_confs_sheffield_pb.png
   :width: 600pt
   :align: center

   **Comparison between Sheffield and PB of differential electrostatic solvation energies between protein conformers.**


Spruce Filter and Design Unit Validation
=========================================

New functionality for preprocessing inputs and validation of the
output ``Design Units`` has been added to **Spruce TK** in the form of two preliminary API classes :oe:cls:`OESpruce::OESpruceFilter` and 
:oe:cls:`OESpruce::OEValidateDesignUnit`.
These classes are also used by the **SPRUCE** application to check and fix, if possible, inputs and to validate the resulting outputs.

:oe:cls:`OESpruce::OESpruceFilter` performs a standardization and a filter check structures to use before prepping
this structure in **SPRUCE**. It automatically checks and fixes invalid component names, invalid residue 
states,and  invalid metal bonds. It also checks for invalid CRYST1 from the header file, low overlap between 
the electron density map and the structure, invalid sequence alignment, blank chain IDs, and invalid
alternate locations. If the structure does not pass these tests, Spruce Filter will fail the structure.

The :oe:cls:`OESpruce::OEValidateDesignUnit` function performs a check on the design units generated by 
**SPRUCE**. It will report if the provided DU contains partial side chains, missing loops, implicit 
hydrogens, incorrect partial charges, termini that are not capped properly, heavy atoms overlapping 
with each other, and broken chains.


Multistate Heuristic pKa Model
===============================

The multistate heuristic pKa model is an extension of the existing pKa model in **OpenEye** software. 
The existing model generates only one probable ionization state based on pKa estimated from 
about 80 rules. With the development of this multistate pKa model, we are enhancing the 
implementation with many more unique functional patterns in the form of pKa rules, identified 
from OpenEye's internal pKa database, allowing generation of all ionization states possible 
at physiological pH.

The model described herein uses SMARTS-based heuristics to assess the pKa of acidic or basic
functional groups of molecules. For all atoms of a query molecule, pKa is assessed as it belongs 
to one of three classes: Acidic (pKa < 6.4), Basic (pKa > 8.4), and Neutral (6.4 < pKa < 8.4). 
According to the assessed pKa class for each atom, it is either ionized and/or un-ionized at
pH 7.4. Any group with Acidic or Basic pKa range will result in one state, ionized or un-ionized 
and a group with Neutral pKa range will result in both states, ionized and un-ionized. Based 
on how many such functional groups there are in the molecule, and how many states are generated 
for them, multiple molecules can be generated for all possible combinations of all possible 
ionization states of acid/base groups.

The multistate heuristic pKa model is currently only available in **Quacpac TK** as preliminary
API classes :oe:cls:`OEProton::OEMultistatepKaModel` and :oe:cls:`OEProton::OEMultistatepKaModelOptions`.


Supported Platforms
*******************

  ===========  ===================  ===========================   ===============================  ====================
  Package      Versions             Linux                         Windows                          macOS
  -----------  -------------------  ---------------------------   -------------------------------  --------------------
  Python       3.7, 3.8, 3.9, 3.10  RHEL7/8, Ubuntu18/20/20-ARM   Win10/11                         10.15, 11, 12
  C++                               RHEL7/8, Ubuntu18/20/20-ARM   Win10/11 (VS2017/19/22)          10.15, 11, 12
  Java         1.8, 11              RHEL7/8, Ubuntu18/20/20-ARM   Win10/11                         10.15, 11, 12
  C#                                                              Win10/11 (VS2017/19/22)
  ===========  ===================  ===========================   ===============================  ====================

Related TK Versions
*******************

  ==============  ===========
  Toolkit         Version
  --------------  -----------
  OEChem TK       3.3.0
  OEDepict TK     2.5.0
  OEDocking TK    4.1.2
  FastROCS TK     2.2.2
  OEFF TK         2.5.0
  Grapheme TK     1.4.5
  Graphsim TK     2.5.4
  Lexichem TK     2.7.5
  OEMedChem TK    1.1.6
  MolProp TK      2.6.0
  Omega TK        4.2.0
  Quacpac TK      2.2.0
  Shape TK        3.5.0
  Spicoli TK      1.5.4
  Spruce TK       1.5.0
  Szmap TK        1.6.4
  Szybki TK       2.5.0
  Zap TK          2.4.3
  ==============  ===========

General Notices
***************

* Support for macOS 12 has been added. macOS 10.14 is no longer supported.

* This is the last release to support macOS 10.15.

* A GCC 7.3 C++ toolkit package has been added for RHEL7.

* Support for Windows 11 has been added.

* Support for Visual Studio 2022 has been added.

* Support for Apple M1 architecture has been added for C++ and Python toolkits.

* Support for Linux ARM architectures has been added.

* This is the last release to support Ubuntu 18.04. Support for Ubuntu 22.04 will be added in the next release.

For detailed release notes, see :ref:`Version 2022.1.0 Release Notes <release_notes_index>`.
