OEToolkits 2020.2.2

The OEToolkits 2020.2.2 is a bug-fix of the OEToolkits 2020.2 release. Click on the below links for more information about the bug fixes in this release:

Release Highlights 2020.2

Omega TK: New fragment library

A new enhanced fragment library has been created for use in Omega TK. The new fragment library consists of more than 500,000 fragments that have been generated using the MMFF force field.

This new library is designed to reduce the run time of large-scale conformer generation, and therefore is not built in to Omega TK but is available as a separate download. To run Omega TK with the enhanced library, simply use the OEOmega method OEOmega.AddFragLib.

The fragments have been collated from commonly used electronic databases, such as Enamine Building Blocks. Consequently, dramatic runtime performance improvements can be seen for the Enamine Real dataset, among others, since less time is spent building fragments on the fly. Runtime performance improvements can be seen for all datasets; the Platinum and Enamine Real sets show particular improvement and run ~50% faster (benchmarks run on Ubuntu 16 with Intel(R) Xeon(R) Gold 6128 CPU 3.40GHz and NVIDIA Tesla V100 GPU).

../_images/Omega_FragLib_Runtime_Perf.png

GPU-OMEGA runtime performance using the built-in fragment library compared to the new enhanced fragment library. Benchmark datasets are a filtered set of 2359 molecules of the Platinum dataset and 10,000 randomly chosen molecules of the Enamine Real, Mcule and WuXi databases.

The enhanced fragment library generates the same high-quality conformers as the built-in library, as shown by the minimum RMSDs of the filtered Platinum dataset.

../_images/Omega_FragLib_Plat_Div_Perf.png

Box plots showing minimum RMSDs of a filtered set of 2359 molecules of the Platinum dataset with the built-in fragment library and the new enhanced fragment library.

Datasets can be downloaded from:

We are eager to get your feedback on the new enhanced fragment library. Please share your experience with us at support@eyesopen.com.

OEDocking TK: Improved receptors

Receptors used in OEDocking TK classes, including FRED, HYBRID, and POSIT, have been improved to take advantage of properly prepared structures from SPRUCE. A new OEReceptor object has been introduced and is created and contained within an OEDesignUnit. Since an OEReceptor is now an integral part of an OEDesignUnit, it is saved into an OEDU file along with a design unit. There is no separate I/O for the receptors.

../_images/2IKO_receptor.png

OEDocking receptor, along with the target structure and the bound ligand, for the 2IKO Human Renin Complexed with Inhibitor.

Having the receptors as part of the design unit with properly prepped structures makes it easier for the docked and posed structures to be used in further downstream modeling, which is especially necessary when working with protein force fields like FF14SB. The new receptors with properly prepped protein also enabled use of modern force fields like FF14SB and Parsley for structure optimization in flexible OEPosit.

Szybki TK: A new protein force field

The AMBER FF14SB protein force field has been implemented in OEFF TK. This new force field has also been made available in SZYBKI for optimizing both protein and protein-ligand complexes.

FF14SB is the most widely used protein force field for molecular dynamics or any other force field-based calculations. It is also used in Orion molecular dynamics package, and is the community gold standard for such calculations.

Supported Platforms

Package

Versions

Linux

Windows

macOS

Python

3.6, 3.7, 3.8

RHEL7/8, Ubuntu18/20

Win10

10.13, 10.14, 10.15

C++

RHEL7/8, Ubuntu18/20

Win10 (VS2017, VS2019)

10.13, 10.14, 10.15

Java

1.8, 11

RHEL7/8, Ubuntu18/20

Win10

10.13, 10.14, 10.15

C#

Win10 (VS2015, VS2017, VS2019)

General Notices

  • Support for Ubuntu20 has been added. Ubuntu16 is no longer supported.

  • C++ CUDA-enabled toolkits and features are now supported on RHEL8 and Ubuntu18.

  • This is the last release to support GCC 4.x and GCC 5.x. GCC 7.4 will be the minimum GCC supported in the next release.

  • This is the last release to support Python 3.6.

  • This is the last release to support VS2015.