OEToolkits 2025.1

OEToolkits 2025.1

Release Highlights 2025.1

FastROCS: Grid Query Support

The 2025.1 release adds support for grid shape queries in FastROCS TK, as is already available in the Shape TK and ROCS. The ability to use shape grids substantially expands the application of FastROCS TK to queries beyond those derived from molecules.

To assess the efficacy of grid-derived versus molecule-derived queries ROCS was run with a series of thirteen molecule-derived queries, as well as the corresponding grid-derived queries, against a database of 1M molecules and the top-ranked 10K hitlist for each of the 13 was retained. FastROCS was then run with the same molecule-derived and grid-derived queries, and the fraction of the top-ranked 10K hits in common with the reference ROCS hitlist was calculated (see Figure 1). The fraction of molecules in common is almost completely constant (compare the black and pink bars in Figure 1), whether FastROCS was used with a molecule-derived query or the corresponding grid-derived query.

../_images/ROCS_vs_FastROCS_shapeonly.png

Figure 1. Thirteen query molecules and their shape grids (left). A graph of ROCS versus FastROCS overlap in finding the same molecules in the hit list (right).

Bioisostere TK: Feature Extensions

New functionality has been added to Bioisostere TK, bringing all features of the BROOD application into the toolkit. Some of the major new features in the toolkit include the ability to build queries in the presence of the target protein binding site and/or a selectivity protein, as well as creating bridging queries that connect two molecules. New APIs have also been added to enable checking whether compounds similar to the analogs generated by BROOD exist in commercial or internal databases, allowing a chemist to consider molecules available for purchase before undertaking the synthesis of a given analog.

With all of the BROOD functionality now available in Bioisostere TK, BROOD has been reengineered with Bioisostere TK to simplify and streamline coordination with the toolkit, ensuring that consistent results are provided between BROOD and Bioisostere TK. The graphical application vBROOD has been also been rewired to the newly engineered BROOD.

Representation of both active site and selectivity proteins in Bioisostere TK, BROOD, and vBROOD has been modified to work with the OEDesignUnit.

../_images/brood_query_in_active_site.png

Figure 2. Building a BROOD query in a protein active site.

SZYBKI: Robust Protein–Ligand Optimization

The robustness of ligand optimization in a protein active site has been improved in both Szybki TK and SZYBKI.

Optimization failures due to non-planarity in aromatic rings of optimized structures are now separated from those due to excess atomic overlap in the starting structures, allowing better handling of failed optimizations. Failures due to a low level of aromatic ring bending, which is a common artifact of force-field-based optimization, can now be ignored, producing a substantial increase in successful ligand optimizations. The criterion for flagging an optimized structure as having bent aromatic rings has also been modified. Overall, the revised algorithm allows more structures to pass through the minimization process, while still giving the user the ability to flag results with deformed aromatic rings.

Besides these algorithmic modifications, there are improvements in integrating optimization of POSIT generated pose and in assigning AM1BCCELF10 charges to ligands, with SZYBKI.

Supported Platforms

Package

Versions

Linux

Windows

macOS

Python

3.10 - 3.12

RHEL8/9, Ubuntu22/22-ARM/24

Win10/11

13, 14, 15

C++

RHEL8/9, Ubuntu22/22-ARM/24

Win10/11 (VS2022)

13, 14, 15

Java

8, 11, 20

RHEL8/9, Ubuntu22/22-ARM/24

Win10/11

13, 14, 15

C#

Win10/11 (VS2022)

General Notices

  • Support for Ubuntu 24.04 has been extended for all applications except VIDA. Support for VIDA will be added in a later release. Support for Ubuntu 24.04 for C++, Python, and Java toolkits was already added in the previous release. Support for Ubuntu 20.04 has been discontinued from this release.

  • Support for macOS 15 has been added for C++, Python, and Java toolkits, as well as applications and VIDA. Support for Ubuntu 20.04 has been discontinued from this release.

  • Support for Python 3.13 is not available in this release but will be added in a future release. Support for Python 3.9 has been discontinued from this release.