Introduction

When transition state embedding is a problem, TSCoDe is the solution.

No idea how your bound structure looks like? Ask TSCoDe.

What it is

TSCoDe is a systematical conformational embedder for small molecules. It helps computational chemists build transition states approximations and binding poses precisely and in an automated way. It is thought as a tool to explore complex multimolecular conformational space fast and systematically, and yield a series of starting points for higher-level calculations.

Since its many subroutines and functionality, it also serves as a computational toolbox to automate various routine tasks, via either MM, semiempirical or DFT methods.

TSCoDe is written in pure Python. The linear algebra module required to translate, rotate, embed and compare conformational ensembles is mostly compiled just-in-time with Numba and is parallelized where possible to achieve the best possible performance and scalability. The program supports various external calculators (at least one required):

• XTB (>=6.3) (recommended)

• ORCA (>=4.2)

• Gaussian (>=9)

• MOPAC2016

What it does

Generate accurately spaced poses for bimolecular and trimolecular transition states of organic molecules. If a transition state is already in hand, the distance between reactive atoms can be specified, so as to obtain all the topologically different poses with precise molecular spacings.

TSCoDe is best suited for modelizations that involve many transition state poses and activation modes, where the combination of reagents conformations is an important aspect in transition state building.

Perform routine tasks and ensemble refinement on conformational ensembles obtained with other software or via the program itself, through different implementations of conformational search algorithms.

First, operators (if provided) are applied to input structures. Then, if more than one input file is provided and the input format conforms to some embedding algorithm (see some examples) a series of poses is created and then refined. It is also possible to perform the refinement on user-provided conformational ensembles.

How the embedding works

Combinations of conformations of transition state molecules are arranged in space using some basic modeling of atomic orbitals and a good dose of linear algebra.

Schematic representation of orbital models used for the embeddings

How the ensemble refinement works

Ensemble refinement starts with a similarity pruning, evaluated through a sequence of:

• RMSD pruning

• TFD (torsion fingerprint deviation) pruning - only for monomolecular embeds/ensembles

• Rotationally-corrected RMSD pruning - invariant for periodic rotation of locally symmetrical known groups, i.e. tBu, Ph

• MOI (moment of inertia) pruning - helps remove enantiomers and rotamers along unknown locally symmetrical groups

Extra features

Transition state searches

TSCoDe implements routines for locating transition states, both for poses generated through the program and as a standalone functionality. The SADDLE and NEB keywords and the saddle> and neb> operators are available:

• With SADDLE, a geometry optimization to the closest energetic maxima is performed on the embedded structures, using the Sella library through ASE.

• With NEB, a climbing image nudged elastic band (CI-NEB) transition state search is performed on each embedded structure. This tends to perform best with the scan> operator, where the initial minimum energy path is extracted from the distance or dihedral scan points.

• The saddle> and neb> operators work in the same way on user-provided structures.

See the operators and keywords page for more details on their usage.