Conformational searches

A conformer represents a unique arrangement of the atoms of a molecule, characterized by different spatial orientations and configurations of the atoms while maintaining a particular connectivity between them. Conformers can differ in their energy levels and stability, and they can interconvert through molecular motions such as rotation, vibration, or even more complex changes.

Exploring the different conformers of a molecule and searching for the lowest energy conformer or set of conformers can be useful for several reasons. The WEASEL conformer search workflow provides a convenient way to find them. This workflow is useful in itself, but it is also the basis for several other workflows that explore the properties and reactivity of a molecule in different conformers. For example, it is possible to calculate different spectra such as NMR, IR, or UV/Vis and CD spectra for a conformer ensemble. The WEASEL conformer search workflow is also the basis for several other chemical search workflows, such as the anion and protomer searcher or the tautomer searcher, to name a few.

For now, let us focus on the basics of the conformer search workflow by taking a look at the conformer ensemble of limonene.

Limonene.

How to run the calculation

To perform the calculation yourself, you can download the file limonene.xyz providing the XYZ coordinates of the structure.

Alternatively, you can use the provided SMILES string:

C=C(C)[C@H]1CC=C(C)CC1

To initiate a conformational search, use the following command:

weasel limonene.xyz -W confsearch

Or, if you prefer using the SMILES string:

weasel -smiles 'C=C(C)[C@H]1CC=C(C)CC1' -W confsearch

Executing one of the commands initiates the conformational search workflow.

Steps of the WEASEL workflow

The WEASEL conformational search workflow consists of several filtering steps to obtain a conformer ensemble for a system. Let's walk through each step of the workflow:

Step 1: Conformer generation

This step generates a set of initial conformers for the system. The conformers are generated using CREST. The energy filter window for this step is 6.0 kcal/mol. This means that WEASEL ranks the resulting conformers by energy, and only conformers that differ by 6.0 kcal/mol or less from the lowest energy conformer are retained in the ensemble. This ensemble is passed to the next step.

Note

You can change the method used to generate the conformer ensemble by applying the keyword -conf-gen-method followed by the desired method. The available methods are READ, CREST, CRESTFF, RDKIT, CREST-RDKIT, CRESTFF-RDKIT, GOAT, GOATFF.

The next three filtering steps (step 2 to step 4) consist of a preoptimization step, followed by an optimization and a final single point DFT calculation using the same methods as in the basic workflow.

Note

You can modify the default methods to suit your needs and preferences using the here provided keywords. Note that they are different to those used to customize the basic Workflow.

Step 2: Preoptimization and clustering

Preoptimization with XTB using ORCA is performed on the ensemble with an energy filter window of 6.0 kcal/mol. This step refines the generated conformations and prepares them for further optimization. In addition, a fine clustering step is applied to group similar conformers together.

Step 3: Optimization

In this step, optimization is performed using ORCA with the r2SCAN-3c method. The energy filter window is set to 5.0 kcal/mol. The optimization ensures that the conformations reach stable energy minima.

Step 4: Single point energy Calculation

After optimization, the single point energy is calculated using the wB97X-V method with the def2-TZVP basis set with ORCA. The energy filter window is set to 4.0 kcal/mol. This calculation provides more accurate energy values for the conformers.

Note

If you want even more accurate energies, you can follow up with a single point calculation using a wavefunction based method. To invoke this additional calculation step, you can use the keyword -conf-spwf.

As mentioned above, throughout the entire workflow, all high-energy structures outside of a specified energy window are eliminated in order to retain only relevant conformers in each step. In addition, duplicates are removed to ensure that only unique conformers are retained. The table provided summarizes the default conditions associated with each filtering step.

Filters applied after each step of the conformation search together with their default values.

Step	Relative energy (kcal/mol)	Remove Identical Conformer	Remove Identical Rotamer
Ensemble generation	no	yes	yes
Preoptimization Filter	6	yes	yes
Optimization Filter	5	yes	yes
DFT SP energy Filter	4	no	no
Wavefunction SP energy Filter	3	no	no

Note

If you want to change the energy windows from step 1 to step 4, you can do so by using the following keywords followed by the desired energy in [kcal/mol]: conformer generation - -conf-gen-enrange, preoptimization - -conf-preopt-enrange, optimization: -conf-preopt-enrange, SP calculation - -conf-spdft-enrange. See also here for more information.

Step 5: Ranking of final energies

After generating the final ensemble, WEASEL evaluates the energies of its conformers and produces a summary file showing the ranking of these energies. In the following section, the summary file and the other output files will be explained in more detail.

Note

By default the conformer search workflow is performed in the solvent water. You can switch to a different solvent by using the keyword -solvent [solvent]. The list of solvents can be found here.

Output files and results

Before discussing the individual steps of the conformational search further below, let us first have a look at the results of the confsearch workflow. The calculation produces the following files:

.
├── limonene.xyz
└── limonene_ConfSearch
    ├── limonene_ConfSearch.input.xyz
    ├── limonene_ConfSearch.report
    ├── limonene_ConfSearch.results.xyz
    ├── limonene_ConfSearch.summary
    ├── limonene_confsearch.summary
    ├── BuildTopo
    │   └── limonene_confsearch_BuildTopo job files
    └── ConfSearch
        ├── CREST job files
        └── ORCA job files

The most important files and a brief description of their contents are listed in the table below.

File	Description
limonene_lowestConf.xyz	lowest-energy conformer
limonene.ensemble.xyz	conformer ensemble
limonene_confsearch.report	Output file of the WEASEL run
limonene_confsearch.summary	Summary file for the WEASEL run

The final conformer ensemble is stored in a multi-xyz file ('limonene_ConfSearch.results.xyz') along with the single point DFT calculated energy of each individual conformer.

Final ensemble of limonene.

During the filtering steps a lot of information was collected, which was stored in the summary file. There, for each filtering step, we find the energies of all conformers that have survived up to that step:

Summary file for job: weasel limonene.xyz -W confsearch
Energy [kcal/mol] / Value  Type            Calculation type    Method      Basis set   Solvent     Charge  Multiplicity    Further tags
-18540.564886              SP-Energy       SP_Filter           XTB         None        ALPB(H2O)   0       1               Conformer 21
-18540.564814              SP-Energy       SP_Filter           XTB         None        ALPB(H2O)   0       1               Conformer 1
...
-18536.471021              SP-Energy       SP_Filter           XTB         None        ALPB(H2O)   0       1               Conformer 38
-18536.064879              SP-Energy       SP_Filter           XTB         None        ALPB(H2O)   0       1               Conformer 20
-18540.566973              SP-Energy       PreOpt              XTB         None        ALPB(H2O)   0       1               Conformer 21
-18540.413787              SP-Energy       PreOpt              XTB         None        ALPB(H2O)   0       1               Conformer 22
...
-18536.473461              SP-Energy       PreOpt              XTB         None        ALPB(H2O)   0       1               Conformer 19
-18536.067621              SP-Energy       PreOpt              XTB         None        ALPB(H2O)   0       1               Conformer 20
-245080.608227             SP-Energy       Opt                 r2SCAN-3c   None        CPCM(Water) 0       1               Conformer 21
-245080.386489             SP-Energy       Opt                 r2SCAN-3c   None        CPCM(Water) 0       1               Conformer 23
...
-245074.615674             SP-Energy       Opt                 r2SCAN-3c   None        CPCM(Water) 0       1               Conformer 11
-245073.865821             SP-Energy       Opt                 r2SCAN-3c   None        CPCM(Water) 0       1               Conformer 15
-245177.283520             SP-Energy       SP_DFT              wB97X-V     def2-TZVP   CPCM(Water) 0       1               Conformer 21
...
-245175.193380             SP-Energy       SP_DFT              wB97X-V     def2-TZVP   CPCM(Water) 0       1               Conformer 19
-245174.939939             SP-Energy       SP_DFT              wB97X-V     def2-TZVP   CPCM(Water) 0       1               Conformer 26
-245177.283520             Final Energy    SP_DFT              wB97X-V     def2-TZVP   CPCM(Water) 0       1               Lowest-energy Conformer

Important

The last column of the summary file contains the ID of the conformers from the initial conformer ensemble.

Note

With each filtering step, more and more conformers are filtered out, thus fewer and fewer conformer IDs are available from step to step. The numbering of these conformers in the last column is not in ascending order, but in the order of their relative energies in the very first step.

Now let us make use of that data. The following graph shows how the relative stability of each conformer evolves with increasingly accurate methods.

Energetic distribution of limonene conformer ensemble in initial ensemble numbering scheme.

For limonene, the lowest energy conformer remains the same through all filter steps. The higher energy conformers change their relative energies more significantly.

The following figure shows the ensemble of conformers that survived to the SP_DFT step, and the relative energies of each of these conformers at each of the filtering steps. From the preoptimization step to the optimization step, the relative energies and even the order can change quite drastically. However, from the Opt step to the SP_DFT step, the results for the limonene conformers are quite similar.

Energetic distribution of limonene conformer ensemble in final ensemble numbering scheme.

Remarks and keywords

Keywords for ensemble generation

Keyword	Description
-conf-gen-method OPTION	Method for providing the initial conformer ensemble, which is then refined using the subsequent filtering steps. Available options are CREST, CRESTFF, GOAT, GOATFF, RDKIT and READ. The READ option allows the user to provide an initial conformer ensemble, e.g. from a different conformer generator. If the READ option is requested, the initial conformer ensemble has to be provided via the structure file as a multi-xyz file.
-conf-gen-maxnconf INT	Maximum number of conformers selected for the next steps. The first INT structures from the initially generated or provided conformer ensemble are considered. The remaining ones are discarded.
-conf-gen-enrange REAL	Energy filter in [kcal/mol]. Energies are computed on GFN2-xTB level. Only conformers with a relative energy of less than REAL compared to the current lowest-energy conformer are considered. The remaining ones are discarded. This is not used for the READ conformer generation method.
-conf-torsionfilter INT1 INT2 INT3 INT4	Use an additional dihedral filter on generated or provided initial conformer ensemble. Default is to not use it. If the four atoms for the definition of the torsion angle are provided via -conf-torsionfilter, the filtering step is switched on.
-conf-torsionfilter-range REAL1 REAL2	Only those conformers, for which this torsion is in the range between REAL1 and REAL2, are considered for the next steps. Torsion angles in degrees.

Keywords for preoptimization filter

Keyword	Description
-conf-preopt	Use preoptimization step. Default is true.
-conf-preopt-enrange REAL	Energy filter in [kcal/mol]. Only conformers with a relative energy of less than REAL compared to the current lowest-energy conformer are considered. The remaining ones are discarded.

Keywords for optimization filter

Keyword	Description
-conf-opt	Use optimization step. Default is true.
-conf-opt-enrange REAL	Energy filter in [kcal/mol]. Only conformers with a relative energy of less than REAL compared to the current lowest-energy conformer are considered. The remaining ones are discarded.
-conf-gibbscorrection	Run frequency calculation after optimization and use the Gibbs correction for the optimization, DFT and wavefunction Single Point filter steps. Default is false.

Keywords for DFT single point filter

Keyword	Description
-conf-spdft	Use DFT SP energy step. Default is true.
-conf-spdft-enrange REAL	Energy filter in [kcal/mol]. Only conformers with a relative energy of less than REAL compared to the current lowest-energy conformer are considered. The remaining ones are discarded.

Keywords for wavefunction single point filter

Keyword	Description
-conf-spwf	Use wavefunction SP energy step. Default is false
-conf-spwf-enrange REAL	Energy filter in [kcal/mol]. Only conformers with a relative energy of less than REAL compared to the current lowest-energy conformer are considered. The remaining ones are discarded.

The default method for the wavefunction single point filter is DLPNO-CCSD(T) with def2-TZVP basis set, and needs to be modified via the workflow file in the CONFORMATIONAL_SEARCH section:

[CONFORMATIONAL_SEARCH]
# Options: see weasel -h
SP_WF_Method = DLPNO-CCSD(T)
# Options: see [SP_WF] Basis
SP_WF_Basis = def2-TZVP

Keywords for changing the maximum number of conformers

Keyword	Description
-conf-maxnconf INT	The maximum number of conformers that is stored in the ensemble file.

Keywords for clustering conformers

Apply an agglomerative hierarchical clustering with complete linkage after the optimization step. The distance matrix is composed of the RMSDs to the lowest energy structure augmented by energy information.

If done, a folder named CLUSTERS will be created inside the Opt folder where each individual cluster can be visualized.

Keyword	Description
-conf-cluster	Apply the clustering after the optimization step.
-conf-cluster-mode OPTION	Mode of clustering. Choose between fine (default) and coarse for more compression of the data.
-conf-cluster-nmax INT	Define an arbitrary maximum number of clusters. The default is -1, meaning that it will be defined automatically.
-conf-cluster-elevel OPTION	Select when to apply the clustering, after the PreOpt step (default) or maybe only after the Opt.