基尼而絲(G.N.A.S)的大世界

Theory - Conformations

There are two closely related and often confusing concepts about molecular 3D structures: conformations and conformers. A conformer of a molecule is a local or global energy minimum structure, while a conformation is any 3D structure of a molecule, which may or may not be an energy minimum. In most applications, such as pharmacophore modeling, the low energy conformations of molecules are used. In Discovery Studio, you can use the Search Small Molecule Conformations tools and the Generate Conformations protocol to generate conformations. The tool panel provides an easy and intuitive way for quick conformation generation and visual inspection of the conformation models. The protocol, on the other hand, provides more advanced controls for conformation search, such as energy minimization in solvation.

Six conformation search methods are available in the Generate Conformations protocol:

• FAST - Quickly provides diverse low-energy conformations. For details, see FAST conformation generation.
• BEST - Ensures the best coverage of conformational space (requires more CPU time). For details, see BEST conformation generation.
• CAESAR - Very quickly provides conformations by sampling torsions. For details, see CAESAR conformation generation.
• Systematic Search - Systematically search on torsion grids of user defined rotatable bonds.
• Random Search - Randomly change torsion angles of user selected rotatable bonds
• Boltzmann Jump - A stochastic search method where random changes in the specified torsion angles are accepted or rejected according to the Metropolis selection criterion.

Note. For details about Systematic, Random, and Boltzmann jump methods, see Torsion search methods.

Search Small Molecule Conformations tools

The Search Small Molecule Conformations tool panel provides a quick way to generate conformations by searching the torsion space of user-defined rotatable bonds.

To display the Search Small Molecule Conformations tool panel

Choose the View | Tool Panels | Search Small Molecule Conformations command from the menu bar.

Conformation Search Setup

Find Rotatable Bonds: Adds Rotatable Bond Attributes to the selected bonds for each molecule, or all rotatable bonds if nothing is selected. During the conformation generation, the torsions will only be varied for bonds that have these attributes.

Conformation Method: Allows you to select the conformation generation method to use: Systematic, Random, or Boltzmann Jump. For details about Systematic, Random, and Boltzmann jump methods, see Torsion search methods.

Energy Threshold: Allows you to select the maximum energy threshold for the final conformations. The value can also be changed using the Energy Threshold property found in the Molecule tab of the Data Table View.

Torsion Kick: Alters the dihedral angles of all of the selected bonds (apart from terminal bonds and bonds in rings) by a random increment. The increment does not exceed +/- 45°.

Note. Torsion Kick ignores terminal bonds and ring bonds. A terminal bond is a bond where at least one of the connected atoms has only one bond to other atoms. A ring bond is a bond that connects two atoms in a ring. For two atoms to be part of a ring, a bond path must exist that connects them after the bond between the two atoms has been removed.

Conformation Search and Minimization

Conformation Generation: Performs the conformation generation according to the conformation setup parameters.

Dreiding Minimization: Applies the Dreiding forcefield with a steepest descent minimizer to optimize all or a selected portion of the current model including conformations. It can be used to both create a good starting conformation and to minimize the conformation candidates. The minimization is constrained by settings specified in the Dreiding Minimize page of the Preference dialog.

Note. The Dreiding forcefield does not provide parameters for all elements. If it is not possible to assign a forcefield type. Molecule and atom properties can be accessed from the respective tabs in the Data Table View.

CAESAR conformation generation

Conformer Algorithm based on Energy Screening And Recursive build-up (CAESAR), is based on a divide-and-conquer and recursive conformation build up approach. This is also combined with consideration of local rotational symmetry so that conformation duplicates due to topological symmetry in a systematic search can be efficiently eliminated. The CAESAR algorithm can be represented schematically as follows:

Figure 1: Schematic presentation of CAESAR algorithm

A molecular tree is recursively partitioned into tree nodes connected by tree edges. Each tree node is either a ring fragment or a rigid structure, and each tree edge represents a rotatable bond. The conformation generation starts from tree node conformation initialization. For the ring nodes, the same ring fragment conformation generation method used in FAST is used in CAESAR. The conformations of the whole molecule are built up recursively from the smallest fragments. At each level of recursive assembling, local rotational symmetry check and energy pruning are performed. The symmetry check removes non-unique torsion increments so that duplicate conformations are effectively eliminated during conformation assembling. The high-energy conformations can be eliminated at the earliest possible stages by using energy filtering at each level of conformation build-up.

Further reading

[Li et al., 2007].

 [Li et al., 2007] Li, J.; Ehlers, T.; Sutter, J.; Varma-OBrien, S.; Kirchmair, J. CAESAR: a new conformer generation algorithm based on recursive build-up and local rotational symmetry consideration. J. Chem. Inf. Model 2007, 47, 1923-1932

http://pubs.acs.org/doi/abs/10.1021/ci700136x

FAST conformation generation

The FAST conformation generation method uses one of three algorithms, depending on the size of the molecule. The conformational space of small molecules is generated using an efficient systematic search. If the molecule is too large, only one conformation is generated for each possible combination of stereocenters. Conformations for molecules that are neither too small nor too large, as measured by the flexibility of the molecule, are generated with a random search method that uses poling. The following is covered here:

• FAST conformation generation for small molecules
• FAST conformation generation for medium-sized molecules
• FAST conformation generation for large molecules

FAST conformation generation for small molecules

A quasi-exhaustive systematic search is used to generate conformations for small molecules. The conformational space is composed of discretized rotations about the rotatable bonds. The following table lists offsets and search increments for the hybridization types:

The conformational space is systematically searched and conformations that have conformation energy above an energy threshold are removed.

FAST conformation generation for medium-sized molecules

A random search method with poling is used to generate conformations for medium-sized molecules if the estimated number of conformations is greater than a threshold. This threshold is 750 by default, but can be modified via a Catalyst.cfg file. To generation conformations, the molecule is split into pieces, a systematic search is performed on each piece, and the pieces are randomly reconnected. Each new conformation is quickly optimized in the torsion space with poling penalty in order to maintain conformation diversity.

FAST conformation generation for large molecules

If the number of rotatable bonds is greater than 30, only one conformation is created for every possible stereocenter.

BEST conformation generation

In general, the best-quality conformation generation method provides improved conformational coverage relative to the fast method. It does this by performing a more rigorous energy minimization in both torsional and Cartesian space and by using poling. Use the best-quality method when conformational coverage is important (e.g., to build conformation models for automatic hypothesis-generation routines).

The following steps are performed by the BEST routine:

• Conjugate-gradient minimization in torsion space
• Conjugate-gradient minimization in Cartesian space
• Quasi-Newton minimization in Cartesian space

 HIPHOP構型理論