The Conformational module consists of a series of algorithms designed to study the conformational space of medium-sized organic molecules.

Questions about Conformations and Energy Profiles




How do I know I have the global minimum?
The only way to be sure you have the global minimum is to do a systematic search with a very large/dense grid. For example, increase the rotation count for sp3-sp3 carbon bond rotors from the default of 3 to 10. This is known to be very inefficient, and chances are that if you have a non-trivial molecule you may not have the patience required for this calculation to complete.

Our experience with medium (drug-size) organic molecules is that the Monte-Carlo algorithm, in general, does a good job of finding the global minimum. See "How do I know my MC result is good?", and "Is the global minimum the best minima?"


Is the systematic algorithm reproducible?

Yes, if you start from the same conformation. In simple molecules the systematic algorithm is almost always inclusive of all (important) minima. As molecules become more complicated the default grid size, while likely correct in finding all classes of minima, may skip some minima. For example, the single "gauche plus-gauche minus" [-120,+120] conformation of olefins might bifurcate into multiple minima as steric bulk increases. It is likely that one of the two new minima would be found using the standard systematic approach. But 'which one' would depend on the initial conformation.

To ensure you find all minima, it is useful to increase the default rotor values. Doing so will increase computational time; an alternate approach would be to use to the Monte-Carlo algorithm.

Another example of the initial conformer affecting systematic results can be found by inspecting an 8 member carbon chain: "C1-C2-C3-C4-C5-C6-C7-C8". For illustrative purposes consider the central "C4-C5" bond as we rotate it 360 degrees; from -180 to 180 degrees. If one starts in the all "trans" conformation, we find that as we rotate the "C4-C5" bond the final conformation of 360 degrees is identical to the initial at 0 degrees. However, if the initial conformation had a number of kinks in it, we might discover that at the 120 degree mark, the C1 and C8 ran into each other. To relieve this steric problem the other dihedral angles, will relax, likely changing by more than 100 degrees and falling into new energy wells. As we continue the coordinate driving of the central C4-C5 angle to trans (180), we might find that the final conformation is not the same as the initial conformation because these other dihedrals have changed.


Is the global minimum the truth?

The global minimum is often the most interesting, and at the very least, is often representative of an equilibrium conformer found at room temperature. However, what conformation is 'best' depends on what you are looking for. There are often many other variables that conformational analysis ignores, including the effect of solvent and the quality of the energy reported at the given theory level.


How should I use QM methods with Conformation Analysis?

Because QM methods take more time than molecular mechanics (by orders of magnitude), it is usually a mistake to try conformational algorithms with QM methods. Typically, one uses the MMFF mechanics force field to generate a list of low energy conformers. This list is then resubmitted at the desired QM level as either an "equilibrium geometry" or "single point energy" calculation.. The original MMFF conformers may change geometry slightly and their relative energies will likely differ.

For small systems (1 or 2 degrees of freedom) it is sometimes possible to use conformational analysis to scan conformer space with a QM method. However, the time required, (as well as the possibility of bad initial conformers with steric problems causing the breaking and forming of bonds), require that users apply caution when applying QM methods to conformational analysis.

A further warning about "bad" initial conformers: It is highly likely that in doing a full conformer search one may start in an unfavorable position. For example, a gauche+/gauche+ conformer in propylene. At the beginning of the minimization two hydrogens may be within 0.25 Angstroms of each other. With molecular mechanics this "bad" starting position is easily handled, but with quantum chemical methods "chemistry" will occur; resulting in the breaking and reforming of bonds. You will likely not end up with the molecule you started with. Given that this is probably not what is intended, and that it will take a very long time to rearrange all atoms into a "new" molecule, you will be lucky if the job runs out of 'geometry cycles' before taking up too much computer time. The best approach is to "know" that you are starting at a good conformer. This is another reason why it is a good idea to start with MMFF conformers.


How do I tell if the Monte Carlo results are correct?
The only way to know for sure is to compare results with a complete, systematic calculation. (see How do I know I have the global minimum?.) You can build your confidence in the Monte Carlo results by restarting the search from different initial conformations. If multiple starting points yield the same "global minimum" you should have confidence that the algorithm is spanning the conformational space fairly well.


What are the details of the algorithm?

How can I modify the algorithm?
Within the 'Set Torsions' mode, you can choose the bonds and ring atoms involved in the conformational search. This is done by double-clicking on either the desired bond or atom. When a bond or atom is selected for rotation, a type-in box will appear. This box is used to enter the number of increments in the rotation. For example, if the value of 3 is entered, rotations of 0, 120 and 240 degrees will be applied. If you do not specify bonds or ring atoms, Spartan will use heuristics to decide which elements are relevant in attaining new conformers. Unless you have some chemical insight as to the relevant rotatable members, Spartan's default selections will typically provide good results. Additionally, several keywords will modify calculation details.



What do the fold numbers mean in the Set-Torsions mode?

On bonds, the fold number is simply the number of gross conformers. So for '3' there are assume 3 states, each 360/3=120 degrees apart. Thus each move is +- 120 degrees. A fold number of 2 would mean two conformers, each +- 360/2=180 degrees.

On atoms the fold number contains more specific information. An atom fold number of 3 means an "Osawa wag" is preformed. (A 4 implies a coupled osowa wag in a cyclohexane like ring.) In the case where the number contains 3 digits, the ones-digit is either 3 or 0 for Osawa wag or no wag, respectively. The hundred's digit, if present indicates that an inversion will be tried. This may occur on asymmetric, non-planar, trivalent nitrogens.

Use the keyword PRINTLEV=2 when the job is run to see more information on what precisely is being flipped and/or rotated.





Interpreting the output file:

Description of the columns:



Why does the output say it is removing molecules from the list and how is it deciding what to remove?

Any conformer that has an energy greater than WINDOW (10.0 kcal/mol) of the lowest energy conformer is thrown away. If there are more than MAXCONFS (100) conformers with acceptable energies the program will discard conformers with the goal of keeping as diverse a group as possible, while, at the same time as keeping the lowest energy conformers.


How many cycles will my molecule take? or
Will my molecule use the Systematic or Monte Carlo method?

The number of cycles a molecule will take depends on the type and number of rotatable bonds. Each rotatable bond has a fold number. (This fold number can be modified in the 'Set Torsions' mode.) For example sp3-sp3 bonds have a default fold number of 3 because these bonds usually have 3 minima 120 degrees apart. For systematic methods the number cycles is the product of all the fold numbers. For Monte Carlo methods the default number of cycles is the square of the sum of all folds. (This equation is a purely empirical formula; experience has shown it to be adequate.) One can limit this value to an upper limit using the
"Maximum Conformers Examined" field in the setup panel, and increase the value above the default using the McConfs= keyword.

The program will choose between the Systematic or Monte Carlo method by choosing the one with the fewest default conformer tries (unless the SEARCHMETHOD keyword is used to override the default).


Keywords:
Keywords specific to the conformational module
CONFANAL*Do a conformational search producing multiple results.
SCONFANAL*Do a conformational search producing only a single, lowest energy, result.
SLCONFANAL*Menu command to generate a library of conformers. Similar to combining the following keywords: SCONFANAL SEARCHMETHOD=SYSTEMATIC REPRUNECONFS=11,15,0.75
SEARCHMETHOD=
  • SYSTEMATIC to force the systematic algorithm.
  • MONTECARLO or MC to force the Monte Carlo algorithm.
  • THOROUGH to increase the default bond-rotation by a factor of three, and also set the KEEPALL keyword.
  • SPARSE Use the Systematic-sparse method searching for 2000 randomly selected conformers. See the SPARSE= keyword for more information.
  • REREAD Re-read a previous run which has deposited result in the proparc file with the SAVEINPROPARC option. This is useful in trying different pruning options (see PRUNEMETHOD) on a KEEPALL run.
SYSTEMATIC for smaller, less flexible systems.


MONTECARLO for larger more flexible systems.
CONFSKEPT=
MAXCONFS=
Set the maximum number of returned conformers to N. The program attempts to select the most diverse set representing the entire population within the energy100
WINDOW=Sets the maximum (delta) energy at which a trial conformer will be saved in the data set. Conformers with an energy greater than the current minimum energy plus this value are rejected. 10.0 (kcal/mol)
KEEPALLKeep all generated conformations. This is similar to maximizing MAXCONFS and WINDOW options, but will also keep some conformers typically thrown away because of bond strain.
MAXITER=
CONFSEXAMINED=*
Maximum number of attempted molecules. Only meaningful for the MONTECARLO method. This is controlled from the "Maximum Conformers Controlled" entry in the setup panel. A complicated function. see How many cycles...
MCCONFS= The Monte Carlo algorithm will do this number of steps, overriding the "Maximum Steps" and the default count.
FINDBOATS
Changes the default 6-member ring move (cyclohexane) to attempt to find the twist-boat conformation. This dramatically slows down the algorithm as the number of possible moves changes from 2 to 27. If FINDBOATS is used the WINDOW= value is increased to 15 kcal/mol as twist-boat conformations are typically higher in energy (~6 kcal/mol for cyclohexane). turned off by default
SKIPBOATS Uses the fast cyclohexane move of correlated flips. on by default
STARTTEMPERATURE= The initial temperature for the monte carlo/simulated-annealing algorithm. 5000 K
NORIGID Do not attempt to do any 'rigid moves' but rely on constraints to get the correct structure. This will likely slow down the algorithm.
NOOPTDo only rigid moves and single points. Do not attempt to re-minimize. (Only useful for small "dynamic constraint" systems).
IGNORENOES Ignore any NOE constraints in the system. The default is to use the NOE constraints as a filter only.
PASSNOES Pass any NOE constraints on to the optimization engine. (Currently only supported by mechanics.) Thus all intermediate results satisfy the NOE constraints. The default is to use the NOE constraints as a filter.
NOEBIASE= Used to bias the energy to favor conformations that satisfy the NOE constraints. By default this is turned off. Must be a positive number. 0.0 (kcal/mol)
SAVEINPROPARC Instead of generating a new list of conformers save the conformers in the property archive for use in Spartan's database applications.
PRUNEMETHOD=i Select a different algorithm when deciding which of the conformers to be saved.
  1. Default method of pruning out higher energy conformers, and attempting to keep a diverse set of the low energy conformers using the RMS-torsion definition of nearness.
  2. Use a binning algorithm using the moments of inertia [(3*Ismall/Itot)^2], polar surface area, dipole, and internal dihedrals as measures of diversity.
  3. 2 .. 9 are small variants on the binning algorithm (not tested yet).
  4. Use a cluster algorithm, which is controlled by the DISTANCEMEASURE keyword.
  5. Modified version of 10 but attempt to keep the central molecule in a cluster, as opposed to the 'lowest' energy in a cluster.
(The default is 0.)
REPRUNECONFS
REPRUNECONFS=i,j,tol
Rerun the pruning algorithm as a property calculation. the optional i. i. tol are values for the PRUNEMETHOD=i, DISTANCEMEASURE=j, and DISTANCEISNEAR=tol keywords, respectively. This keyword is a 'property' keyword, and can only be used after a main conformation run which has been executed with the SAVEINPROPARC keyword.
DISTANCEMEASURE=i A scalar number used to measure distance between conformers.
  1. Using differences in the torsion values of important dihedrals. (The default.)
  2. Use the RMS xyz distance of the heavy atoms of a molecule moved to minimize this RMS value.
  3. Use the RMS xyz distance of the heavy atoms of a molecule aligned to the maximally aligned structure.
  4. Use the score of the maximally aligned structure.
  5. Use the number of atoms not aligned in the maximally aligned structure. Then add (1-normalized-score) to discriminate among conformers with same number of 'aligned' atoms.
  6. Use the number of atoms not aligned in the RMS-xyz distance aligned structure
  7. 11,12,13,14,15,16 Use a pharmacophore structure for alignment, as opposed to the atomic structure used in 1,2,3,4,5,6.
The "maximally aligned structure" and it's score is discussed in the align FAQ.
PRUNETOLERANCE=x In the cluster pruning algorithms conformers are considered "near" if they are within this value. If this value is set to zero, this tolerance is not used, and the nearest pairs will be pruned until the number of conformers desired is reached. See DISTANCEMEASURE for the units and definition of distance used in these algorithms. 0.0
SPARSE=x A modification of the systematic search, where a (random) subset of the systematic conformers are tried. If X > 1 then a total of X (random) conformers will be attempted. If X < 1 then a fraction (x) of the total conformers will be attempted.
CONF_SELECTION_RULE=i* There are a number of default rules used to determine which bonds to rotate. These rules are split into different classes, each of which is useful for different types of problems.
Rules used for all rule sets are:
  • Do not rotate bonds around ligand points (such as the Cp ligand)
  • Do not rotate around double bonds.
  • Do not rotate amide bonds
  • Do not rotate around symmetric bonds such as -CH3 and benzene rotors
There are currently 5 sets of rules:
  1. Legacy Spartan 02 rule set.
    • Do not rotate amide bonds
  2. Legacy Trident rule set used in the pharmacophore searching algorithm
    • Do not rotate ester bonds
    • Do not rotate around terminal xH bonds, such as -OH or -NH2.
  3. Normal.
    • Do Not rotate ester bond
    • Rotate amide bond
    • Do not attempt to find twist boat conformations.
    • Attempt to recognize when different conformation of chair like rings (cyclohexane) can be reproduced by a single rotation.
    • Recognize affects of fused rings
  4. Pruned is an extension of the 'Normal' mode but designed to be quicker by only finding 'important' minima.
    • Do not rotate around terminal xH bonds, such as -OH or -NH2.
    • All bonds to nitrogen are given a fold of 2, even if they formally have a 3-fold rotation.
    • Do not rotate carboxcylic acid groups.
    • Do not rotate methoxy groups and ethane groups when attached to benzene.
    • Do not search for multiple conformations of piperazine.
    • Do not rotate long alkyl chains.
  5. Exhaustive is an extension of 'Normal' mode, but designed to find even high energy minima.
    • Rotate ester bonds
    • Attempt to find boat conformations of cyclohexane like rings.
    • Attempt to find conformations of constrained 5 and 6 member rings.
This value can be set as a keyword, or as a system wide preference in Spartan or Trident's preference dialogue (Options Menu). Selection rules are overridden when users specify atoms/bonds from the 'Set Torsions' mode. The default value is '0'.
IGNORE_USERSELECTION Ignore the bond selection from the user, and use the default rules for determining rotatable bonds. (User defaults are set whenever the user enters the 'set torsion' mode.)
DRYRUN Execute only the setup part of conformational analysis. This is only used in debugging.
DISTANCEMEASURE=i Measure to use in determining distance between two conformers.
CONFSEED=The starting point for the random number generator. different every time
CONSTRAIN*
PARTIAL*
NOSYMTRY*
keywords passed to underlying method
FREQ*Will do a frequency calculation on the final candidates.
PRINTLEV=Control Printing. See Description of the output file for more information.
PRINTLEV=2 displays a label for each conformation.
PRINTLEV=3 dumps the intermediate minimization output to the main output window
...other keywords... Unrecognized keywords will be passed to the underlying method. See below for some commonly used keywords.
Keywords specific to the energy profiles
DYNCON*
DYNCON2
Do dynamic constraints/Energy Profile.
SDYNCON Do a energy profile but only return one result. (Usually, this result is a transition state, but if no local maxima exist the local minima will be returned.) This is not a recommended option, and is used primarily for debugging purposes.
DYNCONMETHOD=If there are multiple constraints how are these constraints applied
  • SEQUENTIALly. First one then the next.
  • On a grid GRID, such as a phi-psi plot. or
  • TOGETHER, where both constraints are adjusted in concert with each other.
TOGETHER
KEEPSYMMETRYAttempt to maintain the starting molecule's symmetry.
NORIGID Do not attempt 'rigid moves' but rely on constraints to get the correct structure.
NOOPTDo only rigid moves and single points. Do not attempt to re-minimize.
NOMECHPREOPT Do not prefix each minimization with a mechanics constrained optimization. This is only meaningful for non-mechanic's methods such as AM1.
RIGIDONLYDo only rigid moves and single points. Do not attempt to re-minimize. (Only useful for small "dynamic constraint" systems).
SAVEFILES
SAVEFILES=2
Save temporary files. Useful for debugging. SAVEFILES=2 (or greater) will save even more intermediate files including those of the sub-jobs.
SAVETEMP=Not-implemented
NAMEPREF=abcSave conformers (or energy profile steps) using the name 'abc'. (This will force the creation of a new file even if executing a 'Equilibrium Conformer' job.
REPLACENot-implemented
*   Keywords marked with an asterisk '*' should not be typed in. They are generated by the setup panel.
Other useful keywords
Keywords not recognized in the conformer/energy-profile module will be passed on the underlying method. Following are some keywords found to be useful.
POSTSOLVENT=...
(mechanics only)
Add the solvent model as a final perturbation. Thus, all minimizations are done with the base force field and a solvation 'correction' is applied to the final energy. SM50R is the most used mechanics solvation model. (POSTSOLVENT=SM50R)


Support@wavefun.com

Last modified: Fri May 13 21:35:34 GMT 2016