MacroModel Conformational Search Methods

A variety of conformational search methods is available. The methods available in MacroModel are described below.

Monte Carlo Multiple Minimum (MCMM)

The Monte Carlo method (MCMM) implemented in MacroModel is highly efficient in performing global searching, exploring close as well as distant areas of the potential energy surface (PES). Random changes are made in torsion angles during the search. Although there is no limit to the number of variable torsions allowed in the search, more than about 10 or 15 flexible torsions greatly increases the complexity of the search. You should therefore expect increased searching times, and possibly also non-convergence of the search. See References 38 and 39 for more information.

Systematic Pseudo-Monte Carlo (SPMC)

If the task is to locate all possible conformational minima on the surface, MacroModel’s systematic method (SPMC) ensures the exploration of unusual torsional values. SPMC allows for high efficiency throughout the search, even at the end where other methods tend to revisit already found conformers. The SPMC searching method is most efficient for smaller molecules. See Reference 40 for more information.

Low-Mode Conformational Search Methods

If you have little or no prior knowledge of the system to be searched, the Low-Mode Conformational Search (LMCS) [30], based on the principles behind saddle point searching, allows for automatic searching without the need to define parameters to be varied during the search. One important consequence of this is that LMCS allows for a serial search of multiple different molecules without user intervention.

A related method, Large Scale Low-Mode (LLMOD) [32, 33], has been developed for large-scale conformational searching, such as protein loop optimization, homology model refinement, and fully flexible docking for induced fit modeling. LLMOD is similar to LMOD, but computes low-mode eigenvectors of a Hessian matrix that is referenced only implicitly, through its product with a series of vectors. LLMOD is the first conformational search method that can be applied to fully flexible, unconstrained protein structures. Serial LLMOD calculations are not currently supported.

Mixed MCMM/Low-Mode Conformational Search Methods

A pure LMCS search is local in scope, but performs exceptionally well in global searches through hybridization with MCMM [31]. By defining key torsions to be varied through MCMM steps, the mixed MCMM/LMCS search has proved faster and more efficient than any other searching algorithm for a variety of systems. MCMM/LMCS has successfully been applied to difficult tasks, such as searching the conformational space of a ligand in the active site of a protein. Mixed MCMM/LMC2 searches are also supported.