Model Loops

Model Loops is an interface to Modeller. Although Modeller is best known as a comparative modeling program, Model Loops currently only interfaces with the loop optimization feature: it generates alternative conformations of peptide segments already present in a structure, but will not build segments that are missing. See also: comparative modeling with Modeller

Modeller is available from the Sali lab for Windows, Mac OS X, Linux, and other UNIX platforms. It is not included with Chimera.

On Windows, Modeller must be run via a script that sets the values of some environmental variables: the script runmod.bat should be text-edited as needed to specify the program version and license key (rem should be removed from the beginning of the license key line), placed in the bin directory of the Modeller installation, and specified instead of the Modeller executable as the Modeller location.

There are several ways to start Model Loops, a tool in the Structure Editing category.

The first step is to select one or more stretches of residues to model. There are several approaches; one is to use the Sequence tool. Stretches of at least four amino acids can be modeled regardless of whether they reside in loops. If multiple stretches are selected, they will be modeled together as a single system. Interactions with the rest of the protein will be taken into account.

Options:

• Restrict selections to molecule - the molecule model of interest should be chosen from the menu. Segments from multiple molecule models cannot be modeled together.
• Restrict selections to turns (true/false) - whether residues assigned as helix or strand should be removed from the selection to model. Helix and strand assignments are taken from the input structure file or generated with ksdssp.
• Include HETATM residues (true/false) - whether in addition to the rest of the protein, HETATM residues should be taken into account during modeling
• Include waters (option not currently available; however, note that waters are usually HETATM residues)
• Include nucleic acids (option not currently available)
• Number of models (default 10) - how many sets of coordinates (alternative conformations) to generate; here, the meaning of "models" is slightly different than in Chimera generally
• Refinement (very fast/fast/slow/very slow/slow large/none) - method of coordinate improvement

The Modeller location is the pathname of the program executable (or the script that runs the program). Input and output files will be placed in a temporary directory (the location will be given in the Reply Log).

Clicking OK starts the calculation. Cancel dismisses the dialog without starting a calculation, and Help opens this manual page in a browser window.

Once the alternative conformations have been generated, they will be displayed and the original coordinates of the segments will be undisplayed (but remain selected). Atoms from residues flanking each modeled segment are included in the output coordinates so that the segment will appear to be connected to the rest of the protein. ViewDock is called to display the conformations. Much like docked ligands, the alternative conformations can be viewed individually by clicking the corresponding lines in the ListBox interface.

Currently, Chimera does not provide a way to replace the original coordinates of the segments with a new conformation from Modeller to produce a new molecule model. This can only be accomplished by editing the text of the coordinate files outside of Chimera.