To run MODELLER from Friend you must have MODELLER (version 8 or 9) installed. Load the template structure(s) and sequence you want to model into Friend. Protein sequences can be loaded into Friend by using the menu "File->Load->Alignment from text box" followed by copying/paste of sequence in fasta view on NCBI's or EBI's web pages into the poped-up form. Protein structures can be loaded by using the menu "File->Load->Structure from PDB" followed by typing a 5-letter code in the poped-up form. The first four letters stand for the PDB-code and the last letter represents the chain identifier. Sequence alignment can be prepared manually of using menu "Tools->Pairwise->Align". To proceed to modeling step one has to use the menu "Tools->MODELLER". Note, names of all templates and sequence you want to model should be marked.

The interface to MODELLER has been implemented as a pop-up window, where a user specifies which sequence he/she is modeling (all other sequences are considered as templates, provided they have structures), desired number of output models, output folder and location of the main executable script of MODELLER. By default the current alignment in Friend is used as an input for MODELLER but it can be specified that MODELLER produce an alignment of the modeled sequence and template(s). Additionally, one can specify to include water and hydrogens into energy calculations, produce a model with all explicit hydrogens and/or produce an approximate but quick model.

Upon modeling startup, a progress bar of calculations is displayed. Modeling can be canceled at any time if needed.

If the modeling was successful, the newly created model(s) will be automatically loaded into Friend.

After these steps have been completed, all the error and log messages of MODELLER can be inspected in a new poped-up window.

You can set up a path to MODELLER executalbe to avoid looking for it every time you are doing comparative modelling.