I'm rather new to using MODELLER and would like to seek out the opinions of other users on how to model a two chain protein complex, the T-cell receptor (TCR). I understand most of the basic formatting for multi-chain.ali files, but a few things are eluding me:

1.) How do I build a profile for multi-chain sequences? When I do this using the script found in the tutorial the build_profile.ali does not include the chain break character ('/') in the alignment output, causing the subsequent multi-template program to produce *_fit.pdb files to read as a single chain. This of course is causing me problems. So I'll repeat my interest again - how do I produce a build_profile.ali that contains the chain break character and the appropriate chain information in the header?

2.) I mentioned this in the first part, but I'd like to use multiple templates for this analysis. Do I used the PDB ID and each chain associated with the TCR, when doing so? Also how should I modify the restraints argument between TCR chains?

3.) Accurate modeling of the CDR loops of the TCR (essentially the active site) is important to my research, should I use loopmodel() or other means for loop refinement?

I realize this is rather ambiguous, but any help will be greatly appreciated, in regard to methods that should be used and/or issues discussed (primarily in Q1).

Best,

Ryan