[modeller_usage] Modeling multiple domains

Hello-

Is there a way to force/encourage Modeller to join domains in a manner with few or no clashes/atom overlaps between domains?

My first test modeling a multiple domain protein, just as in the example from FAQ #1, resulted in a model with many clashes between the CAs and other atoms of the domains.

I realize that Modeller cannot correctly predict domain orientation, but I hope there is a way to strongly discourage inter-domain clashes.

Thanks,

Arlo Randall University of California, Irvine