Aline Thomas wrote: > I have built a 3D model on 4 structures aligned from Clustalw. In the > model built, there is a threonine > that bears a dihedral angle different from all equivalent threonines of > the templates. > Locally I cannot see why this threonine should be different. > Would you help me to explain this unprecedented side-chain orientation ?

This is due to a combination of two factors:

1. There are restraints other than chi1 that act on side chains - e.g. distance restraints and nonbonded interactions. It is entirely possible that one or more of these restraints is acting to force the sidechain away from the chi1 value implied by the chi1 restraint.

2. The dihedral restraints are not as specific as "restrain the angle to the same value as in the template". Instead, they restrain the angle so that the distribution of angles between dihedral angle classes in the models matches that in Modeller's database. This distribution is biased by the type of the aligned residue and its own dihedral class.

Ben Webb, Modeller Caretaker