On 7/10/22 11:15 PM, Li Xue via modeller_usage wrote: > I am trying to refine a loop in a multi-chain complex. During the > refinement/remodelling phase of the loop, does modeller molPDF include > non-bonded energies between chains? Or it will only include the > energies within the loop?

Essentially, both of the above. See https://salilab.org/modeller/10.2/manual/node499.html

First, the loop is optimized without considering interactions with the rest of the system, then a second round of optimization is done which also includes all loop-nonloop interactions.

If you're asking whether the score includes interactions between nonloop atoms in different chains then no, it does not. Only loop atoms are moved during optimization, so any nonloop-nonloop interaction will be constant and so can be ignored. If for some reason you want to include that score (remember, the molpdf is essentially useless for model assessment) then use Selection.energy() to score the final model.

Ben Webb, Modeller Caretaker