On 11/26/11 4:05 AM, Полякова Марина wrote: > The thing is that I use multiple modelling while building the requered > protein. I have 10 templates and there's some misunderstanding about the > MODELLER BEST TEMPLATE. How to define it (from the 10 ones I use) and > what is the main principle of working it out?

If I understand you correctly, you are asking about the REMARK line Modeller puts in model PDB files that it makes. An example such line is:

REMARK 1 MODELLER BEST TEMPLATE % SEQ ID: 37.037

That is not the sequence identity of the "best template" (there is really no way of determining such a thing - one template may be better for one part of the model, and another template for another part). It is simply the best sequence identity of a template - i.e. it is the sequence identity of the template that has the highest sequence identity to the model sequence. This is used only by the GA341 assessment method, which uses template sequence identity as an input parameter.

Ben Webb, Modeller Caretaker