[modeller_usage] In what are the initial structures different?

Dear All:

I have a question about the differences of the initial structure built for homology modeling. Can anyone tell me how exactly are these initial structures different? Do they differ in the secondary structures in uncertain regions? I cannot find this information in the manual.

For example, by using

a = automodel() a.starting_model = 1 a.ending_model = 50

I can built 50 different homology models starting from different initial structures, but I have want to know how these different initial structures are constructed.

Thanks very much for help.

Best, Bin