Optimizing Models

[DEBASHIS MUKHOPADHYAY <deb AT indy.saha.ernet.in>]

Dear Modellers,

    It appears that I have successfully modelled a mutated protein using
'model.top' routine which generated 20 output co-ordinates to choose
from. It is noted that each file contains a parameter called MODELLER
OBJECTIVE FUNCTION - most possibly indicating the model quality(?).
Now, the question is which value of the function to consider - the
lowest one or vice versa?

Also, when I checked the secondary str. etc. of one such model using
PROCHECK, several bad contacts were noted. Can you suggest any
particular 'top' routine which will eventually eliminate all these / or
optimize them, so as to make the model acceptable?

Thanks in advance.

Debashis