If one decides to use models as templates for modeling, must ensure that a good quality model has been built on the first place. This is often not a trivial task.
however a good quality model is usually implies a reasonably closely related experimental structure as a template. Consequently this experimental structure is probably a better starting point than the model itself for future modeling studies.
So, in general I would not consider models as templates, because either they have a highly similar experimental structure or they do not, but in that latter case their quality is highly questionable.
best wishes, Andras
please address your future questions to our discussion forum, many other people could share their experience: http://salilab.org/modeller/discussion_forum.shtml
Andrej Sali wrote: > > -- > Andrej Sali, Associate Professor > Rockefeller University > 1230 York Avenue > New York, NY 10021, USA. > tel: +1 212 327 7550; fax 7540; secretary 7050; labs 7206, 7216 > firstname.lastname@example.org; http://salilab.org > > -----Original Message----- > From: email@example.com [mailto:firstname.lastname@example.org] > Sent: Tuesday, October 29, 2002 11:41 AM > To: email@example.com > Subject: Query > > Dear Prof. Sali, > > I am a student at the University of Mississippi and an > active Modeller user. I had a small doubt about comparative protein > structure modeling and I thought you might be the perfect person to ask > a small doubt that I have. > > Can theoretical models published in PDB be used as one of > the homologs among other crystal structures to build a > homology model? > > If yes, are you aware of any articles where they have > actually used theoretical models as homologs along with > known crystal structures? > > Thank you in advance for your willingness to help. > > Sincerely, > > Yogesh Sabnis > Graduate Research Assistant Tel:(662)915-1286 > Dept. of Medicinal Chemistry Fax:(662)915-5638 > University of Mississippi > E-mail:firstname.lastname@example.org