Dear Ben, thank you for your message, and sorry to bother you again. I tried to use model.patch() to add the disulfide bridge to my model, but I have a supplementary problem. In fact, my protein is a homodimer and the disulfide bridge should link two Cys residues in the two different subunits. I tried to indicate "A" and "B" in the script in this way: # Create the disulfide bond: def patches(mdl): mdl.patch(residue_type='DISU', residues=(mdl.residues['55:A.'], mdl.residues['180:B']))
but it doesn't work. What is the correct way to include this supplementary information about the chains? Many thanks again and best regards Anna
-----Messaggio originale----- Da: Modeller Caretaker [mailto:modeller-care@salilab.org] Inviato: giovedì 20 novembre 2008 18.58 A: Anna Marabotti Cc: modeller_usage@salilab.org Oggetto: Re: [modeller_usage] question for a (putative?) disulfide bridge
Anna Marabotti wrote: > using mutate_model I replaced a residue in a protein with Cys. > Looking at the mutant protein, I saw that this new Cys residue is > near to another Cys residue that was previously in the protein. The > two sulphur moieties are not facing each others as in a disulfide > bridge, but are rotated away. However, manually rotating the side > chains, they can approach their sulphur moieties until less than 3A > of distance, so in my opinion a disulfide bridge could be formed. At > present I have not experimental data to confirm if a disulfide bridge > is forming or not in this mutant. How can I check with MODELLER if > this newly introduced mutation is really able to form a new disulfide > bridge into my protein? How can I model this new disulfide bridge?
You can use model.patch() to add a disulfide bridge (see http://salilab.org/modeller/9v5/manual/node24.html) and build a model in the usual fashion (you could use your original model prior to mutate_model as your template, and use the mutated sequence as your target, or you could use the output from mutate_model as your template and use a 1:1 alignment). The standard assessment scores will give you an idea whether the resulting model is native-like, but you can't be sure that means a disulfide bridge exists.
Ben Webb, Modeller Caretaker