Hullo, am a novice user of modeller... need some help for a problem described below...
my target protein contains 2 chains which need to be modelled. I am currently modelling them independently and then concatenating their pdbs to get the complete protein.
Firstly, is is possible to model the 2 chains together, rather than independently, since that would mean taking one chain into context while modelling the other...
Secondly, after generating the protein, i need to check its binding to another protein, with which it is always found complexed. For instance, in case of immunoglobulins, we have 2 chains to be modelled, and then these chains are always found complexed with a peptide-MHC... So I need that the optimization of the chains and loop refinement of the heavy and the light chains should take place taking into consideration the peptide-MHC as well.
The reason i want to do this is that I tried to generate the structure of some xyz protein through modeller, whose crystallographic structure is also available. so i took xyz(crystallographic) as the template and tried to generate the same structure with modeller. the new struc. has an rmsd value of 0.19 as compared to the actual structure, which is quite good. however, at the binding site, there are some side chains that are not nicely fitting on top of the actual structure. and some interactions have gone down, resulting in poor binding. is there a way to rectify this problem?? i was thinking of doing the simulated annealing part taking into context the entire complex... that is, say, after we concatenate the light and heavy chain of immunoglobulin with p-MHC, and then applying molecular dynamics taking into consideration p-MHC also...but am not sure whether that is possible or not... any say??
sincerely, himan...
Yes, you can use a template containing more than one chain for modelling.
Use a modeller script to extract the protein sequence from the pdb file containing both chains and use that for your alignment.
READ_MODEL FILE = 'pdb_file_name' #Without the extension SEQUENCE_TO_ALI ALIGN_CODES = 'pdb_file_name', ATOM_FILES = ALIGN_CODES WRITE_ALIGNMENT FILE = 'pdb_file_name.seq'
The above script should do that for you. You can also produce a model of a complex provided you have a template containing the two proteins.
Cvetan
Quoting Himanshu Grover hgrover_00@iiita.ac.in:
> Hullo, > am a novice user of modeller... need some help for a problem described > below... > > my target protein contains 2 chains which need to be modelled. I am > currently modelling them independently and then concatenating their pdbs > to get the complete protein. > > Firstly, is is possible to model the 2 chains together, rather than > independently, since that would mean taking one chain into context while > modelling the other... > > Secondly, after generating the protein, i need to check its binding to > another protein, with which it is always found complexed. For instance, in > case of immunoglobulins, we have 2 chains to be modelled, and then these > chains are always found complexed with a peptide-MHC... So I need that the > optimization of the chains and loop refinement of the heavy and the light > chains should take place taking into consideration the peptide-MHC as > well. > > The reason i want to do this is that I tried to generate the structure of > some xyz protein through modeller, whose crystallographic structure is > also available. so i took xyz(crystallographic) as the template and tried > to generate the same structure with modeller. the new struc. has an rmsd > value of 0.19 as compared to the actual structure, which is quite good. > however, at the binding site, there are some side chains that are not > nicely fitting on top of the actual structure. and some interactions have > gone down, resulting in poor binding. is there a way to rectify this > problem?? i was thinking of doing the simulated annealing part taking into > context the entire complex... that is, say, after we > concatenate the light and heavy chain of immunoglobulin with p-MHC, and > then applying molecular dynamics taking into consideration p-MHC > also...but am not sure whether that is possible or not... any say?? > > sincerely, > himan... > > > > _______________________________________________ > modeller_usage mailing list > modeller_usage@salilab.org > http://salilab.org/mailman/listinfo/modeller_usage >
-
Cvetan Stojkoski -
Himanshu Grover