In a homodimer protein, several residues' side chains were missed in A chain or B chain. How can I use the same residue of the other chain to build/complete these missed side chains? How can I think about the coordinates?
Hello Xiao-Yu,
I am assuming that the structure is coming from a crystal structure. Since you will do MD on the system at body temperature (presumably), one could use a program that utilizes internal coordinates (i.e. z-matrix) to artificially build in the missing residues. Some programs (e.g. Amber's tleap) will do this automatically. Then perform an annealing MD simulations where you constrain all of the atoms with the exception of the newly added ones, including the added hydrogen atoms, and allow them to sample/relax the conformational space relative to the atoms with the known crystal coordinates. Once this constrained annealing step is done, then you can release all of the constraints and perform a heating MD, an equilibration MD, and a production MD (often you can do this using a single input file). Remember that the starting structure is a solid state structure usually at a lower temperature than body temperature. Subsequently, many side chains will sample different configuration and/or adopt conformations that are not originally seen in the crystal. The important part for the model building is that you do it logically, reproducibly, and as error free as possible.
Best regards,
Karl
Hello Xiao-Yu,
I am assuming that the structure is coming from a crystal structure. Since you will do MD on the system at body temperature (presumably), one could use a program that utilizes internal coordinates (i.e. z-matrix) to artificially build in the missing residues. Some programs (e.g. Amber's tleap) will do this automatically. Then perform an annealing MD simulations where you constrain all of the atoms with the exception of the newly added ones, including the added hydrogen atoms, and allow them to sample/relax the conformational space relative to the atoms with the known crystal coordinates. Once this constrained annealing step is done, then you can release all of the constraints and perform a heating MD, an equilibration MD, and a production MD (often you can do this using a single input file). Remember that the starting structure is a solid state structure usually at a lower temperature than body temperature. Subsequently, many side chains will sample different configuration and/or adopt conformations that are not originally seen in the crystal. The important part for the model building is that you do it logically, reproducibly, and as error free as possible.
Best regards,
Karl
Hi again. I forgot to mention that prior to each MD stage, make sure to perform a minimization. I often perform the minimizations using the same constraints as will be done in the subsequent MD simulation. This will ensure that erroneous forces are not added at the beginning of each simulation due to high energy structures as determined by the force field.
MD??? I think it is much more premature performing an MD simulation without a model obtained with loop refinement before.
Xiao, see some tutorial about MODELLER (http://salilab.org/modeller/) prior to perform computational intensive calculations such as MD.
As others already advised learning to use MODELLER will help you even in the future.
To your question, you could do that for example in VMD by overlapping one monomer by the other (see "measure fit" and "move") and vice versa. Then save the transformed coordinates and combine the structure to "fill the gaps".
But again, MODELLER would be the preferred way.
thanks so much, i finished this part, yes, coz the main chains are not missing, so i calculated the RMSD for each missing residue in both chains,to make sure they are not so different, and by overlapping one monomer to another, and used the same coordinates to fix the gaps.
Hello Xiao,
I know you have already resolved your problem but i want to know, how you calculate rmsd for each missing residue ? and how you check that there are no huge difference in that?
beacuse i have solved this problem in a different way, first i modeled one chain completely by using modeling tool like Sybyl, SPDBV, and Modeller depending upon the length of missing residue, after completing the structure for one chain,then I used Tk console of VMD software for generating other chain by using symmetry transformation matrix information which is available in pdb file.
- Badges
- Science topic