[Chimera-users] Linking four proteins
meng at cgl.ucsf.edu
Fri Jul 22 10:11:33 PDT 2016
We have a “MultiDomain Assembler” feature that I’d recommend if you were simply starting from the sequence of the multidomain protein and didn’t already have domain structures you wanted to use. I’ll describe that at the end… first I’ll try to answer your question about starting with the domain structures.
I see that what you built already has all the residues (you don’t need to add in linkers).
In that case, you could just connect them with Join Models (menu: Tools… Structure Editing… Build Structure, then go to the Join Models section of the Build Structure tool). It would be an iterative process: first join A to B, then join A-B to C, etc.
However, this will not automatically adjust the structures to make the connections reasonable, it will just make bonds between the ends. The result might be ridiculous. You might try to rotate bonds yourself either before or after joining (see Adjust Torsions section of the Build Structure tool), but considering how many possible bonds there are to rotate, it can be difficult to generate something reasonable just from looking at the structure and tweaking the angles.
Regarding simulations: In the new release (Chimera v1.11) there is a Molecular Dynamics Simulation tool (in menu under Tools… MD/Ensemble Analysis). However, it is somewhat limited in options and may be too slow for anything as large as your system.
You could write out the structure from Chimera and use some other package like AMBER or GROMACS for the simulations. Chimera can write a “prmtop” file for input to AMBER (menu: Tools… Amber… Write Prmtop), and it can display trajectories output from simulations in AMBER and several other packages (menu: Tools… MD/Ensemble Analysis… MD Movie).
… I see I need to update that Write Prmtop manpage with the newer forcefields like AMBER ff14SB, but the actual tool has them. :-)
The MultiDomain Assembler I mentioned earlier is implemented as the “mda” command. It takes the multidomain sequence, BLAST searches against the PDB database to find related structures that could be used to homology-model the domains, then opens those potential template structures and lays them out left to right in sequence order. It also starts up the Modeller interface for building a homology model of the whole thing from thosse different templates. You may want to take a look, there is even a paper (see link in manpage).
With mda, since Modeller is creating the entire protein model (or models, usually you make more than one), I would expect more reasonable overall structure than if you model each domain separately and then stick them together.
I hope this helps,
Elaine C. Meng, Ph.D.
UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco
On Jul 21, 2016, at 8:50 PM, Aditya Padhi <adi.uoh at gmail.com> wrote:
> Dear Chimera users,
> I have predicted a protein structure having 4 different domains (for e.g. 1-94, 95-247, 248-444 and 445-509 residues) and all are part of a single protein from 1-509 residues. They have been predicted from different web-servers.
> Now, I want to join them together to form a single PDB file i.e. make a single PDB file containing the entire protein from 1-509 residues from the above 4 PDB files. Afterwards, I want to run a simulation of the built structure. I searched in google for possible solution but not sure which method would be better. Is there any solution to build this one structure in Chimera?
> Please let me know.
> Thank you so much.
More information about the Chimera-users