You are here

Building a PDB from a set of internal coordinates.

3 posts / 0 new
Last post
Building a PDB from a set of internal coordinates.
#1

I am curious if anyone knows of a program in Rosetta that can build a protein or non-canonical polymer chain based on a set of internal coordinates and other geometric criteria?

Specifically I want to find a program that can take a set of dihedral angle data (with the possibility of including long-range, inter-residue contacts) and use these data to generate the Cartesian coordinates of the molecules.

If anyone knows of a particular program in Rosetta that can accomplish this, please point the way.

Post Situation: 
Mon, 2017-03-06 09:47
jcminerlanl

It's unclear why this is in the Rosetta++ section.  Rosetta++ has been dead for years.

Rosetta can certainly convert internal coordinates to 3D coordinates; it does it all the time, thousands or millions of times for a trajectory.

If you mean 'I have a complete description of my structure in internal coordinates and need 3-D coordinates': I am not aware of a protocol that lets the user externally supply the internal coordinates to get the 3D structure. This would be pretty easy to assemble in PyRosetta, though - it's just a LOT of calls to pose.set_phi, set_psi, set_omega (or set_dof); it ought to automatically update the 3D coordinates when you ask for them.

If you mean "I have some experimental data which makes suggestions about SOME of the torsions and contacts in my pose": you have constraints.  You probably want either ab initio or homology modeling with constraints.  

Mon, 2017-03-06 10:32
smlewis

Here's a script I wrote for conversion from internal to 3D coords based on a 2005 paper. I imagine Rosetta uses something really similar? But I guess you can borrow and modify this script if you need.

https://github.com/rbhatta8/protein-design/blob/master/nerf.py

Tue, 2017-03-07 07:22
rbhatta8