Centroid Rotamer Model adds rotamers in standard Centroid Model for better side-chain representation.

  • Author: Yuan Liu (wendao@uw.edu) (if you need more info or find any problem using it, please contact me, thanks!)
  • last updated: 8/4/2015

Basic idea

For each amino acid, we cluster center of mass of side-chain heavy atoms into rotamer groups. Then we calculate mean value and standard deviation of bond length CEN-CB, bond angle CEN-CB-CA, torsion angle CEN-CB-CA-N for each phi/psi bin using Dunbrack's full-atom rotamer method.

How to use it

First, add -corrections:score:cenrot in your command line for loading centroid rotamer library (this hack should be fixed after Rocco's refactor, but not completely done yet)

The simplest way to use CenRot model is rosetta_scripts, you can easily switch residue type by

<SwitchResidueTypeSetMover name="to_cenrot" set="centroid_rot"/>

Typical mover like packer, minimizer et al. should work properly.

Code example

The lastest version is in my branch (yuan/cenrot_hybird) and will be merged into master very soon. If you want to use it in your code, check pilot/wendao/cenrot_jd2.cc

CenRotRepackMover

normal repack

CenRotRelaxMover

repack + minimize (ramping vdw)

CenRotRigidBodyRelaxMover

repack + minimize (fix bb)

CenRotCanonicalMover

repack + bbGausian

CenRotDockingMover

rigidbody + repack

SmoothFragRepackMover

fragment insertion + repack

RepackMinCenrotMover

repack + minimize (sidechain only)

MinCenrotMover

cartesian/torsion minimize

Score function

\There are couple of different CenRot score functions optimized for different purpose

  • score3_cenrot: for fragment based sampling + repack
  • score4_cenrot_repack: for native backbone repack
  • score4_cenrot_design: for native backbone design
  • score4_cenrot_relax: repack + min in torsion space
  • score4_cenrot_relax_cart: repack + min in cartesian space
  • score4_cenrot_cartmin.wts: minimize in cartesian space (higher vdw)

TODO

more CenRot specific protocol