You are here

Ab-initio modeling on part of the protein

2 posts / 0 new
Last post
Ab-initio modeling on part of the protein
#1

Hi all,

I have a design in which I replaced the sequence of the N-terminus region of protein A with the sequence of protein B, and got the chimera c. Now I would like to use Ab-initio modeling to simulate the possible conformations this protein B portion might adopt in the chimera, and I only want to carry out the simulation on this protein B region, while keeping the rest of the chimera structure fixed.

I would like to know what is the acceptable way of doing Ab-initio on only part of the protein?

Thanks so much for your help!

Post Situation: 
Thu, 2013-12-19 07:45
xiayanlawrence

As I understand it, this should be possible, though you have to use the topology broker instead of the regular abinito applications.

There's something of a demo of using the Topology broker in the rosetta_demos/protocol_capture/2010/TopologyBroker_GPCR directory. What you would want to do is omit the membrane specific portions (including the span and lips file) and be sure you give a Rigid file which makes those sections of the protein which you want to be fixed, fixed. (Note that the unreleased r_broker application is almost entirely identical to using the minirosetta applicaiton with the "-run::protocol broker" option.)

As the documentation on the broker is a little off, I've attached a soluble protein example. You'll need to edit the various files to point to the correct filenames, and to have the rigid file list the correct (pose numbered) start and stop positions. (You can have multiple lines if you have multiple rigid regions interspersed with flexible regions. If you don't have a standard Rosetta constraint file to use, just delete the CLAIMER/END_CLAIMER block for the ConstraintClaimer (so it's just the RigidClaimer). Also keep in mind that the input PDB should have all of the residues (or at least the backbone atoms of the residue) already present, although the coordinates of the residues you're remodeling can be arbitrary. (The sequence of the PDB file, as generated from rosetta_source/src/python/apps/public/pdb2fasta.py, should match the full sequence you're trying to model.)

~/rosetta/rosetta_source/bin/minirosetta.default.linuxgccrelease -database ~/rosetta/rosetta_database/ @test.options -in:file:s test.pdb -in:file:fasta test.fasta -in:file:frag3 test_frags.200.3mers -in:file:frag9 test_frags.200.9mers -broker:setup test.tpb -out:nstruct 100 -out:file:scorefile test.sc -out:file:silent test.out -run::protocol broker

Note that the topology broker typically outputs centroid structures. To get full atom structures, you should pass the structures through the relax application (you can convert from silent files to PDB files at the same time, if you want to). Typically this would allow some flexibility in the "rigid" region, but there's ways to prevent movement in selected regions with relax, if that's what you want to do.

Thu, 2013-12-19 15:16
rmoretti