DimerInterfacePrediction

adavtyan edited this page May 13, 2016 · 5 revisions

Introduction

Starting with two monomers separated, gradually lower the temperature to search for the optimal binding configuration. The monomer structure is usually biased (not very rigidly) to the native monomer structure in the dimer using short range fragment potential with the native structure of the monomer (so-called single memory), a weak center-of-mass bias is added to encourage the monomers to come close to each other. All files mentioned below can be found in trunk/examples/1CTA_Dimer_Binding

Details

Prepare input files for binding simulations.

  • Generate input file 1cta.in:

Download the dimer structure of Troponin C site III (1CTA), copy it to 1cta.pdb, remove the water molecues, remove additional chains other than the two that will be simulated, then run:

$ python $path/create_project_tools/PDBToCoordinates.py 1cta 1cta.coord ## first argument '1cta' is the name of the pdb file. This script generate 1cta.coord $ python $path/create_project_tools/CoordinatesToWorkLammpsDataFile.py 1cta.coord data.1cta -b

The three newly generated files are needed for simulations: 1cta.in 1cta.seq data.1cta. Check to see if the number of residues in 1cta.seq match with that in 1cta.pdb. If not, go to the "Troubleshooting section" below for more details.

  • A weak long-range force is added to keep the two monomers in the mutual vicinity. Please note that two groups (chainA_CA & chainB_CA) are added to the 1cta.in file, and a new line "fix 3 chainA_CA spring couple chainB_CA 0.01(biasing strength k) 0 0 0 10(equilibrium biasing distance d0)" is used to add the biasing force.

  • Build fragment memory file: fragsLAMW.mem.single

$ python $path/frag_mem_tools/Pdb2Gro.py 1cta 1ctaa.gro A ### this will generate 1ctaa.gro using the chain A in pdb file 1cta.pdb.

Now we can use 1ctaa.gro as the single memory for both monomers. Create file fragsLAMW.mem.single.

  • Generate ssweight file

The ssweight file contains the secondary structure information for the monomer native state. It has N rows (N is the total number of residues) and two columns. One way to generate this file is to get the "plain" version (from http://webclu.bio.wzw.tum.de/cgi-bin/stride/db.py) of the stride database entry into a file called ssweight.stride, then run

$ python $path/create_project_tools/stride2ssweight.py > ssweight

list of files: 1cta.pdb, 1cta.in, 1cta.seq, data.1cta, fragsLAMW.mem.single

Run the simulation.

  • Separate two monomers from the bound complex. Use a strong force (k=2) to pull the monomers apart, equilibrium distance between the center of mass of each monomer d0=80 A. e.g., modify line 'fix 3' to:

fix 3 chainA_CA spring couple chainB_CA 2 0 0 0 80 ## k=2, d0=80 Generate restart file (and rename it to 'unfold.separated') from the end of the simulation, by uncommenting the next to the last line of 1cta.in.

  • Run simulated annealing simulations for binding prediction. Use restart file unfold.separated to start the simulation (by replacing 'read_data data.1ctaDimer' with 'read_restart unfold.separated'). Reduce the biasing force (k=0.01), change d0 to 10 so that two monomers can be pulled closer to each other.

fix 3 chainA_CA spring couple chainB_CA 0.02 0 0 0 10 ## k=0.02, d0=80 Recommended annealing schedule: 450K-350K over 4000000 steps.

Troubleshooting

  1. When creating input files (e.g., 1cta.in) and data files (e.g., data.1cta) from the PDB file, a common problem is that seq file doesn't have the same number of residues as the one in the PDB file or contains irregular residues (e.g. '?' or 'X'). The following is an example error message of a typical problem: some residues have missing 'O' atom. Error message: Traceback (most recent call last):

File "/home/wz12/opt/script/PDBToCoordinates.py", line 193, in module

xyz_O = res['O'].get_coord()

File "/apps/python/lib64/python2.6/site-packages/Bio/PDB/Entity.py", line 38, in getitem

return self.child_dictid KeyError: 'O'

Solution: a) Try a different chain; or b)Use other software (e.g. Modeller) to fix the missing atom. Note: a similar error message will be given when CA or CB atoms are missing.

  1. About missing residues: Warning: if an entire residue(s) is missing from the PDB file, the program may not give an error message. This is why it is crucial to double check whether the PDB file is consistent with the generated .seq file in terms of number of residues.
You can’t perform that action at this time.
You signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session.
Press h to open a hovercard with more details.