This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Module to RMSD two proteins with Python?

I have two protein structures in PDB format. The two structures refer to two conformations of the same protein. I want to calculate the RMSD between them.

Is there Python module that will allow me to do this?

python protein-structure
• 21,432 views • link
updated by Aleksandr • Aleksandr • written by bef1 • bef1
You can try pymol or tm align
• 0 views • link
written by Pappu Pappu
You can try pymol or tm align
• 0 views • link
written by Pappu Pappu

4 answers

You can use PyMOL to calculate alignment. To do this follow this steps:

  1. run PyMOL
  2. type this command to load your proteins: fetch pdbId1 and fetch pdbId2 *
  3. Use command: rms_cur pdbId1, pdbId2 to calculate RMSD value betweend your structures

--

*where pdbId1 and pdbId2 is identifier of your protein in PDB database.

• 0 views • link
written by crooveck • crooveck

Hi j.cossio.diaz

I struggled with the same problem and i find two module which can solve structural protein alignment in python.

  1. Please take a look to the link: https://github.com/charnley/rmsd

    You can explore the algorithm by using examples, which are exist in the zip file. The only think you have to do is extract the "ATOMS" with x,y,z coordinates from PDB files.

  2. There is also a tool call biopython, which have a Bio.PDB module to align PDB files. here is the link: http://combichem.blogspot.com.tr/2013/08/aligning-pdb-structures-with-biopython.html

• 0 views • link
updated by Ram Ram • written by Poorya Parvizi • Poorya Parvizi

A UCSF Chimera option:

(Taken from a more complicated script I wrote for use with pychimera here: script, pychimera)

NB, pychimera takes a little more involved installation compared to pymol, but I really like it.

The code below should work inside chimera's python interpreter though I believe with no extra configuration.

from MatchMaker import (match,
                        CP_BEST,
                        GAP_OPEN,
                        GAP_EXTEND,
                        defaults,
                        MATRIX,
                        ITER_CUTOFF)

chimera.openModels.open(model1,type="PDB")
chimera.openModels.open(model2,type="PDB")

all_models = chimera.openModels.list(modelTypes=[chimera.Molecule])
mod0 = all_models[0]
mod1 = all_models[1]

for atoms1, atoms2, rmsd, fullRmsd in match(CP_BEST,[mod0, mod1],defaults[MATRIX],
                                            "nw",defaults[GAP_OPEN],defaults[GAP_EXTEND]):

   mol0_name = atoms2[0].molecule
   mol1_name = atoms1[0].molecule
   printmol0_name.name + "\t" + mol2_name.name + "\t" + str(rmsd))
• 0 views • link
written by Joe Joe

In Biopython, the Bio.SVDSuperimposer class will quickly align the two PDB structures and compute the RMSD. See this Bio.PDB FAQ, and search the page for "RMSD", or read the PDB section of the Biopython tutorial.

• 0 views • link
written by Eric T. Eric T.

Hi, I know this might be very late to comment to it, but Bio.SVDSuperimposer does not seem to rotate the proteins to get the lowest RMSD: i get RMSD = 0.2 for chains that when aligned in PyMOL give RMSD = 0.06

here is my code:

fixed = Bio.PDB.Selection.unfold_entities(Bio.PDB.Polypeptide.PPBuilder().build_peptides(Bio.PDB.PDBParser(QUIET = True).get_structure('X' , '1A00_B.pdb') , aa_only = True)[0], 'A')
moving = Bio.PDB.Selection.unfold_entities(Bio.PDB.Polypeptide.PPBuilder().build_peptides(Bio.PDB.PDBParser(QUIET = True).get_structure('X' , '1A00_D.pdb') , aa_only = True)[0], 'A')

sup = Bio.PDB.Superimposer()
sup.set_atoms(fixed , moving)
print(sup.rms)
sup.apply(moving)
print(sup.rms)


OUTPUT:
0.199679715722
0.199679715722

the "sup.apply(moving)" line does not seem to do anything.

Is this the correct way to find the RMSD between two structures in Biopython? is there a better script to write? or a better module to use than Biopython?

• 0 views • link
written by ac.research • ac.research

The RMS in Bio.PDB.Superimposer is calculated using transformed coordinates, so it shows the RMS of the final superimposed result, not the RMS between input arrays.

By the way, your code is not a correct way to test change of RMS, because .apply function doesn't change anything inside Superimposer class, it just applies computed transformation matrix to the argument. The coordinates you provide to .set_atom method are copied to numpy arrays inside Superimposer class, so any change to them after that won't affect Superimposer members.

• 0 views • link
written by Aleksandr • Aleksandr

Log in to answer this question.

More posts like this