Biopython DSSP "unexpected sidechain number" Warning

I am trying to retrieve residue-specific secondary structure and solvent accessibility related information from a bunch of PDB files using DSSP program from Biopython. However, I am getting a weird warning about some residues having more sidechain atoms than expected. A minimum working code for reproducing the warning for PDB ID: 5PTI is as follows:

 from Bio.PDB import *
 parser = PDBParser(QUIET=True)

 structure = parser.get_structure(id='struct',file="5pti.pdb") 
 model = structure[0]
 dssp = DSSP(model,"5pti.pdb")

For this specific PDB ID, for example, the first warning is, "Residue ARG 1 A has 11 instead of expected 7 sidechain atoms". I have checked the number of atoms for all the arginine residues in this PDB file using the following code:

structure = parser.get_structure(id='struct',file="5pti.pdb") 
model = structure[0]
residues = Selection.unfold_entities(model, 'R')
count = 0

for idx,residue in enumerate(residues):
    if residue.get_resname() == "ARG":
        atoms = Selection.unfold_entities(residue,'A')
        print(len(atoms),atoms[len(atoms)-1],residue.get_id()[1])
        count = count + 1

print(count)

Can someone please point me towards the reason of this warning? Can I ignore it safely for my purpose? Thanks.

pdb biopython dssp

• 1,845 views • link

written 5.1 years ago by sajid •

I don't get that warning with BioPython 1.76.

• 0 views • link

written 5.1 years ago by Mensur Dlakic

Thanks for this pointer. My BioPython version is 1.79. Since you are not getting this warning, I think I should be good to proceed while ignoring this warning. Does that sound like a sensible thing to you?

• 0 views • link

written 5.1 years ago by sajid •

I don't think this will affect secondary structure assignments, but it could affect the calculation of solvent accessibility. I really don't know because I have always used DSSP directly rather than via BioPython.

It may be a good idea to remove the hydrogens because they are not needed by DSSP. That could also take care of the warning. There is a program called reduce which you can find here. It is meant for adding hydrogens, but it can do the opposite.

reduce -Trim 5pti.pdb > 5pti_noH.pdb

• 0 views • link

written 5.1 years ago by Mensur Dlakic

Thanks a lot. I will try to remove hydrogens and will let you know what happens after that. I am extremely grateful for your help.

• 0 views • link

written 5.1 years ago by sajid •

Thanks to your suggestion of removing the hydrogens, the warning is now gone. I have used biopython for removing the hydrogens. I am putting the script here so that it might help someone in the future.

for model in structure:
    for chain in model:
        for residue in chain:
            atoms_to_remove = []
            for atom in residue:
                if atom.element == "H":
                    atoms_to_remove.append(atom.name)
            for atom in atoms_to_remove:
                residue.detach_child(atom)


io = PDBIO()
io.set_structure(structure)
io.save("X_no_hydrogen.pdb")

• 0 views • link

written 5.1 years ago by sajid •

0 answers

No answers yet.