Clashes No Longer Found When Removing Chain

[ovavourakis]

I have two example pdb files (here).
both_chains.pdb contains the backbone atoms (only) of a two-chain antibody Fv.
VH_only.pdb contains the backbone atoms (only) of the just the first chain of the same antibody (i.e. it's missing the second chain, but is otherwise identical to the first file).

Bug: Running backbone on either file does not return the same clashes, regardless of whether it's run from the Python API or the CLI.
This is despite the fact that the same clashes are detected in both files, when running clashes() from the Python API.

To Reproduce

ab = StructureChecking('path/to/biobb_structure_checking', {'input_structure_path':'/path/to/both_chains.pdb'})
vh = StructureChecking('path/to/biobb_structure_checking', {'input_structure_path':'/path/to/VH_only.pdb'})

ab.backbone()

Running backbone.
Found 2 Residues with missing backbone atoms
SER H123   OXT
LYS L112   OXT
No backbone breaks
Found Unexpected backbone links
ASP H58    linked to GLY H60   , expected GLY H59            <----- notice these in the heavy chain 'H'

vh.backbone()

Running backbone.
Found 1 Residues with missing backbone atoms
SER H123   OXT
No backbone breaks
No unexpected backbone links                 <----- they're missing if you run the heavy chain by itself

vh.clashes()

Running clashes.
1 Steric severe clashes detected
ASP H58.C    GLY H60.N       0.125 A                              <----- but they are here!
2 Steric apolar clashes detected
GLY H33.O    GLY H104.C      2.741 A
ARG H71.CB   ASN H88.O       2.673 A
2 Steric polar_acceptor clashes detected
VAL H12.O    SER H122.O      3.067 A
SER H55.O    ASP H58.O       2.942 A
2 Steric polar_donor clashes detected
ASP H58.N    GLY H60.N       2.559 A
GLY H59.N    VAL H61.N       2.515 A
No positive clashes detected
No negative clashes detected

ab.clashes()

Running clashes.
1 Steric severe clashes detected
ASP H58.C    GLY H60.N       0.125 A                              <----- just as they are here!
6 Steric apolar clashes detected
GLY H33.O    GLY H104.C      2.741 A
PRO H65.CB   PRO L100.CB     2.429 A
ARG H71.CB   ASN H88.O       2.673 A
GLY H108.CB  SER L96.CB      2.425 A
GLY L71.CA   ASP L75.O       2.778 A
GLY L71.CB   PHE L76.CA      2.615 A
4 Steric polar_acceptor clashes detected
VAL H12.O    SER H122.O      3.067 A
SER H55.O    ASP H58.O       2.942 A
VAL L12.O    GLU L110.O      3.049 A
SER L70.O    THR L77.O       3.044 A
2 Steric polar_donor clashes detected
ASP H58.N    GLY H60.N       2.559 A
GLY H59.N    VAL H61.N       2.515 A
No positive clashes detected
No negative clashes detected

The same happens if you run backbone from the CLI.

Expected behaviour

I'd expect both backbone() reports to mention the clash between ASP H58.C and GLY H60.N, pointed out above, and I'd expect that to also work from the CLI.

Desktop (please complete the following information):

• OS: Fedora Linux 38 (Thirty Eight)
• Kernel: Linux 6.7.11-100.fc38.x86_64
• Python version: 3.10.13
• BioBB version: 3.13.5

[jlgelpi]

Hi, thanks for the feedback.
You are right, but this is the expected behaviour. To speed up the calculation, backbone() checks only the clashes due to the added or modified backbone atoms or residues (none in this case except for the OXT atom), it does not make a complete clashes analysis (like clashes() does).
On the other hand, the clash is triggering the warning about unexpected peptide bond, as the distance C-N is compatible with such a bond. We will check why it does not appear with the H subunit alone, that is not expected.