Page 1 of 1

Problem with EOM output pdbs

Posted: 2019.04.29 16:49
by jbhopkins
Hi folks,

I'm running EOM on a heterodimer construct connected by a flexible linker. I'm using ATSAS 2.8.4 on MacOS (my local machine). The structure of both domains is known from MX, and was fixed in EOM. The linker and some N and C terminus regions were left flexible. The pool was generated via RANCH using native type conformations. The selected ensemble results (Rg, Dmax, etc) seem to make sense based on what we know from the profile and other biochemical/biophysical data. The problem is that the output sample PDBs are all jumbled. That is, if you open up the PDB file you'll see that the atoms are no longer well ordered by residue number. This means you can't easily visualize it in, for example, PyMOL as the usual cartoon/ribbon (for the fixed domains). A quick spot check of the pool shows this is the case for (seemingly) all of the generated PDBs.

Am I missing a setting to preserve residue order in the output PDBs? Is there a problem with my input PDBs? Is this a bug? Or is it just not possible to maintain that ordering?

I've attached the RANCH log, a sample PDB, and the input PDBs and sequence.

I did try searching the forum for this issue, but it can't use EOM or PDB as search terms, which really limited my ability to find relevant topics.

Thanks in advance for the help.

- Jesse

Re: Problem with EOM output pdbs

Posted: 2019.04.30 11:34
by AL
jbhopkins wrote:
2019.04.29 16:49
Am I missing a setting to preserve residue order in the output PDBs?
Hi Jesse,
unfortunately that's how RANCH works - it calls everything chain A and the order of the residues is a complete chaos. We've opened a bug report for this issue.
Perhaps there is a script that can properly rename the chain IDs, e.g. in haddocking/pdb-tools?

Re: Problem with EOM output pdbs

Posted: 2019.04.30 18:37
by jbhopkins

Thanks for the reply. While you're opening bug reports, can you open one for the fact that the distribution outputs don't report the pool average correctly (it's always 0.00, see attached file for details)?

For those with a similar problem, I've attached a python script that renumbers the residues correctly. To use it, download it and change the file prefix from .txt to .py (you can't attach .py files to this forum). It's had very limited testing, so use at your own risk. You run it from the command line as:

python <path to output pdb directory>

If you want to see what the possible options are, run:

python -h

Hope that helps someone. With the correct numbering, you can at least load into pymol and visualize it easily as a cartoon or otherwise.