pdb files created by PANDORA do not have a properly formatted HEADER

[DaniBodor]

Hi,

When trying to use DSSP4.0 on pdb files created by pandora, I ran into the issue that these files do not have a properly formatted HEADER line. See this comment.

I now used a workaround to add a header line to each file before running dssp on it, but that is not ideal.

[DarioMarzella]

Hi @DaniBodor ,
Unfortunately, the missing header depends on MODELLER. We could add one after the creation of each model and maybe use it to store information like the ranking of the model...
I will see if we can include this in the full v2.0 release, but I am afraid you will have to use your workaround at least for a while still.

[DaniBodor]

OK, once you do get around to it, maybe make sure it adheres fully to PDB standards: https://www.wwpdb.org/documentation/file-format-content/format33/sect1.html#Order. The other minimum requirement we need for DSSP is a CRYST1 entry.

I am not getting DSSP to work on GitHub actions (it works ok locally) on the current files and strongly suspect it has to do with this, potentially in combination with precise DSSP version (am not managing to get exact same DSSP installed in both places).

[drlemmus]

Thank you for looking at this. DSSP still supports PDB format but is now developed with mmCIF in mind. This means that it had to become stricter in formatting requirements. At some point PDB support will be dropped, but not too soon.