How to handle pdb-files with missing ATOM records?

[Bernhard10]

Currently we predend that the missing residues are not there. Is there a better way to do this?

[Bernhard10]

This might effect the fragment library generation in ernwin, if we generate fragments from nr_lists.

[Bernhard10]

See also: https://pdb101.rcsb.org/learn/guide-to-understanding-pdb-data/primary-sequences-and-the-pdb-format

[Bernhard10]

Version 1.0 introduces ftug.get_incomplete_elements and cg.incomplete_elements. The later is a list of element names, inside of which at least one residue is missing.

[ghost]

Could the implementation allow for both?

E. g. to have a switch to allow for reporting (warnings) missing coordinates and also one to "just proceed anyway without notifying the user"?

Either way, it may be useful to put it into the documentation more properly (and what is expected of users; I assume that the ideal way would be if all .pdb files are without error and all users only use error-free .pdb files, even if it may be unrealistic).

[Bernhard10]

I am currently working on this. I have submitted a pullrequest to biopython to include parsing "REMARK 465" lines from PDB: https://github.com/biopython/biopython/pull/1237

Forgi will then be able to act as if the missing residues had never existed, but can also output them when desired. Work on this is in the feature/2.x/missing_residues branch.

[Bernhard10]

The Sequence class in forgi 2.0 now solves this issue.

You can do rna.seq.with_missing or only rna.seq for two versions of the sequence.