repair cif file;

[mlynekg3]

Dear cif-tool developers, I want to "repair" a cif file which comes out of phenix so it is read by alphafold. If I put the file from phenix to PDB-REDO the file is read by alphafold, however chain ID's are changed. I talked to Robbie and he suggested me to use cif-tools. I installed it that way
module load Conda

Create a new environment in a project-specific folder

conda create -p ../cif/ciftools
cd ../cif/ciftools

Activate the environment from the project-specific folder

conda activate ../cif/ciftools

Configure the environment, e.g.:

conda config --add channels bioconda
conda config --add channels conda-forge
conda config --set channel_priority strict

Install packages into the environment, e.g.:

conda install -y bioconda::cif-tools

and it seems to work (I just know a little bit about python).

However I get the error
[ml@login02 ciftools]$ cif-validate 7B04_from_phenix.cif
No validator, please specify a dictionary to use using the --dict option
However, the syntax seems to be OK
I check this:
[ml@login02 ciftools]$ ls "$LIBCIFPP_DATA_DIR"
components.cif mmcif_ddl.dic mmcif_ma.dic mmcif_pdbx.dic
[ml@login02 ciftools]$ cif-validate --dict="$LIBCIFPP_DATA_DIR/mmcif_pdbx.dic" -V 7B04_from_phenix.cif
unknown option

Additionally if I just want to "repair" a cif file. There is no cif2cif command. How to do?

Many thanks, br Georg.

[drlemmus]

Hi Georg,
The '-v' for the verbose option is case sensitive.

To repair a cif-file run 'pdb2cif -v foo.cif bar.cif'

[mlynekg3]

Hi, many thanks for the swift reply.

I made a testrun with the cif from PDB and it works:
[ml@login02 ciftools]$ pdb2cif -v 7B04.cif 7B04_fixed.cif
The asym ID's of ndb_poly_seq_scheme and pdbx_poly_seq_scheme are not equal, dropping ndb_poly_seq_scheme
CPU usage: 2m 46.3s user, 0.9s system, 2m 48.0s wall

However using the file which went through phenix does not work.

[ml@login02 ciftools]$ cif-validate --dict="$LIBCIFPP_DATA_DIR/mmcif_pdbx.dic" -v 7B04_from_phenix.cif
Error while setting validator in datablock nxr_refine
Validator not specified
[ml@login02 ciftools]$ pdb2cif -v 7B04_from_phenix.cif 7B04_fixed.cif
Error converting '7B04_from_phenix.cif'
Cannot reconstruct PDBx file, atom data missing

However PDB-REDO is able to run the file.

I have uploaded the file here with the extension .sh, otherwise it was rejected.

I would be grateful for any help. Br, Georg.
7B04_from_phenix.cif.sh

[drlemmus]

My first guess is that it is because there are no _atom_site.auth_comp_id records.

According to the mmCIF-pdbx dictionary "The collection of asym id, atom id, comp id and seq id components of an author's alternative specification for a macromolecular atom site." This file has 3 of the 4 things. Please check if that is indeed the problem and report back to the Phenix developers if so.

[mhekkel]

The syntax should be:

pdb2cif --dict mmcif_pdbx 7B04.cif 7B04_fixed.cif

I realise now that this is not very obvious. Your call with a full path should perhaps be accepted as well.

[mhekkel]

But the file is broken. It contains an almost empty first datablock, containing only the _entry.id...

The actual data is in the datablock data_nxr_refine. If you comment out (or remove) that data_nxr_refine line the conversion works as expected. Albeit a bit slow...