Write mmcif header

[jamesmkrieger]

Now we can add a full mmCIF header to a new set of coordinates, e.g.

In [1]: from prody import *

In [2]: ag1, header1 = parseMMCIF('D_1292146224_model-review_P1.cif', header=True)
@> 30156 atoms and 1 coordinate set(s) were parsed in 6.96s.

In [3]: ag2 = parsePDB('refined-coot-3.pdb')
@> 29673 atoms and 1 coordinate set(s) were parsed in 0.29s.

In [4]: writeMMCIF('D_1292146224_model-review_P1_chimera-coot-6-refmac-coot-3-prody.cif', ag2, header=header1)

[jamesmkrieger]

Actually, this isn't quite there

[jamesmkrieger]

This now more or less works but it is very very slow. It also doesn't have anything about the order of non-loop tables and the starDict that I've made is not a useful structure for interrogating the metadata.

I'll probably improve it further to address these 3 points

[AnthonyBogetti]

@anupam-banerjee Also can you check this one too?

[karolamik13]

I checked the code before (old_) and after (new_) this improvement.

-rw-rw-r-- 1 xxx xxx 423K lip 29 20:36 old_5ir5_test_header.cif
-rw-rw-r-- 1 xxx xxx 556K lip 29 20:37 new_5ir5_test_header.cif

The new one contains more information (header).

The code:
from prody import *

ag1, header1 = parseMMCIF('5IR5', header=True)
ag2 = ag1.copy()
writeMMCIF('5ir5_test_header.cif', ag2, header=header1)

[karolamik13]

[jamesmkrieger]

Great! And it’s not too slow?

[karolamik13]

Great! And it’s not too slow?

I noticed that it is noticeable slower. With one structure it is not a problem, but with hundrads that might be a problem.