fix filterInteractions to be more general

[jamesmkrieger]

So far, I just have a fix to the single selection mode, which didn't include chain, but I'm close to making the whole thing more general too.

[jamesmkrieger]

Originally, we have filterInteractions and getInteractions using either kwargs selection alone, which is just based on residue numbers and names without chain ID, or both kwargs selection and selection2 that must both have chain IDs and don't take any other information. As such, the standard way to analyse interactions between two regions at present is to make them have separate chain IDs.

With the old code, this is the procedure and the results we get for mGluR1 closed chain (1ewkA modified for comparison to 1ewkB in http://www.bahargroup.org/prody/tutorials/clustenmd_tutorial/clustenmd_tutorial_files.zip):

from prody import *

PDBfile = addMissingAtoms('1ewkA_protein_trim.pdb', method='openbabel')
ag3 = parsePDB(PDBfile)

selstr_x='resnum 36 to 206 346 to 472'
selstr_y='resnum 214 to 341 478 to 511'

ag3.setChids('H')
ag3.select(selstr_x).setChids('U')
ag3.select(selstr_y).setChids('L')

interactions_closed = Interactions()
all_interactions = interactions_closed.calcProteinInteractions(ag3)
lobe_interactions_closed = interactions_closed.getInteractions(selection='chain U', 
                                                               selection2='chain L', replace=True)
lobe_interactions_closed

[[['ASN235', 'ND2_1342', 'L', 'ASP191', 'OD2_979', 'U', 2.6414, 29.5225],
  ['ASN235', 'N_1336', 'L', 'SER166', 'OG_785', 'U', 2.7779, 25.7723],
  ['ARG71', 'NH2_278', 'U', 'GLU325', 'OE2_2040', 'L', 3.0711, 22.601],
  ['LYS409', 'NZ_2742', 'U', 'ASP318', 'OD2_1981', 'L', 3.2217, 29.1795]],
 [['LYS409', 'NZ_2742', 'U', 'ASP318', 'OD1_1980_1981', 'L', 4.2535],
  ['ARG71', 'NH1_277_278', 'U', 'GLU325', 'OE1_2039_2040', 'L', 4.434]],
 [],
 [],
 [['TYR74',
   '302_303_304_305_306_307',
   'U',
   'ARG323',
   'NH1_2022_2023',
   'L',
   4.2377]],
 [['ILE190', 'CD1_971', 'U', 'ALA243', 'CB_1399', 'L', 3.6713, 9.0492]],
 []]

With the updated code, we can still reproduce these results:

[[['ASN235', 'ND2_1342', 'L', 'ASP191', 'OD2_979', 'U', 2.6414, 29.5225],
  ['ASN235', 'N_1336', 'L', 'SER166', 'OG_785', 'U', 2.7779, 25.7723],
  ['ARG71', 'NH2_278', 'U', 'GLU325', 'OE2_2040', 'L', 3.0711, 22.601],
  ['LYS409', 'NZ_2742', 'U', 'ASP318', 'OD2_1981', 'L', 3.2217, 29.1795]],
 [['ASP318', 'OD1_1980_1981', 'L', 'LYS409', 'NZ_2742', 'U', 4.2535],
  ['GLU325', 'OE1_2039_2040', 'L', 'ARG71', 'NH1_277_278', 'U', 4.434]],
 [],
 [],
 [['TYR74',
   '302_303_304_305_306_307',
   'U',
   'ARG323',
   'NH1_2022_2023',
   'L',
   4.2377]],
 [['ALA243', 'CB_1399', 'L', 'ILE190', 'CD1_971', 'U', 3.6713, 9.0492]],
 []]

We can also get them more easily as follows:

In [1]: from prody import *

In [2]: PDBfile = addMissingAtoms('1ewkA_protein_trim.pdb', method='openbabel')
   ...: ag3 = parsePDB(PDBfile)
   ...: ag3
Out[2]: <AtomGroup: addH_1ewkA_protein_trim (7059 atoms)>

In [3]: selstr_x='resnum 36 to 206 346 to 472'
   ...: selstr_y='resnum 214 to 341 478 to 511'

In [4]: interactions_closed = Interactions()

In [5]: all_interactions = interactions_closed.calcProteinInteractions(ag3)

In [6]: lobe_interactions_closed = interactions_closed.getInteractions(selection=selstr_x,
   ...:                                                                selection2=selstr_y, replace=True)
   ...: lobe_interactions_closed
Out[6]: 
[[['ASN235', 'ND2_1342', 'A', 'ASP191', 'OD2_979', 'A', 2.6414, 29.5225],
  ['ASN235', 'N_1336', 'A', 'SER166', 'OG_785', 'A', 2.7779, 25.7723],
  ['ARG71', 'NH2_278', 'A', 'GLU325', 'OE2_2040', 'A', 3.0711, 22.601],
  ['LYS409', 'NZ_2742', 'A', 'ASP318', 'OD2_1981', 'A', 3.2217, 29.1795]],
 [['ASP318', 'OD1_1980_1981', 'A', 'LYS409', 'NZ_2742', 'A', 4.2535],
  ['GLU325', 'OE1_2039_2040', 'A', 'ARG71', 'NH1_277_278', 'A', 4.434]],
 [],
 [],
 [['TYR74',
   '302_303_304_305_306_307',
   'A',
   'ARG323',
   'NH1_2022_2023',
   'A',
   4.2377]],
 [['ILE190', 'CD1_971', 'A', 'ALA243', 'CB_1399', 'A', 3.6713, 9.0492]],
 []]

[karolamik13]

Tested. Works fine. Thank you, James.