IMP logo
IMP Reference Guide  2.10.0
The Integrative Modeling Platform
system_tools.py
1 from __future__ import print_function, division
2 import IMP
3 import IMP.atom
5 import IMP.pmi
6 import IMP.pmi.tools
7 from collections import defaultdict
8 from math import pi
9 import os
10 
11 def resnums2str(res):
12  """Take iterable of TempResidues and return compatified string"""
13  if len(res)==0:
14  return ''
15  idxs = [r.get_index() for r in res]
16  idxs.sort()
17  all_ranges=[]
18  cur_range=[idxs[0],idxs[0]]
19  for idx in idxs[1:]:
20  if idx!=cur_range[1]+1:
21  all_ranges.append(cur_range)
22  cur_range=[idx,idx]
23  cur_range[1]=idx
24  all_ranges.append(cur_range)
25  ret = ''
26  for nr,r in enumerate(all_ranges):
27  ret+='%i-%i'%(r[0],r[1])
28  if nr<len(all_ranges)-1:
29  ret+=', '
30  return ret
31 
32 def get_structure(model,pdb_fn,chain_id,res_range=None,offset=0,model_num=None,ca_only=False):
33  """read a structure from a PDB file and return a list of residues
34  @param model The IMP model
35  @param pdb_fn The file to read (in traditional PDB or mmCIF format)
36  @param chain_id Chain ID to read
37  @param res_range Add only a specific set of residues.
38  res_range[0] is the starting and res_range[1] is the ending residue index
39  The ending residue can be "END", that will take everything to the end of the sequence.
40  None gets you all.
41  @param offset Apply an offset to the residue indexes of the PDB file
42  @param model_num Read multi-model PDB and return that model
43  @param ca_only Read only CA atoms (by default, all non-waters are read)
44  """
46  # Read file in mmCIF format if requested
47  read_file = IMP.atom.read_pdb
48  read_multi_file = IMP.atom.read_multimodel_pdb
49  if pdb_fn.endswith('.cif'):
50  read_file = IMP.atom.read_mmcif
51  read_multi_file = IMP.atom.read_multimodel_mmcif
52  if ca_only:
54  if model_num is None:
55  mh = read_file(pdb_fn,model,
57 
58  else:
59  mhs = read_multi_file(pdb_fn,model,sel)
60  if model_num>=len(mhs):
61  raise Exception("you requested model num "+str(model_num)+\
62  " but the PDB file only contains "+str(len(mhs))+" models")
63  mh = IMP.atom.Selection(mhs[model_num],chain=chain_id,with_representation=True)
64 
65  if res_range==[] or res_range is None:
66  sel = IMP.atom.Selection(mh,chain=chain_id,atom_type=IMP.atom.AtomType('CA'))
67  sel_p = IMP.atom.Selection(mh,chain=chain_id,atom_type=IMP.atom.AT_P)
68  else:
69  start = res_range[0]
70  end = res_range[1]
71  if end=="END":
72  end = IMP.atom.Residue(mh.get_children()[0].get_children()[-1]).get_index()
73  sel = IMP.atom.Selection(mh,chain=chain_id,residue_indexes=range(start,end+1),
74  atom_type=IMP.atom.AtomType('CA'))
75  sel_p = IMP.atom.Selection(mh,chain=chain_id,residue_indexes=range(start,end+1),
76  atom_type=IMP.atom.AT_P)
77  ret = []
78 
79  if sel_p.get_selected_particles():
80  "WARNING: detected nucleotides. Selecting phosphorous instead of CA"
81  sel=sel_p
82 
83  for p in sel.get_selected_particles():
84  res = IMP.atom.Residue(IMP.atom.Atom(p).get_parent())
85  res.set_index(res.get_index() + offset)
86  ret.append(res)
87  if len(ret) == 0:
88  print("WARNING: no residues selected from %s in range %s"
89  % (pdb_fn, res_range))
90  return ret
91 
92 def build_bead(model,residues,input_coord=None):
93  """Generates a single bead"""
94 
95  ds_frag = (residues[0].get_index(), residues[-1].get_index())
96  prt = IMP.Particle(model)
98  ptem = IMP.core.XYZR(prt)
99  mass = IMP.atom.get_mass_from_number_of_residues(len(residues))
100 
101  if ds_frag[0] == ds_frag[-1]:
102  rt = residues[0].get_residue_type()
103  h = IMP.atom.Residue.setup_particle(prt, rt, ds_frag[0])
104  h.set_name('%i_bead' % (ds_frag[0]))
105  prt.set_name('%i_bead' % (ds_frag[0]))
106  try:
108  except IMP.ValueException:
110  IMP.atom.ResidueType("ALA"))
112  ptem.set_radius(radius)
113  else:
115  h.set_name('%i-%i_bead' % (ds_frag[0], ds_frag[-1]))
116  prt.set_name('%i-%i_bead' % (ds_frag[0], ds_frag[-1]))
117  h.set_residue_indexes(range(ds_frag[0], ds_frag[-1] + 1))
118  volume = IMP.atom.get_volume_from_mass(mass)
119  radius = 0.8 * (3.0 / 4.0 / pi * volume) ** (1.0 / 3.0)
120  ptem.set_radius(radius)
121 
123  try:
124  if tuple(input_coord) is not None:
125  ptem.set_coordinates(input_coord)
126  except TypeError:
127  pass
128  return h
129 
130 def build_necklace(model,residues, resolution, input_coord=None):
131  """Generates a string of beads with given length"""
132  out_hiers = []
133  for chunk in list(IMP.pmi.tools.list_chunks_iterator(residues, resolution)):
134  out_hiers.append(build_bead(model,chunk, input_coord=input_coord))
135  return out_hiers
136 
137 def build_ca_centers(model,residues):
138  """Create a bead on the CA position with coarsened size and mass"""
139  out_hiers = []
140  for tempres in residues:
141  residue = tempres.get_hierarchy()
142  rp1 = IMP.Particle(model)
143  rp1.set_name("Residue_%i"%residue.get_index())
144  rt = residue.get_residue_type()
145  this_res = IMP.atom.Residue.setup_particle(rp1,residue)
146  try:
148  except IMP.ValueException:
150  IMP.atom.ResidueType("ALA"))
151  try:
152  mass = IMP.atom.get_mass(rt)
153  except:
155  calpha = IMP.atom.Selection(residue,atom_type=IMP.atom.AT_CA). \
156  get_selected_particles()
157  cp=IMP.atom.Selection(residue,atom_type=IMP.atom.AT_P). \
158  get_selected_particles()
159 
160  if len(calpha)==1:
161  central_atom=calpha[0]
162  elif len(cp)==1:
163  central_atom=cp[0]
164  else:
165  raise("build_ca_centers: weird selection (no Ca, no nucleotide P or ambiguous selection found)")
167  shape = IMP.algebra.Sphere3D(IMP.core.XYZ(central_atom).get_coordinates(),radius)
170  out_hiers.append(this_res)
171  return out_hiers
172 
173 def setup_bead_as_gaussian(mh):
174  """Setup bead as spherical gaussian, using radius as variance"""
175  p = mh.get_particle()
176  center = IMP.core.XYZ(p).get_coordinates()
177  rad = IMP.core.XYZR(p).get_radius()
178  mass = IMP.atom.Mass(p).get_mass()
182 
183 
184 def show_representation(node):
185  print(node)
187  repr = IMP.atom.Representation(node)
188  resolutions = repr.get_resolutions()
189  for r in resolutions:
190  print('---- resolution %i ----' %r)
191  IMP.atom.show_molecular_hierarchy(repr.get_representation(r))
192  return True
193  else:
194  return False
195 
196 def build_representation(parent, rep, coord_finder):
197  """Create requested representation.
198  For beads, identifies continuous segments and sets up as Representation.
199  If any volume-based representations (e.g.,densities) are requested,
200  will instead create a single Representation node.
201  All reps are added as children of the passed parent.
202  @param parent The Molecule to which we'll add add representations
203  @param rep What to build. An instance of pmi::topology::_Representation
204  @param coord_finder A _FindCloseStructure object to help localize beads
205  """
206  built_reps = []
207  atomic_res = 0
208  ca_res = 1
209  model = parent.hier.get_model()
210  if rep.color is not None:
211  if type(rep.color) is float:
212  color = IMP.display.get_rgb_color(rep.color)
213  elif type(rep.color) is str:
214  color = IMP.display.Color(*IMP.pmi.tools.color2rgb(rep.color))
215  elif hasattr(rep.color,'__iter__') and len(rep.color)==3:
216  color = IMP.display.Color(*rep.color)
217  elif type(rep.color) is IMP.display.Color:
218  color = rep.color
219  else:
220  raise Exception("Color must be float or (r,g,b) tuple")
221  else:
222  color = None
223 
224  # first get the primary representation (currently, the smallest bead size)
225  # eventually we won't require beads to be present at all
226  primary_resolution = min(rep.bead_resolutions)
227 
228  # if collective densities, will return single node with everything
229  # below we sample or read the GMMs and add them as representation
230  single_node = False # flag indicating grouping nonlinear segments with one GMM
231  if rep.density_residues_per_component:
232  single_node = True
233  num_components = len(rep.residues)//rep.density_residues_per_component+1
234  rep_dict = defaultdict(list)
235  segp = IMP.Particle(model)
236  root_representation = IMP.atom.Representation.setup_particle(segp,
237  primary_resolution)
238  built_reps.append(root_representation)
239  res_nums = [r.get_index() for r in rep.residues]
240  IMP.atom.Fragment.setup_particle(segp,res_nums)
241  density_frag = IMP.atom.Fragment.setup_particle(IMP.Particle(model),res_nums)
242  density_frag.get_particle().set_name("Densities %i"%rep.density_residues_per_component)
243  density_ps = []
244 
245  if os.path.exists(rep.density_prefix+'.txt') and not rep.density_force_compute:
246  IMP.isd.gmm_tools.decorate_gmm_from_text(rep.density_prefix+'.txt',
247  density_ps,
248  model)
249  if len(density_ps)!=num_components or not os.path.exists(rep.density_prefix+'.txt') or rep.density_force_compute:
250  fit_coords = []
251  total_mass = 0.0
252  for r in rep.residues:
253  for p in IMP.core.get_leaves(r.hier):
254  fit_coords.append(IMP.core.XYZ(p).get_coordinates())
255  total_mass += IMP.atom.Mass(p).get_mass()
256 
257  # fit GMM
258  density_ps = []
260  num_components,
261  model,
262  density_ps,
263  min_covar=4.0,
264  mass_multiplier=total_mass)
265 
266  IMP.isd.gmm_tools.write_gmm_to_text(density_ps,rep.density_prefix+'.txt')
267  if rep.density_voxel_size>0.0:
268  IMP.isd.gmm_tools.write_gmm_to_map(density_ps,rep.density_prefix+'.mrc',
269  rep.density_voxel_size,fast=True)
270 
271  for n, d in enumerate(density_ps):
272  d.set_name('Density #%d' % n)
273  density_frag.add_child(d)
274  root_representation.add_representation(density_frag,
275  IMP.atom.DENSITIES,
276  rep.density_residues_per_component)
277 
278  # get continuous segments from residues
279  segments = []
280  rsort = sorted(list(rep.residues),key=lambda r:r.get_index())
281  prev_idx = rsort[0].get_index()-1
282  prev_structure = rsort[0].get_has_structure()
283  cur_seg = []
284  force_break = False
285  for nr,r in enumerate(rsort):
286  if r.get_index()!=prev_idx+1 or r.get_has_structure()!=prev_structure or force_break:
287  segments.append(cur_seg)
288  cur_seg = []
289  force_break = False
290  cur_seg.append(r)
291  prev_idx = r.get_index()
292  prev_structure = r.get_has_structure()
293  if r.get_index()-1 in rep.bead_extra_breaks:
294  force_break = True
295  if cur_seg!=[]:
296  segments.append(cur_seg)
297 
298  # for each segment, merge into beads
299  name_all = 'frags:'
300  name_count = 0
301  for frag_res in segments:
302  res_nums = [r.get_index() for r in frag_res]
303  rrange = "%i-%i"%(res_nums[0],res_nums[-1])
304  name = "Frag_"+rrange
305  if name_count<3:
306  name_all +=rrange+','
307  elif name_count==3:
308  name_all +='...'
309  name_count+=1
310  segp = IMP.Particle(model,name)
311  this_segment = IMP.atom.Fragment.setup_particle(segp,res_nums)
312  if not single_node:
313  this_representation = IMP.atom.Representation.setup_particle(segp,primary_resolution)
314  built_reps.append(this_representation)
315  for resolution in rep.bead_resolutions:
316  fp = IMP.Particle(model)
317  this_resolution = IMP.atom.Fragment.setup_particle(fp,res_nums)
318  this_resolution.set_name("%s: Res %i"%(name,resolution))
319  if frag_res[0].get_has_structure():
320  # if structured, merge particles as needed
321  if resolution==atomic_res:
322  for residue in frag_res:
323  this_resolution.add_child(residue.get_hierarchy())
324  elif resolution==ca_res and rep.bead_ca_centers:
325  beads = build_ca_centers(model,frag_res)
326  for bead in beads:
327  this_resolution.add_child(bead)
328  else:
329  tempc = IMP.atom.Chain.setup_particle(IMP.Particle(model),"X")
330  for residue in frag_res:
331  tempc.add_child(IMP.atom.create_clone(residue.hier))
332  beads = IMP.atom.create_simplified_along_backbone(tempc,resolution)
333  for bead in beads.get_children():
334  this_resolution.add_child(bead)
335  del tempc
336  del beads
337  else:
338  # if unstructured, create necklace
339  input_coord = coord_finder.find_nearest_coord(min(r.get_index() for r in frag_res))
340  if input_coord is None:
341  input_coord = rep.bead_default_coord
342  beads = build_necklace(model,
343  frag_res,
344  resolution,
345  input_coord)
346  for bead in beads:
347  this_resolution.add_child(bead)
348 
349  # if requested, color all resolutions the same
350  if color:
351  for lv in IMP.core.get_leaves(this_resolution):
353 
354  # finally decide where to put this resolution
355  # if volumetric, collect resolutions from different segments together
356  if single_node:
357  rep_dict[resolution]+=this_resolution.get_children()
358  else:
359  if resolution==primary_resolution:
360  this_representation.add_child(this_resolution)
361  else:
362  this_representation.add_representation(this_resolution,
363  IMP.atom.BALLS,
364  resolution)
365  # if individual beads to be setup as Gaussians:
366  if rep.setup_particles_as_densities:
367  for p in IMP.core.get_leaves(this_resolution):
368  setup_bead_as_gaussian(p)
369  this_resolution.set_name(this_resolution.get_name()+' Densities %i'%resolution)
370  this_representation.add_representation(this_resolution,
371  IMP.atom.DENSITIES,
372  resolution)
373 
374  if single_node:
375  root_representation.set_name(name_all.strip(',')+": Base")
376  d = root_representation.get_representations(IMP.atom.DENSITIES)
377  d[0].set_name('%s: '%name_all + d[0].get_name())
378  for resolution in rep.bead_resolutions:
379  this_resolution = IMP.atom.Fragment.setup_particle(
380  IMP.Particle(model),
381  [r.get_index() for r in rep.residues])
382  this_resolution.set_name("%s: Res %i"%(name_all,resolution))
383  for hier in rep_dict[resolution]:
384  this_resolution.add_child(hier)
385  if resolution==primary_resolution:
386  root_representation.add_child(this_resolution)
387  else:
388  root_representation.add_representation(this_resolution,
389  IMP.atom.BALLS,
390  resolution)
391  return built_reps
Tools for handling Gaussian Mixture Models.
Definition: gmm_tools.py:1
Add mass to a particle.
Definition: Mass.h:23
double get_volume_from_residue_type(ResidueType rt)
Return an estimate for the volume of a given residue.
Simple 3D transformation class.
Represent an RGB color.
Definition: Color.h:24
static Gaussian setup_particle(Model *m, ParticleIndex pi)
Definition: core/Gaussian.h:48
void show_molecular_hierarchy(Hierarchy h)
Print out the molecular hierarchy.
static Fragment setup_particle(Model *m, ParticleIndex pi)
Definition: Fragment.h:63
double get_mass(const Selection &s)
Get the total mass of a hierarchy, in Daltons.
static XYZR setup_particle(Model *m, ParticleIndex pi)
Definition: XYZR.h:48
Select atoms which are selected by both selectors.
Definition: pdb.h:348
double get_mass(ResidueType c)
Get the mass from the residue type.
Color get_rgb_color(double f)
Return the color for f from the RGB color map.
double get_mass_from_number_of_residues(unsigned int num_aa)
Estimate the mass of a protein from the number of amino acids.
Miscellaneous utilities.
Definition: tools.py:1
double get_ball_radius_from_volume_3d(double volume)
Return the radius of a sphere with a given volume.
Definition: Sphere3D.h:35
The type of an atom.
static Residue setup_particle(Model *m, ParticleIndex pi, ResidueType t, int index, int insertion_code)
Definition: Residue.h:157
static Representation setup_particle(Model *m, ParticleIndex pi)
GenericHierarchies get_leaves(Hierarchy mhd)
Get all the leaves of the bit of hierarchy.
A reference frame in 3D.
A Gaussian distribution in 3D.
Definition: Gaussian3D.h:24
def fit_gmm_to_points
fit a GMM to some points.
Definition: gmm_tools.py:231
A decorator for a representation.
double get_volume_from_mass(double m, ProteinDensityReference ref=ALBER)
Estimate the volume of a protein from its mass.
Ints get_index(const ParticlesTemp &particles, const Subset &subset, const Subsets &excluded)
A decorator for a particle representing an atom.
Definition: atom/Atom.h:234
static Mass setup_particle(Model *m, ParticleIndex pi, Float mass)
Definition: Mass.h:44
The type for a residue.
PDBSelector * get_default_pdb_selector()
Definition: pdb.h:473
A decorator for a particle with x,y,z coordinates.
Definition: XYZ.h:30
static Colored setup_particle(Model *m, ParticleIndex pi, Color color)
Definition: Colored.h:62
def write_gmm_to_map
write density map from GMM.
Definition: gmm_tools.py:113
A decorator for a residue.
Definition: Residue.h:134
static bool get_is_setup(const IMP::ParticleAdaptor &p)
Hierarchy create_simplified_along_backbone(Chain input, const IntRanges &residue_segments, bool keep_detailed=false)
Rotation3D get_identity_rotation_3d()
Return a rotation that does not do anything.
Definition: Rotation3D.h:302
Class to handle individual particles of a Model object.
Definition: Particle.h:41
Select all CA ATOM records.
Definition: pdb.h:77
Python classes to represent, score, sample and analyze models.
def write_gmm_to_text
write a list of gaussians to text.
Definition: gmm_tools.py:62
Functionality for loading, creating, manipulating and scoring atomic structures.
static Chain setup_particle(Model *m, ParticleIndex pi, std::string id)
Definition: Chain.h:81
An exception for an invalid value being passed to IMP.
Definition: exception.h:137
Select hierarchy particles identified by the biological name.
Definition: Selection.h:66
Select all ATOM and HETATM records with the given chain ids.
Definition: pdb.h:189
def decorate_gmm_from_text
read the output from write_gmm_to_text, decorate as Gaussian and Mass
Definition: gmm_tools.py:22
A decorator for a particle with x,y,z coordinates and a radius.
Definition: XYZR.h:27