File: trajectory_analysis.py

package info (click to toggle)
openstructure 2.11.1-1
  • links: PTS, VCS
  • area: main
  • in suites: forky, sid
  • size: 206,240 kB
  • sloc: cpp: 188,571; python: 36,686; ansic: 34,298; fortran: 3,275; sh: 312; xml: 146; makefile: 29
file content (289 lines) | stat: -rw-r--r-- 9,950 bytes parent folder | download | duplicates (2)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
"""
**This Module requires numpy**

This module contains functions to analyze trajectories, mainly 
similiraty measures baed on RMSDS and pairwise distances.

Author: Niklaus Johner (niklaus.johner@unibas.ch)
"""

import ost.mol.alg
import ost.geom
from ost import LogError
import os

def smooth(vec,n):
#Function to smooth a vector or a list of floats
#for each element it takes the average over itself and the
#n elements on each side, so over (2n+1) elements
  try:
    vec2=vec.copy()
  except:
    vec2=vec[:]
  for i in range(n):
    v=0.0
    count=1.0
    v+=vec[i]
    for j in range(n):
      count+=1
      v+=vec[i+j+1]
    for j in range(i):
      count+=1
      v+=vec[i-(j+1)]
    vec2[i]=v/float(count)
  for i in range(1,n+1):
    v=0.0
    count=1.0
    v+=vec[-i]
    for j in range(n):
      count+=1
      v+=vec[-(i+j+1)]
    for j in range(i-1):
      count+=1
      v+=vec[-i+j+1]
    vec2[-i]=v/float(count)
  for i in range(n,len(vec2)-n):
    v=vec[i]
    for j in range(n):
      v+=vec[i+j+1]
      v+=vec[i-j-1]
    vec2[i]=v/float(2.*n+1.)
  return vec2


"""
From here on the module needs numpy
"""

def RMSD_Matrix_From_Traj(t,sele,first=0,last=-1,align=True,align_sele=None):
  """
  This function calculates a matrix M such that M[i,j] is the
  RMSD (calculated on **sele**) between frames i and j of the trajectory **t**
  aligned on sele.

  :param t: the trajectory
  :param sele: the selection used for alignment and RMSD calculation
  :param first: the first frame of t to be used
  :param last: the last frame of t to be used
  :type t: :class:`~ost.mol.CoordGroupHandle`
  :type sele: :class:`~ost.mol.EntityView`
  :type first: :class:`int`
  :type last: :class:`int`

  :return: Returns a numpy N\\ :subscript:`frames`\\ xN\\ :subscript:`frames` matrix,
   where N\\ :subscript:`frames` is the number of frames.
  """
  if not align_sele:align_sele=sele
  try:
    import numpy as npy
    if last==-1:last=t.GetFrameCount()
    n_frames=last-first
    rmsd_matrix=npy.identity(n_frames)
    for i in range(n_frames):
      if align:
        t=ost.mol.alg.SuperposeFrames(t,align_sele,begin=first,end=last,ref=i)
        eh=t.GetEntity()
      t.CopyFrame(i)
      rmsd_matrix[i,:]=ost.mol.alg.AnalyzeRMSD(t,sele,sele)
      if i==0:
        last=last-first
        first=0
    return rmsd_matrix
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise


def PairwiseDistancesFromTraj(t,sele,first=0,last=-1,seq_sep=1):
  """
  This function calculates the distances between any pair of atoms in **sele**  
  with sequence separation larger than **seq_sep** from a trajectory **t**.
  It return a matrix containing one line for each atom pair and N\\ :subscript:`frames` columns, where
  N\\ :subscript:`frames` is the number of frames in the trajectory.
  
  :param t: the trajectory
  :param sele: the selection used to determine the atom pairs
  :param first: the first frame of t to be used
  :param last: the last frame of t to be used
  :param seq_sep: The minimal sequence separation between atom pairs
  :type t: :class:`~ost.mol.CoordGroupHandle`
  :type sele: :class:`~ost.mol.EntityView`
  :type first: :class:`int`
  :type last: :class:`int`
  :type seq_sep: :class:`int`

  :return: a numpy N\\ :subscript:`pairs`\\ xN\\ :subscript:`frames` matrix.
  """
  try:
    import numpy as npy
    if last==-1:last=t.GetFrameCount()
    n_frames=last-first
    n_var=0
    for i,a1 in enumerate(sele.atoms):
      for j,a2 in enumerate(sele.atoms):
        if not j-i<seq_sep:n_var+=1
    #n_var=sele.GetAtomCount()
    #n_var=(n_var-1)*(n_var)/2.
    dist_matrix=npy.zeros(n_frames*n_var)
    dist_matrix=dist_matrix.reshape(n_var,n_frames)
    k=0
    for i,a1 in enumerate(sele.atoms):
      for j,a2 in enumerate(sele.atoms):
        if j-i<seq_sep:continue
        dist_matrix[k]=ost.mol.alg.AnalyzeDistanceBetwAtoms(t,a1.GetHandle(),a2.GetHandle())[first:last]
        k+=1
    return dist_matrix
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise
    
def DistanceMatrixFromPairwiseDistances(distances,p=2):
  """
  This function calculates an distance matrix M(N\\ :subscript:`frames`\\ xN\\ :subscript:`frames`\\ ) from
  the pairwise distances matrix D(N\\ :subscript:`pairs`\\ xN\\ :subscript:`frames`\\ ), where
  N\\ :subscript:`frames` is the number of frames in the trajectory
  and N\\ :subscript:`pairs` the number of atom pairs.
  M[i,j] is the distance between frame i and frame j
  calculated as a p-norm of the differences in distances
  from the two frames (distance-RMSD for p=2).

  :param distances: a pairwise distance matrix as obtained from 
   :py:func:`~mol.alg.trajectory_analysis.PairwiseDistancesFromTraj`
  :param p: exponent used for the p-norm.

  :return: a numpy N\\ :subscript:`frames`\\ xN\\ :subscript:`frames` matrix, where N\\ :subscript:`frames`
   is the number of frames.
  """
  try:
    import numpy as npy
    n1=distances.shape[0]
    n2=distances.shape[1]
    dist_mat=npy.identity(n2)
    for i in range(n2):
      for j in range(n2):
        if j<=i:continue
        d=(((abs(distances[:,i]-distances[:,j])**p).sum())/float(n1))**(1./p)
        dist_mat[i,j]=d
        dist_mat[j,i]=d
    return dist_mat
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise

def DistRMSDFromTraj(t,sele,ref_sele,radius=7.0,average=False,seq_sep=4,first=0,last=-1):
  """
  This function calculates the distance RMSD from a trajectory.
  The distances selected for the calculation are all the distances
  between pair of atoms from residues that are at least **seq_sep** apart
  in the sequence and that are smaller than **radius** in **ref_sel**.
  The number and order of atoms in **ref_sele** and **sele** should be the same.

  :param t: the trajectory
  :param sele: the selection used to calculate the distance RMSD
  :param ref_sele: the reference selection used to determine the atom pairs and reference distances
  :param radius: the upper limit of distances in ref_sele considered for the calculation
  :param seq_sep: the minimal sequence separation between atom pairs considered for the calculation 
  :param average: use the average distance in the trajectory as reference instead of the distance obtained from ref_sele
  :param first: the first frame of t to be used
  :param last: the last frame of t to be used
  
  :type t: :class:`~ost.mol.CoordGroupHandle`
  :type sele: :class:`~ost.mol.EntityView`
  :type ref_sele: :class:`~ost.mol.EntityView`
  :type radius: :class:`float`
  :type average: :class:`bool`
  :type first: :class:`int`
  :type last: :class:`int`
  :type seq_sep: :class:`int`

  :return: a numpy vecor dist_rmsd(N\\ :subscript:`frames`).
  """
  if not sele.GetAtomCount()==ref_sele.GetAtomCount():
    print('Not same number of atoms in the two views')
    return
  try:
    import numpy as npy
    if last==-1:last=t.GetFrameCount()
    n_frames=last-first
    dist_rmsd=npy.zeros(n_frames)
    pair_count=0.0
    for i,a1 in enumerate(ref_sele.atoms):
      for j,a2 in enumerate(ref_sele.atoms):
        if j<=i:continue
        r1=a1.GetResidue()
        c1=r1.GetChain()
        r2=a2.GetResidue()
        c2=r2.GetChain()      
        if c1==c2 and abs(r2.GetNumber().num-r1.GetNumber().num)<seq_sep:continue
        d=ost.geom.Distance(a1.pos,a2.pos)
        if d<radius:
          a3=sele.atoms[i]
          a4=sele.atoms[j]
          d_traj=ost.mol.alg.AnalyzeDistanceBetwAtoms(t,a3.GetHandle(),a4.GetHandle())[first:last]
          if average:d=npy.mean(d_traj)
          for k,el in enumerate(d_traj):
            dist_rmsd[k]+=(el-d)**2.0
          pair_count+=1.0
    return (dist_rmsd/float(pair_count))**0.5
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise
    
def AverageDistanceMatrixFromTraj(t,sele,first=0,last=-1):
  """
  This function calcultes the distance between each pair of atoms
  in **sele**, averaged over the trajectory **t**.

  :param t: the trajectory
  :param sele: the selection used to determine the atom pairs
  :param first: the first frame of t to be used
  :param last: the last frame of t to be used
  :type t: :class:`~ost.mol.CoordGroupHandle`
  :type sele: :class:`~ost.mol.EntityView`
  :type first: :class:`int`
  :type last: :class:`int`

  :return: a numpy N\\ :subscript:`pairs`\\ xN\\ :subscript:`pairs` matrix, where N\\ :subscript:`pairs`
   is the number of atom pairs in **sele**.
  """
  try:
    import numpy as npy
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise
  n_atoms=sele.GetAtomCount()
  M=npy.zeros([n_atoms,n_atoms])
  for i,a1 in enumerate(sele.atoms):
    for j,a2 in enumerate(sele.atoms):
      if j>i:continue
      d=ost.mol.alg.AnalyzeDistanceBetwAtoms(t,a1.GetHandle(),a2.GetHandle())[first:last]
      M[i,j]=npy.mean(d)
      M[j,i]=npy.mean(d)
  return M

def AnalyzeDistanceFluctuationMatrix(t,sele,first=0,last=-1):
  try:
    import numpy as npy
  except ImportError:
    LogError("Function needs numpy, but I could not import it.")
    raise
  n_atoms=sele.GetAtomCount()
  M=npy.zeros([n_atoms,n_atoms])
  for i,a1 in enumerate(sele.atoms):
    for j,a2 in enumerate(sele.atoms):
      if i>j:continue
      d=ost.mol.alg.AnalyzeDistanceBetwAtoms(t,a1.GetHandle(),a2.GetHandle())[first:last]
      M[j,i]=npy.std(d)
      M[i,j]=npy.std(d)
  return M
  
def IterativeSuperposition(t,sele,threshold=1.0,initial_sele=None,iterations=5,ref_frame=0):
  if initial_sele:current_sele=initial_sele
  else: current_sele=sele
  for i in range(iterations):
    t=ost.mol.alg.SuperposeFrames(t,current_sele,ref=ref_frame)
    al=[a for a in sele.atoms if ost.mol.alg.AnalyzeRMSF(t,ost.mol.CreateViewFromAtoms([a]))<threshold]
    if len(al)==0:return
    current_sele=ost.mol.CreateViewFromAtoms(al)
  return current_sele