Mercurial > repos > marpiech > norwich_tools_docking
comparison tools/rdock/bin/sdrmsd @ 3:b02d74d22d05 draft default tip
planemo upload
author | marpiech |
---|---|
date | Mon, 29 Aug 2016 08:23:52 -0400 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
2:bd50f811878f | 3:b02d74d22d05 |
---|---|
1 #! /usr/bin/env python | |
2 # | |
3 # Calculate SMART RMSD with or without molecular superposition (FIT or NOFIT) | |
4 # Script distributed under GNU LGPL 3.0 along rDock software. | |
5 # | |
6 # This algorithm takes into account molecular automorphism. That is, it identifies | |
7 # molecules which are the same but might have atom orders changed and still be able to | |
8 # match the pairs and correctly calculate the RMSD. | |
9 # | |
10 # Author: Daniel Alvarez-Garcia | |
11 # Date: 08-11-2013 | |
12 | |
13 import math | |
14 import pybel | |
15 import numpy as npy | |
16 import optparse | |
17 | |
18 def superpose3D(ref, target, weights=None,refmask=None,targetmask=None,returnRotMat=False): | |
19 """superpose3D performs 3d superposition using a weighted Kabsch algorithm : http://dx.doi.org/10.1107%2FS0567739476001873 & doi: 10.1529/biophysj.105.066654 | |
20 definition : superpose3D(ref, target, weights,refmask,targetmask) | |
21 @parameter 1 : ref - xyz coordinates of the reference structure (the ligand for instance) | |
22 @type 1 : float64 numpy array (nx3) | |
23 --- | |
24 @parameter 2 : target - theoretical target positions to which we should move (does not need to be physically relevant. | |
25 @type 2 : float 64 numpy array (nx3) | |
26 --- | |
27 @parameter 3: weights - numpy array of atom weights (usuallly between 0 and 1) | |
28 @type 3 : float 64 numpy array (n) | |
29 @parameter 4: mask - a numpy boolean mask for designating atoms to include | |
30 Note ref and target positions must have the same dimensions -> n*3 numpy arrays where n is the number of points (or atoms) | |
31 Returns a set of new coordinates, aligned to the target state as well as the rmsd | |
32 """ | |
33 if weights == None : | |
34 weights=1.0 | |
35 if refmask == None : | |
36 refmask=npy.ones(len(ref),"bool") | |
37 if targetmask == None : | |
38 targetmask=npy.ones(len(target),"bool") | |
39 #first get the centroid of both states | |
40 ref_centroid = npy.mean(ref[refmask]*weights,axis=0) | |
41 #print ref_centroid | |
42 refCenteredCoords=ref-ref_centroid | |
43 #print refCenteredCoords | |
44 target_centroid=npy.mean(target[targetmask]*weights,axis=0) | |
45 targetCenteredCoords=target-target_centroid | |
46 #print targetCenteredCoords | |
47 #the following steps come from : http://www.pymolwiki.org/index.php/OptAlign#The_Code and http://en.wikipedia.org/wiki/Kabsch_algorithm | |
48 # Initial residual, see Kabsch. | |
49 E0 = npy.sum( npy.sum(refCenteredCoords[refmask] * refCenteredCoords[refmask]*weights,axis=0),axis=0) + npy.sum( npy.sum(targetCenteredCoords[targetmask] * targetCenteredCoords[targetmask]*weights,axis=0),axis=0) | |
50 reftmp=npy.copy(refCenteredCoords[refmask]) | |
51 targettmp=npy.copy(targetCenteredCoords[targetmask]) | |
52 #print refCenteredCoords[refmask] | |
53 #single value decomposition of the dotProduct of both position vectors | |
54 try: | |
55 dotProd = npy.dot( npy.transpose(reftmp), targettmp* weights) | |
56 V, S, Wt = npy.linalg.svd(dotProd ) | |
57 except Exception: | |
58 try: | |
59 dotProd = npy.dot( npy.transpose(reftmp), targettmp) | |
60 V, S, Wt = npy.linalg.svd(dotProd ) | |
61 except Exception: | |
62 print >> sys.stderr,"Couldn't perform the Single Value Decomposition, skipping alignment" | |
63 return ref, 0 | |
64 # we already have our solution, in the results from SVD. | |
65 # we just need to check for reflections and then produce | |
66 # the rotation. V and Wt are orthonormal, so their det's | |
67 # are +/-1. | |
68 reflect = float(str(float(npy.linalg.det(V) * npy.linalg.det(Wt)))) | |
69 if reflect == -1.0: | |
70 S[-1] = -S[-1] | |
71 V[:,-1] = -V[:,-1] | |
72 rmsd = E0 - (2.0 * sum(S)) | |
73 rmsd = npy.sqrt(abs(rmsd / len(ref[refmask]))) #get the rmsd | |
74 #U is simply V*Wt | |
75 U = npy.dot(V, Wt) #get the rotation matrix | |
76 # rotate and translate the molecule | |
77 new_coords = npy.dot((refCenteredCoords), U)+ target_centroid #translate & rotate | |
78 #new_coords=(refCenteredCoords + target_centroid) | |
79 #print U | |
80 if returnRotMat : | |
81 return new_coords,rmsd, U | |
82 return new_coords,rmsd | |
83 | |
84 | |
85 def squared_distance(coordsA, coordsB): | |
86 """Find the squared distance between two 3-tuples""" | |
87 sqrdist = sum( (a-b)**2 for a, b in zip(coordsA, coordsB) ) | |
88 return sqrdist | |
89 | |
90 def rmsd(allcoordsA, allcoordsB): | |
91 """Find the RMSD between two lists of 3-tuples""" | |
92 deviation = sum(squared_distance(atomA, atomB) for | |
93 (atomA, atomB) in zip(allcoordsA, allcoordsB)) | |
94 return math.sqrt(deviation / float(len(allcoordsA))) | |
95 | |
96 def mapToCrystal(xtal, pose): | |
97 """Some docking programs might alter the order of the atoms in the output (like Autodock Vina does...) | |
98 this will mess up the rmsd calculation with OpenBabel""" | |
99 query = pybel.ob.CompileMoleculeQuery(xtal.OBMol) | |
100 mapper=pybel.ob.OBIsomorphismMapper.GetInstance(query) | |
101 mappingpose = pybel.ob.vvpairUIntUInt() | |
102 exit=mapper.MapUnique(pose.OBMol,mappingpose) | |
103 return mappingpose[0] | |
104 | |
105 def parseArguments(): | |
106 optparse.OptionParser.format_epilog = lambda self, formatter: self.epilog | |
107 epilog = """Args: | |
108 reference.sdf SDF file with the reference molecule. | |
109 input.sdf SDF file with the molecules to be compared to reference.\n""" | |
110 parser = optparse.OptionParser("usage: %prog [options] reference.sdf input.sdf", epilog=epilog) | |
111 parser.add_option("-f", "--fit",dest="fit", action="store_true", default=False, | |
112 help="Superpose molecules before RMSD calculation") | |
113 parser.add_option("--threshold","-t",dest="threshold", action="store", nargs=1, | |
114 help="Discard poses with RMSD < THRESHOLD with respect previous poses which where not rejected based on same principle. A Population SDField will be added to output SD with the population number.", type=float) | |
115 parser.add_option("-o","--out", dest="outfilename", metavar="FILE", default=False, | |
116 help="If declared, write an output SDF file with the input molecules with a new sdfield <RMSD>. If molecule was fitted, the fitted molecule coordinates will be saved.") | |
117 (options, args) = parser.parse_args() | |
118 | |
119 #Check we have two arguments | |
120 if len(args) < 2: | |
121 parser.error("Incorrect number of arguments. Use -h or --help options to print help.") | |
122 | |
123 return options, args | |
124 | |
125 def updateCoords(obmol, newcoords): | |
126 "Update OBMol coordinates. newcoords is a numpy array" | |
127 for i,atom in enumerate(obmol): | |
128 atom.OBAtom.SetVector(*newcoords[i]) | |
129 | |
130 def getAutomorphRMSD(target, molec, fit=False): | |
131 """ | |
132 Use Automorphism to reorder target coordinates to match ref coordinates atom order | |
133 for correct RMSD comparison. Only the lowest RMSD will be returned. | |
134 | |
135 Returns: | |
136 If fit=False: bestRMSD (float) RMSD of the best matching mapping. | |
137 If fit=True: (bestRMSD, molecCoordinates) (float, npy.array) RMSD of best match and its molecule fitted coordinates. | |
138 """ | |
139 mappings = pybel.ob.vvpairUIntUInt() | |
140 bitvec = pybel.ob.OBBitVec() | |
141 lookup = [] | |
142 for i, atom in enumerate(target): | |
143 lookup.append(i) | |
144 success = pybel.ob.FindAutomorphisms(target.OBMol, mappings) | |
145 targetcoords = [atom.coords for atom in target] | |
146 mappose = npy.array(mapToCrystal(target, molec)) | |
147 mappose = mappose[npy.argsort(mappose[:,0])][:,1] | |
148 posecoords = npy.array([atom.coords for atom in molec])[mappose] | |
149 resultrmsd = 999999999999 | |
150 for mapping in mappings: | |
151 automorph_coords = [None] * len(targetcoords) | |
152 for x, y in mapping: | |
153 automorph_coords[lookup.index(x)] = targetcoords[lookup.index(y)] | |
154 mapping_rmsd = rmsd(posecoords, automorph_coords) | |
155 if mapping_rmsd < resultrmsd: | |
156 resultrmsd = mapping_rmsd | |
157 fitted_result = False | |
158 if fit: | |
159 fitted_pose, fitted_rmsd = superpose3D(npy.array(automorph_coords), npy.array(posecoords)) | |
160 if fitted_rmsd < resultrmsd: | |
161 resultrmsd = fitted_rmsd | |
162 fitted_result = fitted_pose | |
163 | |
164 if fit: | |
165 return (resultrmsd, fitted_pose) | |
166 else: | |
167 return resultrmsd | |
168 | |
169 def saveMolecWithRMSD(outsdf, molec, rmsd, population=False): | |
170 newData = pybel.ob.OBPairData() | |
171 newData.SetAttribute("RMSD") | |
172 newData.SetValue('%.3f'%rmsd) | |
173 | |
174 if population: | |
175 popData = pybel.ob.OBPairData() | |
176 popData.SetAttribute("Population") | |
177 popData.SetValue('%i'%population) | |
178 molec.OBMol.CloneData(popData) | |
179 | |
180 molec.OBMol.CloneData(newData) # Add new data | |
181 outsdf.write(molec) | |
182 | |
183 if __name__ == "__main__": | |
184 import sys, os | |
185 | |
186 (opts, args) = parseArguments() | |
187 | |
188 xtal = args[0] | |
189 poses = args[1] | |
190 | |
191 if not os.path.exists(xtal) or not os.path.exists(poses): | |
192 sys.exit("Input files not found. Please check the path given is correct.") | |
193 | |
194 fit = opts.fit | |
195 outfname = opts.outfilename | |
196 threshold = opts.threshold | |
197 | |
198 # Read crystal pose | |
199 crystal = next(pybel.readfile("sdf", xtal)) | |
200 crystal.removeh() | |
201 crystalnumatoms = len(crystal.atoms) | |
202 | |
203 #If outfname is defined, prepare an output SDF sink to write molecules | |
204 if outfname: | |
205 outsdf = pybel.Outputfile('sdf', outfname, overwrite=True) | |
206 | |
207 # Find the RMSD between the crystal pose and each docked pose | |
208 dockedposes = pybel.readfile("sdf", poses) | |
209 if fit: print "POSE\tRMSD_FIT" | |
210 else: print "POSE\tRMSD_NOFIT" | |
211 skipped = [] | |
212 moleclist = {} # Save all poses with their dockid | |
213 population = {} # Poses to be written | |
214 outlist = {} | |
215 for docki, dockedpose in enumerate(dockedposes): | |
216 dockedpose.removeh() | |
217 natoms = len(dockedpose.atoms) | |
218 if natoms != crystalnumatoms: | |
219 skipped.append(docki+1) | |
220 continue | |
221 if fit: | |
222 resultrmsd, fitted_result = getAutomorphRMSD(crystal, dockedpose, fit=True) | |
223 updateCoords(dockedpose, fitted_result) | |
224 else: | |
225 resultrmsd = getAutomorphRMSD(crystal, dockedpose, fit=False) | |
226 | |
227 if threshold: | |
228 # Calculate RMSD between all previous poses | |
229 # Discard if rmsd < FILTER threshold | |
230 if moleclist: | |
231 match = None | |
232 bestmatchrmsd = 999999 | |
233 for did,prevmol in moleclist.iteritems(): | |
234 tmprmsd = getAutomorphRMSD(prevmol, dockedpose) | |
235 if tmprmsd < threshold: | |
236 if tmprmsd < bestmatchrmsd: | |
237 bestmatchrmsd = tmprmsd | |
238 match = did | |
239 | |
240 if match != None: | |
241 # Do not write this one | |
242 # sum one up to the matching previous molecule id | |
243 print >> sys.stderr, "Pose %i matches pose %i with %.3f RMSD"%(docki+1, match+1, bestmatchrmsd) | |
244 population[match] += 1 | |
245 else: | |
246 # There's no match. Print info for this one and write to outsdf if needed | |
247 # Save this one! | |
248 if outfname: outlist[docki] = (dockedpose, resultrmsd) | |
249 print "%d\t%.2f"%((docki+1),resultrmsd) | |
250 moleclist[docki] = dockedpose | |
251 population[docki] = 1 | |
252 else: | |
253 # First molecule in list. Append for sure | |
254 moleclist[docki] = dockedpose | |
255 population[docki] = 1 | |
256 if outfname: outlist[docki] = (dockedpose, resultrmsd) | |
257 else: | |
258 # Just write best rmsd found and the molecule to outsdf if demanded | |
259 if outfname: saveMolecWithRMSD(outsdf, dockedpose, resultrmsd) | |
260 print "%d\t%.2f"%((docki+1),resultrmsd) | |
261 | |
262 if outlist: | |
263 # Threshold applied and outlist need to be written | |
264 for docki in sorted(outlist.iterkeys()): | |
265 molrmsd = outlist[docki] | |
266 # Get number of matchs in thresholding operation | |
267 pop = population.get(docki) | |
268 if not pop: pop = 1 | |
269 # Save molecule | |
270 saveMolecWithRMSD(outsdf, molrmsd[0], molrmsd[1], pop) | |
271 | |
272 if skipped: print >> sys.stderr, "SKIPPED input molecules due to number of atom missmatch: %s"%skipped |