Mercurial > repos > bgruening > openbabel_subsearch

comparison ob_spectrophore_search.py @ 0:98e12cc1f3a8 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/bgruening/galaxytools/tree/master/chemicaltoolbox/openbabel commit 01da22e4184a5a6f6a3dd4631a7b9c31d1b6d502
author bgruening
date Sat, 20 May 2017 08:40:52 -0400
parents
children 85689710a226
comparison
equal deleted inserted replaced
-1:000000000000 0:98e12cc1f3a8
1 #!/usr/bin/env python
2 """
3 Input: tabular format file with one column storing the unique id for the compounds and any other with the Spectrophores(TM) descriptors.
4 Output: parse the target file using the same protocol used to generate the databases in our servers. Physico-chemical properties are computed and stored as metadata in the sdf output file.
5 Copyright 2012, Bjoern Gruening and Xavier Lucas
6 """
7 import sys, os
8 import argparse
9 import openbabel
10 openbabel.obErrorLog.StopLogging()
11 import pybel
12 import math
13 import numpy as np
14
15 #TODO get rid of eval()
16
17 global spectrophore
18 spectrophore = pybel.ob.OBSpectrophore()
19
20 def parse_command_line():
21 parser = argparse.ArgumentParser()
22 parser.add_argument('--target', required=True, help='target file name in sdf format with Spectrophores(TM) descriptors stored as meta-data')
23 parser.add_argument('--library', required=True, help='library of compounds with pre-computed physico-chemical properties, including Spectrophores(TM) in tabular format')
24 parser.add_argument('-c', '--column', required=True, type=int, help='#column containing the Spectrophores(TM) descriptors in the library file')
25 parser.add_argument('-o', '--output', required=True, help='output file name')
26 parser.add_argument('-n', '--normalization', default="ZeroMeanAndUnitStd", choices=['No', 'ZeroMean', 'UnitStd', 'ZeroMeanAndUnitStd'], help='Normalization method')
27 parser.add_argument('-a', '--accuracy', default="20", choices=['1', '2', '5', '10', '15', '20', '30', '36', '45', '60'], help='Accuracy expressed as angular stepsize')
28 parser.add_argument('-s', '--stereo', default="No", choices=['No', 'Unique', 'Mirror', 'All'], help='Stereospecificity of the cage')
29 parser.add_argument('-r', '--resolution', type=float, default="3.0", help='Resolution')
30 return parser.parse_args()
31
32 def set_parameters(args):
33 if args.normalization == 'No':
34 spectrophore.SetNormalization( spectrophore.NoNormalization )
35 else:
36 spectrophore.SetNormalization( eval('spectrophore.NormalizationTowards' + args.normalization) )
37 spectrophore.SetAccuracy( eval('spectrophore.AngStepSize' + args.accuracy) )
38 spectrophore.SetStereo( eval('spectrophore.' + args.stereo + 'StereoSpecificProbes') )
39 spectrophore.SetResolution( args.resolution )
40 return True
41
42 def Compute_Spectrophores_distance(target_spectrophore, args):
43 outfile = open(args.output, 'w')
44 for mol in open(args.library, 'r'):
45 try:
46 distance = ( ( np.asarray( target_spectrophore, dtype=float ) - np.asarray( mol.split('\t')[ args.column - 1 ].strip().split(', '), dtype=float) )**2).sum()
47 except ValueError:
48 distance = 0
49 outfile.write( '%s\t%f\n' % (mol.strip(), distance ) )
50 outfile.close()
51
52 def __main__():
53 """
54 Computation of Spectrophores(TM) distances to a target molecule.
55 """
56 args = parse_command_line()
57 # This sets up the parameters for the Spectrophore generation. Parameters are set to fit those of our standard parsing tool
58 set_parameters(args)
59
60 mol = pybel.readfile('sdf', args.target).next()
61 target_spectrophore = mol.data["Spectrophores(TM)"].strip().split(', ')
62 # Compute the paired-distance between every molecule in the library and the target
63 distances = Compute_Spectrophores_distance(target_spectrophore, args)
64
65 if __name__ == "__main__" :
66 __main__()