Mercurial > repos > bgruening > chemfp
annotate chemfp_clustering/nxn_clustering.py @ 29:c0acc635080c
ChemicalToolBoX update.
| author | Bjoern Gruening <bjoern.gruening@gmail.com> |
|---|---|
| date | Tue, 23 Jul 2013 01:02:05 +0200 |
| parents | 1868005213a1 |
| children |
| rev | line source |
|---|---|
| 0 | 1 #!/usr/bin/env python |
| 2 """ | |
| 3 Modified version of code examples from the chemfp project. | |
| 4 http://code.google.com/p/chem-fingerprints/ | |
| 5 Thanks to Andrew Dalke of Andrew Dalke Scientific! | |
| 6 """ | |
| 7 import matplotlib | |
| 8 matplotlib.use('Agg') | |
| 6 | 9 import argparse |
| 0 | 10 import os |
| 11 import chemfp | |
| 12 import scipy.cluster.hierarchy as hcluster | |
| 13 import pylab | |
| 14 import numpy | |
| 15 | |
| 6 | 16 def distance_matrix(arena, tanimoto_threshold = 0.0): |
| 0 | 17 n = len(arena) |
| 6 | 18 # Start off a similarity matrix with 1.0s along the diagonal |
| 19 try: | |
| 20 similarities = numpy.identity(n, "d") | |
| 21 except: | |
| 22 raise Exception('Input dataset is to large!') | |
| 23 chemfp.set_num_threads( args.processors ) | |
| 0 | 24 |
| 6 | 25 ## Compute the full similarity matrix. |
| 26 # The implementation computes the upper-triangle then copies | |
| 27 # the upper-triangle into lower-triangle. It does not include | |
| 28 # terms for the diagonal. | |
| 29 results = chemfp.search.threshold_tanimoto_search_symmetric(arena, threshold=tanimoto_threshold) | |
| 0 | 30 |
| 6 | 31 # Copy the results into the NumPy array. |
| 32 for row_index, row in enumerate(results.iter_indices_and_scores()): | |
| 33 for target_index, target_score in row: | |
| 34 similarities[row_index, target_index] = target_score | |
| 0 | 35 |
| 6 | 36 # Return the distance matrix using the similarity matrix |
| 37 return 1.0 - similarities | |
| 0 | 38 |
| 39 | |
| 6 | 40 if __name__ == "__main__": |
| 41 parser = argparse.ArgumentParser(description="""NxN clustering for fps files. | |
| 42 For more details please see the chemfp documentation: | |
| 43 https://chemfp.readthedocs.org | |
| 44 """) | |
| 45 | |
| 46 parser.add_argument("-i", "--input", dest="input_path", | |
| 47 required=True, | |
| 48 help="Path to the input file.") | |
| 49 | |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
50 parser.add_argument("-c", "--cluster", dest="cluster_image", |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
51 help="Path to the output cluster image.") |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
52 |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
53 parser.add_argument("-s", "--smatrix", dest="similarity_matrix", |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
54 help="Path to the similarity matrix output file.") |
| 6 | 55 |
| 56 parser.add_argument("-t", "--threshold", dest="tanimoto_threshold", | |
| 57 type=float, default=0.0, | |
| 58 help="Tanimoto threshold [0.0]") | |
| 59 | |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
60 parser.add_argument("--oformat", default='png', help="Output format (png, svg)") |
| 6 | 61 |
| 62 parser.add_argument('-p', '--processors', type=int, | |
| 63 default=4) | |
| 64 | |
| 65 args = parser.parse_args() | |
| 66 | |
|
22
6c496b524b41
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
21
diff
changeset
|
67 targets = chemfp.open( args.input_path, format='fps' ) |
|
6c496b524b41
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
21
diff
changeset
|
68 arena = chemfp.load_fingerprints( targets ) |
| 6 | 69 distances = distance_matrix( arena, args.tanimoto_threshold ) |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
70 |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
71 if args.similarity_matrix: |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
72 distances.tofile( args.similarity_matrix ) |
| 6 | 73 |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
74 if args.cluster_image: |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
75 linkage = hcluster.linkage( distances, method="single", metric="euclidean" ) |
| 6 | 76 |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
77 hcluster.dendrogram(linkage, labels=arena.ids) |
| 0 | 78 |
|
23
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
79 pylab.savefig( args.cluster_image, format=args.oformat ) |
|
1868005213a1
ChemicalToolBoX update.
Bjoern Gruening <bjoern.gruening@gmail.com>
parents:
22
diff
changeset
|
80 |