-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcreate_data.py
342 lines (242 loc) · 10.6 KB
/
create_data.py
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
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
"""
TODO:
- Load LC50 data from ECOTOX.
- Take median per chemical species pairs.
- Defined chemical groups.
- Export files per chemical groups and each species.
- Forall chemicals and species export relevant KGs.
"""
from tera.DataAggregation import Taxonomy, Effects, Traits
from tera.DataAccess import EffectsAPI
from tera.DataIntegration import DownloadedWikidata, LogMapMapping
from tera.utils import strip_namespace, unit_conversion
from tqdm import tqdm
from rdflib import Graph, URIRef, Literal, BNode, Namespace
from rdflib.namespace import RDFS, RDF
cco = Namespace('http://rdf.ebi.ac.uk/terms/chembl#')
skos = Namespace('http://www.w3.org/2004/02/skos/core#')
obo = Namespace('http://purl.obolibrary.org/obo/')
import pandas as pd
from collections import defaultdict
import pubchempy as pcp
import numpy as np
def get_subgraph(to_visit, graph, backtracking=0):
out = Graph()
visited = set()
while to_visit:
curr = to_visit.pop()
visited.add(curr)
tmp = set(graph.triples((curr,None,None)))
for t in tmp:
out.add(t)
to_visit |= set([o for _,_,o in tmp if not isinstance(o,Literal)])
to_visit -= visited
if backtracking > 0:
tmp = set()
for s in set([s for s,_,_ in out]):
tmp |= set(graph.subjects(object=s))
for t in out:
graph.remove(t)
return out + get_subgraph(tmp, graph, backtracking-1)
return out
def load_endpoint_data():
ed = Effects(directory='../ecotox_data/',verbose=False)
species_mapping = LogMapMapping(filename='./data/final_mappings.txt')
chemicals_mappings = DownloadedWikidata(filename='./data/cas_to_chembl.csv')
species_mapping.load()
chemicals_mappings.load()
ncbi_namespace = Namespace('https://www.ncbi.nlm.nih.gov/taxonomy/')
species_mapping = [(ed.namespace['taxon/'+k],ncbi_namespace['taxon/'+i.pop(0)]) for k,i in species_mapping.mappings.items()]
ed.replace(species_mapping)
chembl_namespace = Namespace('http://rdf.ebi.ac.uk/resource/chembl/molecule/')
chemicals_mappings = [(ed.namespace['cas/'+k],chembl_namespace[i.pop(0)]) for k,i in chemicals_mappings.mappings.items()]
ed.replace(chemicals_mappings)
endpoints = EffectsAPI(dataobject=ed, verbose=True).get_endpoint(c=None, s=None)
d = defaultdict(list)
for c,s,cc,cu,ep,ef,sd,sdu in endpoints:
try:
sd = float(sd)
except:
continue
if 'day' in str(sdu).lower():
sd *= 24
elif 'week' in str(sdu).lower():
sd *= (7*24)
elif 'hour' in str(sdu).lower():
sd *= 1
else:
continue
if ('LC50' in str(ep) or 'LD50' in str(ep) or ('EC50' in str(ep) and 'MOR' in str(ef))) and ('ncbi' in str(s) and 'chembl' in str(c)):
try:
factor = unit_conversion(str(cu),'http://qudt.org/vocab/unit#MilligramPerLitre')
except:
factor = 0
if factor > 0:
cc = float(cc)
cc = cc*factor
d['chemical'].append(str(c))
d['species'].append(str(s))
d['concentration'].append(cc)
d['study_duration'].append(sd)
df = pd.DataFrame(data=d)
df.to_csv('./data/experiments.csv')
def fingerprints():
df = pd.read_csv('./data/experiments.csv')
mapping = DownloadedWikidata(filename='./data/chembl_to_cid.csv')
to_look_for = mapping.convert(set(df['chemical']),reverse=False,strip=True)
to_look_for = set([URIRef('http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID'+str(i)) for k,i in to_look_for.items() if i != 'no mapping'])
out = []
fp = []
for c,c2 in tqdm(zip(to_look_for,set(df['chemical'])),total=len(to_look_for)):
try:
compound = pcp.Compound.from_cid(int(c.split('CID')[-1]))
fp.append(bin(int(compound.fingerprint,16))[2:])
out.append(c2)
except:
pass
df = pd.DataFrame(data={'chemical':out,'fingerprint':fp})
df.to_csv('./data/chemicals_fingerprints.csv')
def chemical_features():
df = pd.read_csv('./data/experiments.csv')
mapping = DownloadedWikidata(filename='./data/chembl_to_cid.csv')
to_look_for = mapping.convert(set(df['chemical']),reverse=False,strip=True)
to_look_for = set([URIRef('http://rdf.ncbi.nlm.nih.gov/pubchem/compound/CID'+str(i)) for k,i in to_look_for.items() if i != 'no mapping'])
out = defaultdict(list)
fp = []
fs = ['xlogp','exact_mass','tpsa','complexity','charge']
for c,c2 in tqdm(zip(to_look_for,set(df['chemical'])),total=len(to_look_for)):
try:
compound = pcp.Compound.from_cid(int(c.split('CID')[-1]))
tmp = compound.to_dict(fs)
for k in tmp:
out[k].append(tmp[k])
out['chemical'].append(c2)
except:
pass
df = pd.DataFrame(data=out)
df.to_csv('./data/chemicals_features.csv')
def load_chemical_groups():
"""
Split (in)organic
CHEBI:50860 (organic)
CHEBI:24835 (inorganic)
"""
df = pd.read_csv('./data/experiments.csv')
chemicals = df['chemical']
graph = Graph()
graph.parse('../chembl/chembl_26.0_molecule_chebi_ls.ttl',format='ttl')
mapping = defaultdict()
for c in chemicals:
c = URIRef(c)
m = list(graph.objects(subject=c,predicate=skos['exactMatch']))
if m:
mapping[c]=m.pop(0)
chebi_graph = Graph()
chebi_graph.parse('../chebi/chebi.ttl',format='ttl')
chebi_graph = replace(chebi_graph,{i:k for k,i in mapping.items()})
for steps in range(1,20):
out = defaultdict(set)
desendents=set()
for c in map(lambda x:URIRef(x), chemicals):
p = ' / '.join('<'+r+'>' for r in [RDFS.subClassOf]*steps)
qres = chebi_graph.query(
"""SELECT DISTINCT ?parent
WHERE {
<%s> %s ?parent
}""" % (str(c),p))
for r in qres:
desendents.add((c,r[-1]))
for c,p in desendents:
out[p].add(c)
df = pd.DataFrame(data={'parent':list(out.keys()),'children':[','.join(out[k]) for k in out]})
df.to_csv('./data/chemical_group_%s.csv' % str(steps))
def load_species_groups():
df = pd.read_csv('./data/experiments.csv')
t = Graph()
t.load('./data/taxonomy_0.ttl',format='ttl')
species = set(df['species'])
for steps in range(0,20):
out_taxon = defaultdict(set)
out_division = defaultdict(set)
desendents = set()
for c in map(URIRef,species):
p = ' / '.join('<'+r+'>' for r in [RDF.type,*[RDFS.subClassOf]*steps])
qres = t.query(
"""SELECT DISTINCT ?parent
WHERE {
<%s> %s ?parent
}""" % (str(c),p))
for r in qres:
desendents.add((c,r[-1]))
for c,p in desendents:
if 'division' in str(p):
out_division[p].add(c)
else:
out_taxon[p].add(c)
df = pd.DataFrame(data={'parent':list(out_division.keys()),'children':[','.join(out_division[k]) for k in out_division]})
df.to_csv('./data/species_divisions.csv')
df = pd.DataFrame(data={'parent':list(out_taxon.keys()),'children':[','.join(out_taxon[k]) for k in out_taxon]})
df.to_csv('./data/species_groups_%s.csv' % str(steps))
def replace(graph,mapping):
for s,p,o in graph:
if s in mapping:
graph.remove((s,p,o))
graph.add((mapping[s],p,o))
if o in mapping:
graph.remove((s,p,o))
graph.add((s,p,mapping[o]))
if p in mapping:
graph.remove((s,p,o))
graph.add((s,mapping[p],o))
return graph
def load_chemical_graph():
df = pd.read_csv('./data/experiments.csv')
mapping = DownloadedWikidata(filename='./data/chembl_to_mesh.csv')
mapping.load()
mapping = {URIRef('http://id.nlm.nih.gov/mesh/'+i.pop(0)):URIRef('http://rdf.ebi.ac.uk/resource/chembl/molecule/'+k) for k,i in mapping.mappings.items()}
mesh_graph = Graph()
mesh_graph.parse('../mesh/mesh.nt',format='nt')
mesh_graph = replace(mesh_graph,mapping)
graph = Graph()
graph.parse('../chembl/chembl_26.0_molecule_chebi_ls.ttl',format='ttl')
mapping = defaultdict()
for c in df['chemical']:
c = URIRef(c)
m = list(graph.objects(subject=c,predicate=skos['exactMatch']))
if m:
mapping[c]=m.pop(0)
chebi_graph = Graph()
chebi_graph.parse('../chebi/chebi.ttl',format='ttl')
chebi_graph = replace(chebi_graph,{i:k for k,i in mapping.items()})
chembl_graph = Graph()
for f in [#'../chembl/chembl_26.0_molecule.ttl',
'../chembl/chembl_26.0_molhierarchy.ttl',
'../chembl/chembl_26.0_target.ttl',
'../chembl/chembl_26.0_targetrel.ttl',
'../chembl/chembl_26.0_moa.ttl']:
chembl_graph.parse(f,format='ttl')
for i,g in enumerate([mesh_graph,chebi_graph,chembl_graph]):
graph = get_subgraph(set([URIRef(a) for a in set(df['chemical'])]), g, backtracking=0)
graph.serialize('./data/chemicals_%s.ttl' % str(i),format='ttl')
def load_taxonomy_graph():
df = pd.read_csv('./data/experiments.csv')
t = Taxonomy(directory='../taxdump/', verbose=True, taxon_namespace='http://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?mode=Info&id=')
ne = DownloadedWikidata(filename='./data/ncbi_to_eol.csv', verbose=False)
n = list(set(t.graph.subjects(predicate=t.namespace['rank'],
object=t.namespace['rank/species'])))
tr = Traits(directory='../eol/', verbose=True)
conv = ne.convert(n, strip=True)
converted = [(tr.namespace[i],k) for k,i in conv.items() if i != 'no mapping']
tr.replace(converted)
for i,g in enumerate([t.graph,tr.graph]):
tmp = set([URIRef(a) for a in set(df['species'])])
graph = get_subgraph(tmp, g, backtracking=0)
graph.serialize('./data/taxonomy_%s.ttl' % str(i),format='ttl')
if __name__ == '__main__':
#load_endpoint_data()
#fingerprints()
#chemical_features()
#load_species_groups()
#load_chemical_groups()
#load_chemical_graph()
load_taxonomy_graph()