models.py 28.6 KB
Newer Older
Hervé  MENAGER's avatar
Hervé MENAGER committed
1
2
3
from __future__ import unicode_literals

from django.db import models
4
from django.db.models import Max
Hervé  MENAGER's avatar
Hervé MENAGER committed
5

6
from .utils import FingerPrinter
7
from .ws import get_pubmed_info, get_google_patent_info, get_uniprot_info, get_taxonomy_info, get_go_info, get_pfam_info
8

Hervé  MENAGER's avatar
Hervé MENAGER committed
9

10
class AutoFillableModel(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
11

12
13
14
15
16
17
18
19
20
21
22
23
24
    """
    AutoFillableModel makes it possible to automatically fill model fields from
    external sources in the autofill() method
    The save method allows to either include autofill or not. in autofill kwarg is
    set to True, save() will first call autofill(), otherwise it won't
    """

    class Meta:
        abstract = True

    def save(self, *args, **kwargs):
        if kwargs.get('autofill') is True:
            self.autofill()
25
26
        if 'autofill' in kwargs:
            del kwargs['autofill']
Hervé  MENAGER's avatar
Hervé MENAGER committed
27
        super(AutoFillableModel, self).save(*args, **kwargs)
28
29
30


class Bibliography(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
31

Hervé  MENAGER's avatar
Hervé MENAGER committed
32
33
34
35
    """
    Bibliography data table
    """
    SOURCES = (
36
37
38
        ('PM', 'PubMed ID'),
        ('PT', 'Patent'),
        ('DO', 'DOI ID')
Hervé  MENAGER's avatar
Hervé MENAGER committed
39
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
40
41
    source = models.CharField(
        'Bibliographic type', max_length=2, choices=SOURCES, default='PM')
Hervé  MENAGER's avatar
Hervé MENAGER committed
42
43
    id_source = models.CharField('Bibliographic ID', max_length=25)
    title = models.CharField('Title', max_length=300)
44
    journal_name = models.CharField('Journal name', max_length=50, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
45
46
    authors_list = models.CharField('Authors list', max_length=500)
    biblio_year = models.PositiveSmallIntegerField('Year')
47
    cytotox = models.BooleanField('Cytotoxicity data', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
48
49
50
51
    in_silico = models.BooleanField('in silico study', default=False)
    in_vitro = models.BooleanField('in vitro study', default=False)
    in_vivo = models.BooleanField('in vivo study', default=False)
    in_cellulo = models.BooleanField('in cellulo study', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
52
53
    pharmacokinetic = models.BooleanField(
        'pharmacokinetic study', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
54
    xray = models.BooleanField('X-Ray data', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
55

56
    def autofill(self):
57
58
59
        if self.source == 'PM':
            info = get_pubmed_info(self.id_source)
        else:
60
            info = get_google_patent_info(self.id_source)
61
62
63
64
        self.title = info['title']
        self.journal_name = info['journal_name']
        self.authors_list = info['authors_list']
        self.biblio_year = info['biblio_year']
Hervé  MENAGER's avatar
Hervé MENAGER committed
65

Hervé  MENAGER's avatar
Hervé MENAGER committed
66
67
68
    class Meta:
        verbose_name_plural = "bibliographies"

69
70
    def __str__(self):
        return '{}, {}'.format(self.source, self.id_source)
71

Hervé  MENAGER's avatar
Hervé MENAGER committed
72

73
class Taxonomy(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
74
75
    taxonomy_id = models.DecimalField(
        'NCBI TaxID', unique=True, max_digits=9, decimal_places=0)
Hervé  MENAGER's avatar
Hervé MENAGER committed
76
    name = models.CharField('Organism name', max_length=200)
77

78
    def autofill(self):
79
80
81
        info = get_taxonomy_info(self.taxonomy_id)
        self.name = info['scientific_name']

82
83
84
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
85
86
    class Meta:
        verbose_name_plural = "taxonomies"
Hervé  MENAGER's avatar
Hervé MENAGER committed
87

Hervé  MENAGER's avatar
Hervé MENAGER committed
88

89
class MolecularFunction(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
90
    go_id = models.CharField('Gene Ontology ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
91
    # GO term id format: 'GO:0000000'
Hervé  MENAGER's avatar
Hervé MENAGER committed
92
93
    description = models.CharField('description', max_length=500)

94
    def autofill(self):
95
96
97
        info = get_go_info(self.go_id)
        self.description = info['label']

98
99
100
    def __str__(self):
        return self.description

Hervé  MENAGER's avatar
Hervé MENAGER committed
101

102
class Protein(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
103
    uniprot_id = models.CharField('Uniprot ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
104
105
    recommended_name_long = models.CharField(
        'Uniprot Recommended Name (long)', max_length=75)
Hervé  MENAGER's avatar
Hervé MENAGER committed
106
107
108
    short_name = models.CharField('Short name', max_length=50)
    gene_name = models.CharField('Gene name', unique=True, max_length=30)
    entry_name = models.CharField('Entry name', max_length=30)
109
    organism = models.ForeignKey('Taxonomy', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
110
111
    molecular_functions = models.ManyToManyField(MolecularFunction)

112
    def autofill(self):
113
        info = get_uniprot_info(self.uniprot_id)
114
        self.recommended_name_long = info['recommended_name']
115
116
        self.gene_name = info['gene']
        self.entry_name = info['entry_name']
117
        self.short_name = info['short_name']
118
119
120
121
122
        try:
            taxonomy = Taxonomy.objects.get(taxonomy_id=info['organism'])
        except Taxonomy.DoesNotExist:
            taxonomy = Taxonomy()
            taxonomy.taxonomy_id = info['organism']
123
            taxonomy.save(autofill=True)
124
        self.organism = taxonomy
125
        super(Protein, self).save()
126
127
128
129
130
131
        for go_id in info['molecular_functions']:
            try:
                mol_function = MolecularFunction.objects.get(go_id=go_id)
            except MolecularFunction.DoesNotExist:
                mol_function = MolecularFunction()
                mol_function.go_id = go_id
132
                mol_function.save(autofill=True)
133
            self.molecular_functions.add(mol_function)
134

135
136
137
    def __str__(self):
        return '{} ({})'.format(self.uniprot_id, self.recommended_name_long)

Hervé  MENAGER's avatar
Hervé MENAGER committed
138

139
class Domain(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
140
141
    pfam_acc = models.CharField('Pfam Accession', max_length=10, unique=True)
    pfam_id = models.CharField('Pfam Family Identifier', max_length=20)
Hervé  MENAGER's avatar
Hervé MENAGER committed
142
    pfam_description = models.CharField('Pfam Description', max_length=100)
Hervé  MENAGER's avatar
Hervé MENAGER committed
143
    domain_family = models.CharField('Domain family', max_length=25)
Hervé  MENAGER's avatar
Hervé MENAGER committed
144
145
    # TODO: what is this field? check database
    # contents
146

147
    def autofill(self):
148
149
150
        info = get_pfam_info(self.pfam_acc)
        self.pfam_id = info['id']
        self.pfam_description = info['description']
Hervé  MENAGER's avatar
Hervé MENAGER committed
151

152
153
154
    def __str__(self):
        return '{} ({}-{})'.format(self.pfam_acc, self.pfam_id, self.pfam_description)

Hervé  MENAGER's avatar
Hervé MENAGER committed
155

156
class ProteinDomainComplex(models.Model):
157
158
    protein = models.ForeignKey('Protein', models.CASCADE)
    domain = models.ForeignKey('Domain', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
159
160
161
    ppc_copy_nb = models.IntegerField(
        'Number of copies of the protein in the complex')

Hervé  MENAGER's avatar
Hervé MENAGER committed
162
163
    class Meta:
        verbose_name_plural = "complexes"
164

165
166
167
    def __str__(self):
        return '{}-{}'.format(self.protein_id, self.domain_id)

168
169
    def name(self):
        return self.protein.short_name
Hervé  MENAGER's avatar
Hervé MENAGER committed
170

171

172
class ProteinDomainBoundComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
173
174
175
    ppp_copy_nb_per_p = models.IntegerField(
        'Number of copies of the protein in the pocket')

176
177
    class Meta:
        verbose_name_plural = "bound complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
178
179


180
class ProteinDomainPartnerComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
181

182
183
    class Meta:
        verbose_name_plural = "partner complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
184

Hervé  MENAGER's avatar
Hervé MENAGER committed
185

186
187
188
class Symmetry(models.Model):
    code = models.CharField('Symmetry code', max_length=2)
    description = models.CharField('Description', max_length=300)
Hervé  MENAGER's avatar
Hervé MENAGER committed
189

190
191
192
    class Meta:
        verbose_name_plural = "symmetries"

193
194
195
    def __str__(self):
        return '{} ({})'.format(self.code, self.description)

196
197

class Disease(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
198
    name = models.CharField('Disease', max_length=30, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
199
    # is there any database/nomenclature for diseases?
200
201
202
203

    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
204

Hervé  MENAGER's avatar
Hervé MENAGER committed
205
206
207
class PpiFamily(models.Model):
    name = models.CharField('Name', max_length=30, unique=True)

208
209
210
    class Meta:
        verbose_name_plural = "PPI Families"

Hervé  MENAGER's avatar
Hervé MENAGER committed
211
212
    def __str__(self):
        return self.name
Hervé  MENAGER's avatar
Hervé MENAGER committed
213

Hervé  MENAGER's avatar
Hervé MENAGER committed
214

215
class Ppi(AutoFillableModel):
216
    pdb_id = models.CharField('PDB ID', max_length=4, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
217
218
    pockets_nb = models.IntegerField(
        'Total number of pockets in the complex', default=1)
219
    symmetry = models.ForeignKey(Symmetry, models.CASCADE)
220
    diseases = models.ManyToManyField(Disease)
221
    family = models.ForeignKey(PpiFamily, models.CASCADE, null=True, blank=True)
222
    name = models.TextField('PPI name', null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
223

224
225
226
    def __str__(self):
        return '{} PPI, PDB:{}'.format(self.symmetry.description, self.pdb_id or 'unknown')

227
228
229
    def autofill(self):
        # name is denormalized and stored in the database to reduce SQL queries in query mode
        self.name = self.compute_name_from_protein_names()
230
231
232
233
234

    def get_ppi_bound_complexes(self):
        """
        return bound ppi complexes belonging to this ppi
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
235
        # this is the less efficient query ever seen, FIXME
236
237
        return PpiComplex.objects.filter(ppi=self, complex__in=ProteinDomainBoundComplex.objects.all())

238
    def compute_name_from_protein_names(self):
Hervé  MENAGER's avatar
Hervé MENAGER committed
239
        all_protein_names = set(
240
            [ppi_complex.complex.protein.short_name for ppi_complex in self.ppicomplex_set.all()])
Hervé  MENAGER's avatar
Hervé MENAGER committed
241
242
        bound_protein_names = set(
            [ppi_complex.complex.protein.short_name for ppi_complex in self.get_ppi_bound_complexes()])
243
244
245
246
        partner_protein_names = all_protein_names - bound_protein_names
        bound_str = ','.join(bound_protein_names)
        partner_str = ','.join(partner_protein_names)
        name = bound_str
Hervé  MENAGER's avatar
Hervé MENAGER committed
247
        if partner_str != '':
248
249
            name += ' / ' + partner_str
        return name
250

Hervé  MENAGER's avatar
Hervé MENAGER committed
251

Hervé  MENAGER's avatar
Hervé MENAGER committed
252
class PpiComplex(models.Model):
253
254
    ppi = models.ForeignKey(Ppi, models.CASCADE)
    complex = models.ForeignKey(ProteinDomainComplex, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
255
256
    cc_nb = models.IntegerField(
        'Number of copies of the complex in the PPI', default=1)
Hervé  MENAGER's avatar
Hervé MENAGER committed
257
258
259
260

    class Meta:
        verbose_name_plural = "Ppi complexes"

261
262
263
    def __str__(self):
        return 'PPI {}, Complex {} ({})'.format(self.ppi, self.complex, self.cc_nb)

Hervé  MENAGER's avatar
Hervé MENAGER committed
264

265
class Compound(models.Model):
266
267
    canonical_smile = models.TextField(
        'Canonical Smile', unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
268
269
270
271
272
    is_macrocycle = models.BooleanField('Contains one or more macrocycles')
    aromatic_ratio = models.DecimalField(
        'Aromatic ratio', max_digits=3, decimal_places=2)
    balaban_index = models.DecimalField(
        'Balaban index', max_digits=3, decimal_places=2)
273
    fsp3 = models.DecimalField('Fsp3', max_digits=3, decimal_places=2)
Hervé  MENAGER's avatar
Hervé MENAGER committed
274
275
276
277
278
279
280
281
282
283
    gc_molar_refractivity = models.DecimalField(
        'GC Molar Refractivity', max_digits=5, decimal_places=2)
    log_d = models.DecimalField(
        'LogD (Partition coefficient octanol-1/water, with pKa information)', max_digits=4, decimal_places=2)
    a_log_p = models.DecimalField(
        'ALogP (Partition coefficient octanol-1/water)', max_digits=4, decimal_places=2)
    mean_atom_vol_vdw = models.DecimalField(
        'Mean atom volume computed with VdW radii', max_digits=4, decimal_places=2)
    molecular_weight = models.DecimalField(
        'Molecular weight', max_digits=6, decimal_places=2)
284
285
286
287
    nb_acceptor_h = models.IntegerField('Number of hydrogen bond acceptors')
    nb_aliphatic_amines = models.IntegerField('Number of aliphatics amines')
    nb_aromatic_bonds = models.IntegerField('Number of aromatic bonds')
    nb_aromatic_ether = models.IntegerField('Number of aromatic ethers')
Hervé  MENAGER's avatar
Hervé MENAGER committed
288
289
290
291
    nb_aromatic_sssr = models.IntegerField(
        'Number of aromatic Smallest Set of System Rings (SSSR)')
    nb_atom = models.IntegerField('Number of atoms')
    nb_atom_non_h = models.IntegerField('Number of non hydrogen atoms')
292
293
    nb_benzene_like_rings = models.IntegerField('Number of benzene-like rings')
    nb_bonds = models.IntegerField('Number of bonds')
Hervé  MENAGER's avatar
Hervé MENAGER committed
294
295
296
297
298
299
300
301
302
303
304
305
306
307
    nb_bonds_non_h = models.IntegerField(
        'Number of bonds not involving a hydrogen')
    nb_br = models.IntegerField('Number of Bromine atoms')
    nb_c = models.IntegerField('Number of Carbon atoms')
    nb_chiral_centers = models.IntegerField('Number of chiral centers')
    nb_circuits = models.IntegerField('Number of circuits')
    nb_cl = models.IntegerField('Number of Chlorine atoms')
    nb_csp2 = models.IntegerField('Number of sp2-hybridized carbon atoms')
    nb_csp3 = models.IntegerField('Number of sp3-hybridized carbon atoms')
    nb_donor_h = models.IntegerField('Number of hydrogen bond donors')
    nb_double_bonds = models.IntegerField('Number of double bonds')
    nb_f = models.IntegerField('Number of fluorine atoms')
    nb_i = models.IntegerField('Number of iodine atoms')
    nb_multiple_bonds = models.IntegerField('Number of multiple bonds')
308
    nb_n = models.IntegerField('Number of nitrogen atoms')
Hervé  MENAGER's avatar
Hervé MENAGER committed
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
    nb_o = models.IntegerField('Number of oxygen atoms')
    nb_rings = models.IntegerField('Number of rings')
    nb_rotatable_bonds = models.IntegerField('Number of rotatable bonds')
    randic_index = models.DecimalField(
        'Randic index', max_digits=4, decimal_places=2)
    rdf070m = models.DecimalField(
        'RDF070m, radial distribution function weighted by the atomic masses at 7Å', max_digits=5, decimal_places=2)
    rotatable_bond_fraction = models.DecimalField(
        'Fraction of rotatable bonds', max_digits=3, decimal_places=2)
    sum_atom_polar = models.DecimalField(
        'Sum of atomic polarizabilities', max_digits=5, decimal_places=2)
    sum_atom_vol_vdw = models.DecimalField(
        'Sum of atom volumes computed with VdW radii', max_digits=6, decimal_places=2)
    tpsa = models.DecimalField(
        'Topological Polar Surface Area (TPSA)', max_digits=5, decimal_places=2)
    ui = models.DecimalField(
        'Unsaturation index', max_digits=4, decimal_places=2)
    wiener_index = models.IntegerField('Wiener index')
    common_name = models.CharField(
        'Common name', unique=True, max_length=20, blank=True, null=True)
    pubchem_id = models.CharField(
        'Pubchem ID', max_length=10, blank=True, null=True)
    chemspider_id = models.CharField(
        'Chemspider ID', unique=True, max_length=10, blank=True, null=True)
    chembl_id = models.CharField(
        'Chembl ID', max_length=30, blank=True, null=True)
    iupac_name = models.CharField(
        'IUPAC name', max_length=255, blank=True, null=True)
    mddr_compound = models.ForeignKey(
338
        'MDDRCompoundImport', models.CASCADE, blank=True, null=True)
339

Hervé  MENAGER's avatar
Hervé MENAGER committed
340
341
342
    @property
    def biblio_refs(self):
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
343
        return all RefCompoundBiblio related to this compound
Hervé  MENAGER's avatar
Hervé MENAGER committed
344
345
        """
        return RefCompoundBiblio.objects.filter(compound=self)
346

347
348
349
350
351
352
353
354
355
356
357
358
359
    @property
    def pfam_ids(self):
        """
        return all PFAM ids for the domain of the proteins of the bound
        complexes in the PPIs this compound has an action on
        """
        pfam_ids = set()
        for ca in self.compoundaction_set.all():
            ca.get_complexes()
            for bound_complex in ca.ppi.get_ppi_bound_complexes():
                pfam_ids.add(bound_complex.complex.domain.pfam_id)
        return pfam_ids

360
361
362
363
364
365
366
367
368
369
    @property
    def compound_action_pdb_ids(self):
        """
        return all PDB codes of the corresponding compound actions
        """
        pdb_ids = set()
        for ca in self.compoundaction_set.all():
            pdb_ids.add(ca.pdb_id)
        return pdb_ids

370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
    @property
    def hba_hbd(self):
        return self.nb_acceptor_h + self.nb_donor_h

    @property
    def lipinsky_mw(self):
        return self.molecular_weight <= 500

    @property
    def lipinsky_hba(self):
        return self.nb_acceptor_h <= 10

    @property
    def lipinsky_hbd(self):
        return self.nb_donor_h <= 5

    @property
    def lipinsky_a_log_p(self):
        return self.a_log_p <= 5

    @property
    def lipinsky_global(self):
        return int(self.lipinsky_mw) + int(self.lipinsky_hba) + \
            int(self.lipinsky_hbd) + int(self.lipinsky_a_log_p) >= 3

    @property
    def veber_hba_hbd(self):
        return self.nb_acceptor_h + self.nb_donor_h <= 12

    @property
    def veber_tpsa(self):
        return self.tpsa <= 140

    @property
    def veber_rb(self):
        return self.nb_rotatable_bonds <= 10

    @property
    def veber_global(self):
        return self.veber_rb and (self.veber_hba_hbd or self.veber_tpsa)

    @property
    def pfizer_a_log_p(self):
        return self.a_log_p <= 3

    @property
    def pfizer_tpsa(self):
        return self.tpsa >= 75

    @property
    def pfizer_global(self):
        return self.pfizer_a_log_p and self.pfizer_tpsa

423
424
    @property
    def best_pXC50_activity(self):
425
        return self.compoundactivityresult_set.aggregate(Max('activity'))['activity__max']
426
427
428
429
430
431

    @property
    def best_pXC50_compound_activity_result(self):
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
432
        return self.compoundactivityresult_set.filter(activity=best_pXC50_activity)[0]
433

434
435
436
437
438
    @property
    def le(self):
        """
        LE: Ligand Efficiency
        """
439
440
441
442
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
        return (1.37 * float(best_pXC50_activity))/self.nb_atom_non_h
443
444
445
446
447
448

    @property
    def lle(self):
        """
        LLE: Lipophilic Efficiency
        """
449
450
451
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
Hervé  MENAGER's avatar
Hervé MENAGER committed
452
        return float(best_pXC50_activity - self.a_log_p)
453

454
    @property
455
    def best_pXC50_activity_ppi_name(self):
456
        """
457
        Name of the PPI corresponding to the best PXC50 activity
458
        """
459
460
461
        best_activity_car = self.best_pXC50_compound_activity_result
        if best_activity_car is None:
            return None
462
463
464
        ppi_name = best_activity_car.test_activity_description.ppi.name
        return ppi_name

465
466
467
468
469
470
471
472
473
474
475
    @property
    def best_pXC50_activity_ppi_family(self):
        """
        Family of the PPI corresponding to the best PXC50 activity
        """
        best_activity_car = self.best_pXC50_compound_activity_result
        if best_activity_car is None:
            return None
        ppi_family = best_activity_car.test_activity_description.ppi.family.name
        return ppi_family

476
477
478
479
480
481
482
483
484
485
486
487
488
489
    @property
    def bioch_tests_count(self):
        """
        return the number of associated biochemical tests
        """
        return self.compoundactivityresult_set.all().filter(test_activity_description__test_type='BIOCH').count()

    @property
    def cell_tests_count(self):
        """
        return the number of associated cell tests
        """
        return self.compoundactivityresult_set.all().filter(test_activity_description__test_type='CELL').count()

490
491
492
493
494
495
496
    @property
    def families(self):
        """
        return the all PPI families for PPIs involved in the compound activity of the compound
        """
        return list(set([ca.ppi.family for ca in self.compoundaction_set.all()]))

497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
class CompoundTanimoto(models.Model):
    canonical_smiles = models.TextField(
        'Canonical Smile')
    compound = models.ForeignKey(Compound, models.CASCADE)
    tanimoto = models.DecimalField(
        'Tanimoto value', max_digits=5, decimal_places=4)

    class Meta:
        unique_together = (
            ('canonical_smiles', 'compound'))

def create_tanimoto(smiles_query):
    if CompoundTanimoto.objects.filter(canonical_smiles=smiles_query).count()==0:
        smiles_dict = {c.id:c.canonical_smile for c in Compound.objects.all()}
        fingerprinter = FingerPrinter("FP4")
        #1. compute tanimoto for SMILES query vs all compounds
        tanimoto_dict = fingerprinter.tanimoto_smiles(smiles_query, smiles_dict)
        #2. insert results in a table with three fields: SMILES query, compound id, tanimoto index
        cts = []
        for id_, smiles in smiles_dict.items():
            cts.append(CompoundTanimoto(canonical_smiles=smiles_query, compound=Compound.objects.get(id=id_), tanimoto=tanimoto_dict[id_]))
        CompoundTanimoto.objects.bulk_create(cts)
519

520
class PcaBiplotData(models.Model):
521
522
    pca_biplot_data = models.TextField(
        'PCA biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
523

524

525
class LeLleBiplotData(models.Model):
526
527
    le_lle_biplot_data = models.TextField(
        'LE-LLE biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
528

529

530
class MDDRActivityClass(models.Model):
531
    name = models.TextField('Activity Class', max_length=100, unique=True)
532
533
534
535
536

    class Meta:
        verbose_name_plural = "MDDR activity classes"

    def __str__(self):
537
        return self.name
538

Hervé  MENAGER's avatar
Hervé MENAGER committed
539

540
class MDDRCompoundImport(models.Model):
541
542

    MDDR_DEVELOPMENT_PHASES = (
Hervé  MENAGER's avatar
Hervé MENAGER committed
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
        ('Biological Testing', ''),
        ('Preclinical', ''),
        ('Phase III', ''),
        ('Phase II', ''),
        ('Phase I/II', ''),
        ('Phase I', ''),
        ('Launched', ''),
        ('Pre-Registered', ''),
        ('Not Applicable', ''),
        ('Discontinued', ''),
        ('Clinical', ''),
        ('Withdrawn', ''),
        ('Registered', ''),
        ('Not Determined', ''),
        ('Phase II/III', ''),
        ('IND Filed', ''),
559
560
        ('Phase 0', ''),
        ('Suspended', ''),
561
562
    )

563
    mddr_name = models.TextField('MDDR name')
Hervé  MENAGER's avatar
Hervé MENAGER committed
564
565
    dvpmt_phase = models.CharField(
        'Development phase', max_length=20, choices=MDDR_DEVELOPMENT_PHASES)
566
567
    canonical_smile = models.TextField(
        'Canonical Smile', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
568
569
    # TODO index this table on canonical_smile
    db_import_date = models.DateTimeField('MDDR release year/month')
570
    activity_classes = models.ManyToManyField(MDDRActivityClass)
Hervé  MENAGER's avatar
Hervé MENAGER committed
571

572
573
    class Meta:
        # over multiple releases of the MDDR database, the same compound can evolve in its development phase
Hervé  MENAGER's avatar
Hervé MENAGER committed
574
575
        # the same compound can have different names and development phases in
        # the same MDDR release
576
        unique_together = (('mddr_name', 'dvpmt_phase', 'canonical_smile'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
577
        verbose_name_plural = "MDDR compound imports"
Hervé  MENAGER's avatar
Hervé MENAGER committed
578

579
580
581
    def __str__(self):
        return "{}, {}".format(self.mddr_name, self.dvpmt_phase)

Hervé  MENAGER's avatar
Hervé MENAGER committed
582

583
class MDDRSimilarity(models.Model):
584
585
586
587
    canonical_smile_ippidb = models.TextField(
        'Canonical Smile for IPPIDB compound', unique=True, blank=True, null=True)
    canonical_smile_mddr = models.TextField(
        'Canonical Smile for MDDR Compound', unique=True, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
588
    tanimoto = models.DecimalField('Tanimoto', max_digits=6, decimal_places=5)
Hervé  MENAGER's avatar
Hervé MENAGER committed
589
590

    class Meta:
591
        unique_together = (('canonical_smile_ippidb', 'canonical_smile_mddr'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
592
        verbose_name_plural = "MDDR similarities"
593

Hervé  MENAGER's avatar
Hervé MENAGER committed
594

595
596
class CellLine(models.Model):
    name = models.CharField('Name', max_length=50, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
597

598
599
600
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
601

602
603
604
605
606
607
608
609
610
611
class TestActivityDescription(models.Model):
    TEST_TYPES = (
        ('BIOCH', 'Biochemical assay'),
        ('CELL', 'Cellular assay')
    )
    TEST_MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
612
613
614
615
    PROTEIN_BOUND_CONSTRUCTS = (
        ('F', 'Full length'),
        ('U', 'Unspecified')
    )
616
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
617
618
    protein_domain_bound_complex = models.ForeignKey(
        ProteinDomainBoundComplex, models.CASCADE)
619
    ppi = models.ForeignKey(Ppi, models.CASCADE, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
620
    test_name = models.CharField('Test name', max_length=100)
621
    is_primary = models.BooleanField('Is primary')
Hervé  MENAGER's avatar
Hervé MENAGER committed
622
623
    protein_bound_construct = models.CharField(
        'Protein bound construct', max_length=5, choices=PROTEIN_BOUND_CONSTRUCTS, blank=True, null=True)
624
    test_type = models.CharField('Test type', max_length=5, choices=TEST_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
625
626
627
628
    test_modulation_type = models.CharField(
        'Test modulation type', max_length=1, choices=TEST_MODULATION_TYPES)
    nb_active_compounds = models.IntegerField(
        'Total number of active compounds')
Hervé  MENAGER's avatar
Hervé MENAGER committed
629
630
    cell_line = models.ForeignKey(
        CellLine, models.CASCADE, blank=True, null=True)
631

Hervé  MENAGER's avatar
Hervé MENAGER committed
632
    def get_complexes(self):
633
634
635
636
        """
        get the complexes tested for this PPI
        depends on the modulation type
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
637
        if self.test_modulation_type == 'I':
638
            return self.ppi.ppicomplex_set.all()
639
640
        else:
            return self.ppi.get_ppi_bound_complexes()
641

642
643
644
    @property
    def protein_domain_partner_complex(self):
        for ppic in self.ppi.ppicomplex_set.all():
Hervé  MENAGER's avatar
Hervé MENAGER committed
645
            if hasattr(ppic.complex, 'proteindomainpartnercomplex'):
646
647
648
                return ppic.complex.proteindomainpartnercomplex
        return None

Hervé  MENAGER's avatar
Hervé MENAGER committed
649

650
class CompoundActivityResult(models.Model):
651
652
653
654
655
    MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
656
    ACTIVITY_TYPES = (
Hervé  MENAGER's avatar
Hervé MENAGER committed
657
658
659
660
        ('pIC50', 'pIC50 (half maximal inhibitory concentration, -log10)'),
        ('pEC50', 'pEC50 (half maximal effective concentration, -log10)'),
        ('pKd', 'pKd (dissociation constant, -log10)'),
        ('pKi', 'pKi (inhibition constant, -log10)'),
Hervé  MENAGER's avatar
Hervé MENAGER committed
661
    )
662
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
663
664
    test_activity_description = models.ForeignKey(
        TestActivityDescription, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
665
666
667
668
    activity_type = models.CharField(
        'Activity type', max_length=5, choices=ACTIVITY_TYPES)
    activity = models.DecimalField(
        'Activity', max_digits=12, decimal_places=10)
669
    inhibition_percentage = models.DecimalField(
670
        'Inhibition percentage', max_digits=3, decimal_places=0, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
671
672
    modulation_type = models.CharField(
        'Modulation type', max_length=1, choices=MODULATION_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
673
674

    class Meta:
Hervé  MENAGER's avatar
Hervé MENAGER committed
675
676
677
        unique_together = (
            ('compound', 'test_activity_description', 'activity_type'),)

678
679
    def __str__(self):
        return 'Compound activity result for {} test {} on {}'.format(self.activity_type, self.test_activity_description.id, self.compound.id)
Hervé  MENAGER's avatar
Hervé MENAGER committed
680

681
682
683
    def is_best(self):
        return self.compound.best_pXC50_compound_activity_result.id == self.id

Hervé  MENAGER's avatar
Hervé MENAGER committed
684

685
class TestCytotoxDescription(models.Model):
686
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
687
    test_name = models.CharField('Cytotoxicity test name', max_length=100)
688
    cell_line = models.ForeignKey(CellLine, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
689
690
691
    compound_concentration = models.DecimalField(
        'Compound concentration in μM', max_digits=7, decimal_places=3, blank=True, null=True)

Hervé  MENAGER's avatar
Hervé MENAGER committed
692

693
class CompoundCytotoxicityResult(models.Model):
694
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
695
696
    test_cytotoxicity_description = models.ForeignKey(
        TestCytotoxDescription, models.CASCADE)
697
    toxicity = models.BooleanField('Toxicity', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
698
699

    class Meta:
700
        unique_together = (('compound', 'test_cytotoxicity_description'),)
701

702
703
704
    def __str__(self):
        return 'Compound cytotoxicity result for test {} on {}'.format(self.test_cytotoxicity_description.id, self.compound.id)

Hervé  MENAGER's avatar
Hervé MENAGER committed
705

706
class TestPKDescription(models.Model):
707
708
709
710
711
712
    ADMINISTRATION_MODES = (
        ('IV', ''),
        ('PO', ''),
        ('IP', ''),
        ('SL', 'SL')
    )
713
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
714
    test_name = models.CharField('Pharmacokinetic test name', max_length=100)
715
    organism = models.ForeignKey(Taxonomy, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
716
717
    administration_mode = models.CharField(
        'Administration mode', max_length=2, choices=ADMINISTRATION_MODES, blank=True, null=True)
718
719
    concentration = models.DecimalField(
        'Concentration in mg/l', max_digits=7, decimal_places=3, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
720
    dose = models.DecimalField(
721
        'Dose in mg/kg', max_digits=9, decimal_places=4, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
722
723
724
    dose_interval = models.IntegerField(
        'Dose interval, in hours', blank=True, null=True)

725
726

class CompoundPKResult(models.Model):
727
728
    compound = models.ForeignKey(Compound, models.CASCADE)
    test_pk_description = models.ForeignKey(TestPKDescription, models.CASCADE)
729
    tolerated = models.NullBooleanField('Tolerated', null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
730
731
732
733
    auc = models.IntegerField(
        'Area under curve (ng.mL-1.hr)', blank=True, null=True)
    clearance = models.DecimalField(
        'Clearance (mL/hr)', max_digits=7, decimal_places=3, blank=True, null=True)
734
    c_max = models.DecimalField(
Hervé  MENAGER's avatar
Hervé MENAGER committed
735
736
737
738
739
740
741
        'Maximal concentration (ng/mL)', max_digits=7, decimal_places=3, blank=True, null=True)
    oral_bioavailability = models.IntegerField(
        'Oral Bioavailability (%F)', blank=True, null=True)
    t_demi = models.IntegerField('t½', blank=True, null=True)
    t_max = models.IntegerField('tmax', blank=True, null=True)
    voldistribution = models.DecimalField(
        'Volume distribution (Vd)', max_digits=5, decimal_places=2, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
742
743

    class Meta:
744
        unique_together = (('compound', 'test_pk_description'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
745

746
747
748
    def __str__(self):
        return 'Compound PK result for test {} on {}'.format(self.test_pk_description.id, self.compound.id)

Hervé  MENAGER's avatar
Hervé MENAGER committed
749

750
class CompoundAction(models.Model):
751
752
    ACTIVATION_MODES = (
        ('O', 'Orthosteric'),
753
754
        ('A', 'Allosteric'),
        ('U', 'Unspecified')
755
    )
756
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
757
758
    activation_mode = models.CharField(
        'Activation mode', max_length=1, choices=ACTIVATION_MODES)
759
    ppi = models.ForeignKey(Ppi, models.CASCADE)
760
    pdb_id = models.CharField('PDB ID', max_length=4, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
761
762
    nb_copy_compounds = models.IntegerField(
        'Number of copies for the compound')
Hervé  MENAGER's avatar
Hervé MENAGER committed
763
764

    class Meta:
Hervé  MENAGER's avatar
Hervé MENAGER committed
765
766
767
        unique_together = (('ppi', 'compound', 'activation_mode', 'pdb_id'),)

    def get_complexes(self):
768
769
770
771
        """
        get the complexes involved in the compound action
        which are always the bound complexes
        """
772
        return self.ppi.get_ppi_bound_complexes()
Hervé  MENAGER's avatar
Hervé MENAGER committed
773

Hervé  MENAGER's avatar
Hervé MENAGER committed
774

775
class RefCompoundBiblio(models.Model):
776
777
    compound = models.ForeignKey(Compound, models.CASCADE)
    bibliography = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
778
779
    compound_name = models.CharField(
        'Compound name in the publication', max_length=50)
Hervé  MENAGER's avatar
Hervé MENAGER committed
780
781

    class Meta:
782
        unique_together = (('compound', 'bibliography'),)