models.py 28.4 KB
Newer Older
Hervé  MENAGER's avatar
Hervé MENAGER committed
1
from __future__ import unicode_literals
2
import operator
Hervé  MENAGER's avatar
Hervé MENAGER committed
3
4

from django.db import models
5
from django.db.models import Max
Hervé  MENAGER's avatar
Hervé MENAGER committed
6

7
from .utils import FingerPrinter, smi2inchi, smi2inchikey
8
from .ws import get_pubmed_info, get_google_patent_info, get_uniprot_info, get_taxonomy_info, get_go_info, get_pfam_info
9

Hervé  MENAGER's avatar
Hervé MENAGER committed
10

11
class AutoFillableModel(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
12

13
14
15
16
17
18
19
20
21
22
23
24
25
    """
    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()
26
27
        if 'autofill' in kwargs:
            del kwargs['autofill']
Hervé  MENAGER's avatar
Hervé MENAGER committed
28
        super(AutoFillableModel, self).save(*args, **kwargs)
29
30
31


class Bibliography(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
32

Hervé  MENAGER's avatar
Hervé MENAGER committed
33
34
35
36
    """
    Bibliography data table
    """
    SOURCES = (
37
38
39
        ('PM', 'PubMed ID'),
        ('PT', 'Patent'),
        ('DO', 'DOI ID')
Hervé  MENAGER's avatar
Hervé MENAGER committed
40
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
41
42
    source = models.CharField(
        'Bibliographic type', max_length=2, choices=SOURCES, default='PM')
Hervé  MENAGER's avatar
Hervé MENAGER committed
43
44
    id_source = models.CharField('Bibliographic ID', max_length=25)
    title = models.CharField('Title', max_length=300)
45
    journal_name = models.CharField('Journal name', max_length=50, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
46
47
    authors_list = models.CharField('Authors list', max_length=500)
    biblio_year = models.PositiveSmallIntegerField('Year')
48
    cytotox = models.BooleanField('Cytotoxicity data', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
49
50
51
52
    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
53
54
    pharmacokinetic = models.BooleanField(
        'pharmacokinetic study', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
55
    xray = models.BooleanField('X-Ray data', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
56

57
    def autofill(self):
58
59
60
        if self.source == 'PM':
            info = get_pubmed_info(self.id_source)
        else:
61
            info = get_google_patent_info(self.id_source)
62
63
64
65
        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
66

Hervé  MENAGER's avatar
Hervé MENAGER committed
67
68
69
    class Meta:
        verbose_name_plural = "bibliographies"

70
71
    def __str__(self):
        return '{}, {}'.format(self.source, self.id_source)
72

Hervé  MENAGER's avatar
Hervé MENAGER committed
73

74
class Taxonomy(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
75
76
    taxonomy_id = models.DecimalField(
        'NCBI TaxID', unique=True, max_digits=9, decimal_places=0)
Hervé  MENAGER's avatar
Hervé MENAGER committed
77
    name = models.CharField('Organism name', max_length=200)
78

79
    def autofill(self):
80
81
82
        info = get_taxonomy_info(self.taxonomy_id)
        self.name = info['scientific_name']

83
84
85
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
86
87
    class Meta:
        verbose_name_plural = "taxonomies"
Hervé  MENAGER's avatar
Hervé MENAGER committed
88

Hervé  MENAGER's avatar
Hervé MENAGER committed
89

90
class MolecularFunction(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
91
    go_id = models.CharField('Gene Ontology ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
92
    # GO term id format: 'GO:0000000'
Hervé  MENAGER's avatar
Hervé MENAGER committed
93
94
    description = models.CharField('description', max_length=500)

95
    def autofill(self):
96
97
98
        info = get_go_info(self.go_id)
        self.description = info['label']

99
100
101
    def __str__(self):
        return self.description

Hervé  MENAGER's avatar
Hervé MENAGER committed
102

103
class Protein(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
104
    uniprot_id = models.CharField('Uniprot ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
105
106
    recommended_name_long = models.CharField(
        'Uniprot Recommended Name (long)', max_length=75)
Hervé  MENAGER's avatar
Hervé MENAGER committed
107
108
109
    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)
110
    organism = models.ForeignKey('Taxonomy', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
111
112
    molecular_functions = models.ManyToManyField(MolecularFunction)

113
    def autofill(self):
114
        info = get_uniprot_info(self.uniprot_id)
115
        self.recommended_name_long = info['recommended_name']
116
117
        self.gene_name = info['gene']
        self.entry_name = info['entry_name']
118
        self.short_name = info['short_name']
119
120
121
122
123
        try:
            taxonomy = Taxonomy.objects.get(taxonomy_id=info['organism'])
        except Taxonomy.DoesNotExist:
            taxonomy = Taxonomy()
            taxonomy.taxonomy_id = info['organism']
124
            taxonomy.save(autofill=True)
125
        self.organism = taxonomy
126
        super(Protein, self).save()
127
128
129
130
131
132
        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
133
                mol_function.save(autofill=True)
134
            self.molecular_functions.add(mol_function)
135

136
137
138
    def __str__(self):
        return '{} ({})'.format(self.uniprot_id, self.recommended_name_long)

Hervé  MENAGER's avatar
Hervé MENAGER committed
139

140
class Domain(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
141
142
    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
143
    pfam_description = models.CharField('Pfam Description', max_length=100)
Hervé  MENAGER's avatar
Hervé MENAGER committed
144
    domain_family = models.CharField('Domain family', max_length=25)
Hervé  MENAGER's avatar
Hervé MENAGER committed
145
146
    # TODO: what is this field? check database
    # contents
147

148
    def autofill(self):
149
150
151
        info = get_pfam_info(self.pfam_acc)
        self.pfam_id = info['id']
        self.pfam_description = info['description']
Hervé  MENAGER's avatar
Hervé MENAGER committed
152

153
154
155
    def __str__(self):
        return '{} ({}-{})'.format(self.pfam_acc, self.pfam_id, self.pfam_description)

Hervé  MENAGER's avatar
Hervé MENAGER committed
156

157
class ProteinDomainComplex(models.Model):
158
159
    protein = models.ForeignKey('Protein', models.CASCADE)
    domain = models.ForeignKey('Domain', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
160
161
162
    ppc_copy_nb = models.IntegerField(
        'Number of copies of the protein in the complex')

Hervé  MENAGER's avatar
Hervé MENAGER committed
163
164
    class Meta:
        verbose_name_plural = "complexes"
165

166
167
168
    def __str__(self):
        return '{}-{}'.format(self.protein_id, self.domain_id)

169
170
    def name(self):
        return self.protein.short_name
Hervé  MENAGER's avatar
Hervé MENAGER committed
171

172

173
class ProteinDomainBoundComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
174
175
176
    ppp_copy_nb_per_p = models.IntegerField(
        'Number of copies of the protein in the pocket')

177
178
    class Meta:
        verbose_name_plural = "bound complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
179
180


181
class ProteinDomainPartnerComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
182

183
184
    class Meta:
        verbose_name_plural = "partner complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
185

Hervé  MENAGER's avatar
Hervé MENAGER committed
186

187
188
189
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
190

191
192
193
    class Meta:
        verbose_name_plural = "symmetries"

194
195
196
    def __str__(self):
        return '{} ({})'.format(self.code, self.description)

197
198

class Disease(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
199
    name = models.CharField('Disease', max_length=30, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
200
    # is there any database/nomenclature for diseases?
201
202
203
204

    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
205

Hervé  MENAGER's avatar
Hervé MENAGER committed
206
207
208
class PpiFamily(models.Model):
    name = models.CharField('Name', max_length=30, unique=True)

209
210
211
    class Meta:
        verbose_name_plural = "PPI Families"

Hervé  MENAGER's avatar
Hervé MENAGER committed
212
213
    def __str__(self):
        return self.name
Hervé  MENAGER's avatar
Hervé MENAGER committed
214

Hervé  MENAGER's avatar
Hervé MENAGER committed
215

216
class Ppi(AutoFillableModel):
217
    pdb_id = models.CharField('PDB ID', max_length=4, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
218
219
    pockets_nb = models.IntegerField(
        'Total number of pockets in the complex', default=1)
220
    symmetry = models.ForeignKey(Symmetry, models.CASCADE)
221
    diseases = models.ManyToManyField(Disease)
222
    family = models.ForeignKey(PpiFamily, models.CASCADE, null=True, blank=True)
223
    name = models.TextField('PPI name', null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
224

225
    def __str__(self):
226
        return 'PPI #{} on {}'.format(self.id, self.name)
227

228
229
230
    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()
231
232
233
234
235

    def get_ppi_bound_complexes(self):
        """
        return bound ppi complexes belonging to this ppi
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
236
        # this is the less efficient query ever seen, FIXME
237
238
        return PpiComplex.objects.filter(ppi=self, complex__in=ProteinDomainBoundComplex.objects.all())

239
    def compute_name_from_protein_names(self):
Hervé  MENAGER's avatar
Hervé MENAGER committed
240
        all_protein_names = set(
241
            [ppi_complex.complex.protein.short_name for ppi_complex in self.ppicomplex_set.all()])
Hervé  MENAGER's avatar
Hervé MENAGER committed
242
243
        bound_protein_names = set(
            [ppi_complex.complex.protein.short_name for ppi_complex in self.get_ppi_bound_complexes()])
244
245
246
247
        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
248
        if partner_str != '':
249
250
            name += ' / ' + partner_str
        return name
251

Hervé  MENAGER's avatar
Hervé MENAGER committed
252

Hervé  MENAGER's avatar
Hervé MENAGER committed
253
class PpiComplex(models.Model):
254
255
    ppi = models.ForeignKey(Ppi, models.CASCADE)
    complex = models.ForeignKey(ProteinDomainComplex, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
256
257
    cc_nb = models.IntegerField(
        'Number of copies of the complex in the PPI', default=1)
Hervé  MENAGER's avatar
Hervé MENAGER committed
258
259
260
261

    class Meta:
        verbose_name_plural = "Ppi complexes"

262
263
264
    def __str__(self):
        return 'PPI {}, Complex {} ({})'.format(self.ppi, self.complex, self.cc_nb)

Hervé  MENAGER's avatar
Hervé MENAGER committed
265

266
class Compound(AutoFillableModel):
267
268
    canonical_smile = models.TextField(
        'Canonical Smile', unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
269
270
271
272
273
    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)
274
    fsp3 = models.DecimalField('Fsp3', max_digits=3, decimal_places=2)
Hervé  MENAGER's avatar
Hervé MENAGER committed
275
276
277
278
279
280
281
282
283
284
    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)
285
286
287
288
    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
289
290
291
292
    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')
293
294
    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
295
296
297
298
299
300
301
302
303
304
305
306
307
308
    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')
309
    nb_n = models.IntegerField('Number of nitrogen atoms')
Hervé  MENAGER's avatar
Hervé MENAGER committed
310
311
312
    nb_o = models.IntegerField('Number of oxygen atoms')
    nb_rings = models.IntegerField('Number of rings')
    nb_rotatable_bonds = models.IntegerField('Number of rotatable bonds')
313
314
    inchi = models.TextField('InChi')
    inchikey = models.TextField('InChiKey')
Hervé  MENAGER's avatar
Hervé MENAGER committed
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
    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)
340

341
342
343
    class Meta:
       ordering = ['id']

344
345
346
347
348
349
350
351
352
353
354
355
356
    def compute_drugbank_compound_similarity(self):
        """ compute Tanimoto similarity to existing DrugBank compounds """
        self.save()
        fingerprinter = FingerPrinter("FP4")
        #1. compute tanimoto for SMILES query vs all compounds
        smiles_dict = {c.id:c.canonical_smiles for c in DrugBankCompound.objects.all()}
        tanimoto_dict = fingerprinter.tanimoto_smiles(self.canonical_smile, smiles_dict)
        tanimoto_dict = dict(sorted(tanimoto_dict.items(), key=operator.itemgetter(1), reverse=True)[:15])
        dbcts = []
        for id_, tanimoto in tanimoto_dict.items():
            dbcts.append(DrugbankCompoundTanimoto(compound=self, drugbank_compound=DrugBankCompound.objects.get(id=id_), tanimoto=tanimoto))
        DrugbankCompoundTanimoto.objects.bulk_create(dbcts)

Hervé  MENAGER's avatar
Hervé MENAGER committed
357
358
359
    @property
    def biblio_refs(self):
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
360
        return all RefCompoundBiblio related to this compound
Hervé  MENAGER's avatar
Hervé MENAGER committed
361
362
        """
        return RefCompoundBiblio.objects.filter(compound=self)
363

364
365
366
367
368
369
370
371
372
373
374
375
376
    @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

377
    @property
Hervé  MENAGER's avatar
Hervé MENAGER committed
378
    def compound_action_ligand_ids(self):
379
380
381
        """
        return all PDB codes of the corresponding compound actions
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
382
        ligand_ids = set()
383
        for ca in self.compoundaction_set.all():
Hervé  MENAGER's avatar
Hervé MENAGER committed
384
385
            ligand_ids.add(ca.ligand_id)
        return ligand_ids
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
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
    @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

440
441
    @property
    def best_pXC50_activity(self):
442
        return self.compoundactivityresult_set.aggregate(Max('activity'))['activity__max']
443
444
445
446
447
448

    @property
    def best_pXC50_compound_activity_result(self):
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
449
        return self.compoundactivityresult_set.filter(activity=best_pXC50_activity)[0]
450

451
452
453
454
455
    @property
    def le(self):
        """
        LE: Ligand Efficiency
        """
456
457
458
459
        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
460
461
462
463
464
465

    @property
    def lle(self):
        """
        LLE: Lipophilic Efficiency
        """
466
467
468
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
Hervé  MENAGER's avatar
Hervé MENAGER committed
469
        return float(best_pXC50_activity - self.a_log_p)
470

471
    @property
472
    def best_pXC50_activity_ppi_name(self):
473
        """
474
        Name of the PPI corresponding to the best PXC50 activity
475
        """
476
477
478
        best_activity_car = self.best_pXC50_compound_activity_result
        if best_activity_car is None:
            return None
479
480
481
        ppi_name = best_activity_car.test_activity_description.ppi.name
        return ppi_name

482
483
484
485
486
487
488
489
490
491
492
    @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

493
494
495
496
497
498
499
500
501
502
503
504
505
506
    @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()

507
508
509
510
511
512
    @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()]))
513
514
515
516
517
    
    @property
    def sorted_similar_drugbank_compounds(self):
        return self.drugbankcompoundtanimoto_set.order_by('-tanimoto')
        
518
519
520
    def autofill(self):
        # compute InChi and InChiKey
        self.inchi = smi2inchi(self.canonical_smile)
521
        self.inchikey = smi2inchikey(self.canonical_smile)
522
        self.compute_drugbank_compound_similarity()
523

524
525
526
    def __str__(self):
        return 'Compound #{}'.format(self.id)

527
528
529
class CompoundTanimoto(models.Model):
    canonical_smiles = models.TextField(
        'Canonical Smile')
530
    fingerprint = models.TextField('Fingerprint')
531
532
533
534
535
536
    compound = models.ForeignKey(Compound, models.CASCADE)
    tanimoto = models.DecimalField(
        'Tanimoto value', max_digits=5, decimal_places=4)

    class Meta:
        unique_together = (
537
            ('canonical_smiles', 'fingerprint', 'compound'))
538

539
def create_tanimoto(smiles_query, fingerprint):
540
    if CompoundTanimoto.objects.filter(canonical_smiles=smiles_query, fingerprint=fingerprint).count()==0:
541
        smiles_dict = {c.id:c.canonical_smile for c in Compound.objects.all()}
542
        fingerprinter = FingerPrinter(fingerprint)
543
544
545
546
547
        #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():
548
            cts.append(CompoundTanimoto(canonical_smiles=smiles_query, fingerprint=fingerprint, compound=Compound.objects.get(id=id_), tanimoto=tanimoto_dict[id_]))
549
        CompoundTanimoto.objects.bulk_create(cts)
550

551
class PcaBiplotData(models.Model):
552
553
    pca_biplot_data = models.TextField(
        'PCA biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
554

555

556
class LeLleBiplotData(models.Model):
557
558
    le_lle_biplot_data = models.TextField(
        'LE-LLE biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
559

560

561
562
class CellLine(models.Model):
    name = models.CharField('Name', max_length=50, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
563

564
565
566
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
567

568
569
570
571
572
573
574
575
576
577
class TestActivityDescription(models.Model):
    TEST_TYPES = (
        ('BIOCH', 'Biochemical assay'),
        ('CELL', 'Cellular assay')
    )
    TEST_MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
578
579
580
581
    PROTEIN_BOUND_CONSTRUCTS = (
        ('F', 'Full length'),
        ('U', 'Unspecified')
    )
582
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
583
584
    protein_domain_bound_complex = models.ForeignKey(
        ProteinDomainBoundComplex, models.CASCADE)
585
    ppi = models.ForeignKey(Ppi, models.CASCADE, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
586
    test_name = models.CharField('Test name', max_length=100)
587
    is_primary = models.BooleanField('Is primary')
Hervé  MENAGER's avatar
Hervé MENAGER committed
588
589
    protein_bound_construct = models.CharField(
        'Protein bound construct', max_length=5, choices=PROTEIN_BOUND_CONSTRUCTS, blank=True, null=True)
590
    test_type = models.CharField('Test type', max_length=5, choices=TEST_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
591
592
593
594
    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
595
596
    cell_line = models.ForeignKey(
        CellLine, models.CASCADE, blank=True, null=True)
597

Hervé  MENAGER's avatar
Hervé MENAGER committed
598
    def get_complexes(self):
599
600
601
602
        """
        get the complexes tested for this PPI
        depends on the modulation type
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
603
        if self.test_modulation_type == 'I':
604
            return self.ppi.ppicomplex_set.all()
605
606
        else:
            return self.ppi.get_ppi_bound_complexes()
607

608
609
610
    @property
    def protein_domain_partner_complex(self):
        for ppic in self.ppi.ppicomplex_set.all():
Hervé  MENAGER's avatar
Hervé MENAGER committed
611
            if hasattr(ppic.complex, 'proteindomainpartnercomplex'):
612
613
614
                return ppic.complex.proteindomainpartnercomplex
        return None

Hervé  MENAGER's avatar
Hervé MENAGER committed
615

616
class CompoundActivityResult(models.Model):
617
618
619
620
621
    MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
622
    ACTIVITY_TYPES = (
Hervé  MENAGER's avatar
Hervé MENAGER committed
623
624
625
626
        ('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
627
    )
628
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
629
630
    test_activity_description = models.ForeignKey(
        TestActivityDescription, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
631
632
633
634
    activity_type = models.CharField(
        'Activity type', max_length=5, choices=ACTIVITY_TYPES)
    activity = models.DecimalField(
        'Activity', max_digits=12, decimal_places=10)
635
    inhibition_percentage = models.DecimalField(
636
        'Inhibition percentage', max_digits=3, decimal_places=0, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
637
638
    modulation_type = models.CharField(
        'Modulation type', max_length=1, choices=MODULATION_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
639
640

    class Meta:
Hervé  MENAGER's avatar
Hervé MENAGER committed
641
642
643
        unique_together = (
            ('compound', 'test_activity_description', 'activity_type'),)

644
645
    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
646

647
648
649
    def is_best(self):
        return self.compound.best_pXC50_compound_activity_result.id == self.id

Hervé  MENAGER's avatar
Hervé MENAGER committed
650

651
class TestCytotoxDescription(models.Model):
652
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
653
    test_name = models.CharField('Cytotoxicity test name', max_length=100)
654
    cell_line = models.ForeignKey(CellLine, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
655
656
657
    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
658

659
class CompoundCytotoxicityResult(models.Model):
660
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
661
662
    test_cytotoxicity_description = models.ForeignKey(
        TestCytotoxDescription, models.CASCADE)
663
    toxicity = models.BooleanField('Toxicity', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
664
665

    class Meta:
666
        unique_together = (('compound', 'test_cytotoxicity_description'),)
667

668
669
670
    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
671

672
class TestPKDescription(models.Model):
673
674
675
676
677
678
    ADMINISTRATION_MODES = (
        ('IV', ''),
        ('PO', ''),
        ('IP', ''),
        ('SL', 'SL')
    )
679
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
680
    test_name = models.CharField('Pharmacokinetic test name', max_length=100)
681
    organism = models.ForeignKey(Taxonomy, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
682
683
    administration_mode = models.CharField(
        'Administration mode', max_length=2, choices=ADMINISTRATION_MODES, blank=True, null=True)
684
685
    concentration = models.DecimalField(
        'Concentration in mg/l', max_digits=7, decimal_places=3, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
686
    dose = models.DecimalField(
687
        'Dose in mg/kg', max_digits=9, decimal_places=4, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
688
689
690
    dose_interval = models.IntegerField(
        'Dose interval, in hours', blank=True, null=True)

691
692

class CompoundPKResult(models.Model):
693
694
    compound = models.ForeignKey(Compound, models.CASCADE)
    test_pk_description = models.ForeignKey(TestPKDescription, models.CASCADE)
695
    tolerated = models.NullBooleanField('Tolerated', null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
696
697
698
699
    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)
700
    c_max = models.DecimalField(
Hervé  MENAGER's avatar
Hervé MENAGER committed
701
702
703
704
705
706
707
        '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
708
709

    class Meta:
710
        unique_together = (('compound', 'test_pk_description'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
711

712
713
714
    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
715

716
class CompoundAction(models.Model):
717
718
    ACTIVATION_MODES = (
        ('O', 'Orthosteric'),
719
720
        ('A', 'Allosteric'),
        ('U', 'Unspecified')
721
    )
722
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
723
724
    activation_mode = models.CharField(
        'Activation mode', max_length=1, choices=ACTIVATION_MODES)
725
    ppi = models.ForeignKey(Ppi, models.CASCADE)
726
    ligand_id = models.CharField('PDB Ligand ID', max_length=3, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
727
728
    nb_copy_compounds = models.IntegerField(
        'Number of copies for the compound')
Hervé  MENAGER's avatar
Hervé MENAGER committed
729
730

    class Meta:
731
        unique_together = (('ppi', 'compound', 'activation_mode', 'ligand_id'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
732
733

    def get_complexes(self):
734
735
736
737
        """
        get the complexes involved in the compound action
        which are always the bound complexes
        """
738
        return self.ppi.get_ppi_bound_complexes()
Hervé  MENAGER's avatar
Hervé MENAGER committed
739

740
741
    def __str__(self):
        return 'Action of {} on {}'.format(self.compound, self.ppi)
Hervé  MENAGER's avatar
Hervé MENAGER committed
742

743
class RefCompoundBiblio(models.Model):
744
745
    compound = models.ForeignKey(Compound, models.CASCADE)
    bibliography = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
746
747
    compound_name = models.CharField(
        'Compound name in the publication', max_length=50)
Hervé  MENAGER's avatar
Hervé MENAGER committed
748
749

    class Meta:
750
        unique_together = (('compound', 'bibliography'),)
751

752
class DrugBankCompound(models.Model):
753
754
755
756
757
    id = models.TextField(
        'Drugbank ID', unique=True, primary_key=True)
    common_name = models.TextField('Common name')
    canonical_smiles = models.TextField(
        'Canonical SMILES')
758
759
760
761
762
763

class DrugbankCompoundTanimoto(models.Model):
    compound = models.ForeignKey(Compound, models.CASCADE)
    drugbank_compound = models.ForeignKey(DrugBankCompound, models.CASCADE)
    tanimoto = models.DecimalField(
        'Tanimoto value', max_digits=5, decimal_places=4)