models.py 28.5 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
6
from django.conf import settings
Hervé  MENAGER's avatar
Hervé MENAGER committed
7

8
from .utils import FingerPrinter, smi2inchi, smi2inchikey
9
from .ws import get_pubmed_info, get_google_patent_info, get_uniprot_info, get_taxonomy_info, get_go_info, get_pfam_info
10

Hervé  MENAGER's avatar
Hervé MENAGER committed
11

12
class AutoFillableModel(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
13

14
15
16
17
18
19
20
21
22
23
24
25
26
    """
    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()
27
28
        if 'autofill' in kwargs:
            del kwargs['autofill']
Hervé  MENAGER's avatar
Hervé MENAGER committed
29
        super(AutoFillableModel, self).save(*args, **kwargs)
30
31
32


class Bibliography(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
33

Hervé  MENAGER's avatar
Hervé MENAGER committed
34
35
36
37
    """
    Bibliography data table
    """
    SOURCES = (
38
39
40
        ('PM', 'PubMed ID'),
        ('PT', 'Patent'),
        ('DO', 'DOI ID')
Hervé  MENAGER's avatar
Hervé MENAGER committed
41
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
42
43
    source = models.CharField(
        'Bibliographic type', max_length=2, choices=SOURCES, default='PM')
Hervé  MENAGER's avatar
Hervé MENAGER committed
44
45
    id_source = models.CharField('Bibliographic ID', max_length=25)
    title = models.CharField('Title', max_length=300)
46
    journal_name = models.CharField('Journal name', max_length=50, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
47
48
    authors_list = models.CharField('Authors list', max_length=500)
    biblio_year = models.PositiveSmallIntegerField('Year')
49
    cytotox = models.BooleanField('Cytotoxicity data', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
50
51
52
53
    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
54
55
    pharmacokinetic = models.BooleanField(
        'pharmacokinetic study', default=False)
Rachel TORCHET's avatar
Rachel TORCHET committed
56
    xray = models.BooleanField('X-Ray data', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
57

58
    def autofill(self):
59
60
61
        if self.source == 'PM':
            info = get_pubmed_info(self.id_source)
        else:
62
            info = get_google_patent_info(self.id_source)
63
64
65
66
        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
67

Hervé  MENAGER's avatar
Hervé MENAGER committed
68
69
70
    class Meta:
        verbose_name_plural = "bibliographies"

71
72
    def __str__(self):
        return '{}, {}'.format(self.source, self.id_source)
73

Hervé  MENAGER's avatar
Hervé MENAGER committed
74

75
class Taxonomy(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
76
77
    taxonomy_id = models.DecimalField(
        'NCBI TaxID', unique=True, max_digits=9, decimal_places=0)
Hervé  MENAGER's avatar
Hervé MENAGER committed
78
    name = models.CharField('Organism name', max_length=200)
79

80
    def autofill(self):
81
82
83
        info = get_taxonomy_info(self.taxonomy_id)
        self.name = info['scientific_name']

84
85
86
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
87
88
    class Meta:
        verbose_name_plural = "taxonomies"
Hervé  MENAGER's avatar
Hervé MENAGER committed
89

Hervé  MENAGER's avatar
Hervé MENAGER committed
90

91
class MolecularFunction(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
92
    go_id = models.CharField('Gene Ontology ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
93
    # GO term id format: 'GO:0000000'
Hervé  MENAGER's avatar
Hervé MENAGER committed
94
95
    description = models.CharField('description', max_length=500)

96
    def autofill(self):
97
98
99
        info = get_go_info(self.go_id)
        self.description = info['label']

100
101
102
    def __str__(self):
        return self.description

Hervé  MENAGER's avatar
Hervé MENAGER committed
103

104
class Protein(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
105
    uniprot_id = models.CharField('Uniprot ID', unique=True, max_length=10)
Hervé  MENAGER's avatar
Hervé MENAGER committed
106
107
    recommended_name_long = models.CharField(
        'Uniprot Recommended Name (long)', max_length=75)
Hervé  MENAGER's avatar
Hervé MENAGER committed
108
109
110
    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)
111
    organism = models.ForeignKey('Taxonomy', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
112
113
    molecular_functions = models.ManyToManyField(MolecularFunction)

114
    def autofill(self):
115
        info = get_uniprot_info(self.uniprot_id)
116
        self.recommended_name_long = info['recommended_name']
117
118
        self.gene_name = info['gene']
        self.entry_name = info['entry_name']
119
        self.short_name = info['short_name']
120
121
122
123
124
        try:
            taxonomy = Taxonomy.objects.get(taxonomy_id=info['organism'])
        except Taxonomy.DoesNotExist:
            taxonomy = Taxonomy()
            taxonomy.taxonomy_id = info['organism']
125
            taxonomy.save(autofill=True)
126
        self.organism = taxonomy
127
        super(Protein, self).save()
128
129
130
131
132
133
        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
134
                mol_function.save(autofill=True)
135
            self.molecular_functions.add(mol_function)
136

137
138
139
    def __str__(self):
        return '{} ({})'.format(self.uniprot_id, self.recommended_name_long)

Hervé  MENAGER's avatar
Hervé MENAGER committed
140

141
class Domain(AutoFillableModel):
Hervé  MENAGER's avatar
Hervé MENAGER committed
142
143
    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
144
    pfam_description = models.CharField('Pfam Description', max_length=100)
Hervé  MENAGER's avatar
Hervé MENAGER committed
145
    domain_family = models.CharField('Domain family', max_length=25)
Hervé  MENAGER's avatar
Hervé MENAGER committed
146
147
    # TODO: what is this field? check database
    # contents
148

149
    def autofill(self):
150
151
152
        info = get_pfam_info(self.pfam_acc)
        self.pfam_id = info['id']
        self.pfam_description = info['description']
Hervé  MENAGER's avatar
Hervé MENAGER committed
153

154
155
156
    def __str__(self):
        return '{} ({}-{})'.format(self.pfam_acc, self.pfam_id, self.pfam_description)

Hervé  MENAGER's avatar
Hervé MENAGER committed
157

158
class ProteinDomainComplex(models.Model):
159
160
    protein = models.ForeignKey('Protein', models.CASCADE)
    domain = models.ForeignKey('Domain', models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
161
162
163
    ppc_copy_nb = models.IntegerField(
        'Number of copies of the protein in the complex')

Hervé  MENAGER's avatar
Hervé MENAGER committed
164
165
    class Meta:
        verbose_name_plural = "complexes"
166

167
168
169
    def __str__(self):
        return '{}-{}'.format(self.protein_id, self.domain_id)

170
171
    def name(self):
        return self.protein.short_name
Hervé  MENAGER's avatar
Hervé MENAGER committed
172

173

174
class ProteinDomainBoundComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
175
176
177
    ppp_copy_nb_per_p = models.IntegerField(
        'Number of copies of the protein in the pocket')

178
179
    class Meta:
        verbose_name_plural = "bound complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
180
181


182
class ProteinDomainPartnerComplex(ProteinDomainComplex):
Hervé  MENAGER's avatar
Hervé MENAGER committed
183

184
185
    class Meta:
        verbose_name_plural = "partner complexes"
Hervé  MENAGER's avatar
Hervé MENAGER committed
186

Hervé  MENAGER's avatar
Hervé MENAGER committed
187

188
189
190
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
191

192
193
194
    class Meta:
        verbose_name_plural = "symmetries"

195
196
197
    def __str__(self):
        return '{} ({})'.format(self.code, self.description)

198
199

class Disease(models.Model):
Hervé  MENAGER's avatar
Hervé MENAGER committed
200
    name = models.CharField('Disease', max_length=30, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
201
    # is there any database/nomenclature for diseases?
202
203
204
205

    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
206

Hervé  MENAGER's avatar
Hervé MENAGER committed
207
208
209
class PpiFamily(models.Model):
    name = models.CharField('Name', max_length=30, unique=True)

210
211
212
    class Meta:
        verbose_name_plural = "PPI Families"

Hervé  MENAGER's avatar
Hervé MENAGER committed
213
214
    def __str__(self):
        return self.name
Hervé  MENAGER's avatar
Hervé MENAGER committed
215

Hervé  MENAGER's avatar
Hervé MENAGER committed
216

217
class Ppi(AutoFillableModel):
218
    pdb_id = models.CharField('PDB ID', max_length=4, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
219
220
    pockets_nb = models.IntegerField(
        'Total number of pockets in the complex', default=1)
221
    symmetry = models.ForeignKey(Symmetry, models.CASCADE)
222
    diseases = models.ManyToManyField(Disease)
223
    family = models.ForeignKey(PpiFamily, models.CASCADE, null=True, blank=True)
224
    name = models.TextField('PPI name', null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
225

226
    def __str__(self):
227
        return 'PPI #{} on {}'.format(self.id, self.name)
228

229
230
231
    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()
232
233
234
235
236

    def get_ppi_bound_complexes(self):
        """
        return bound ppi complexes belonging to this ppi
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
237
        # this is the less efficient query ever seen, FIXME
238
239
        return PpiComplex.objects.filter(ppi=self, complex__in=ProteinDomainBoundComplex.objects.all())

240
    def compute_name_from_protein_names(self):
Hervé  MENAGER's avatar
Hervé MENAGER committed
241
        all_protein_names = set(
242
            [ppi_complex.complex.protein.short_name for ppi_complex in self.ppicomplex_set.all()])
Hervé  MENAGER's avatar
Hervé MENAGER committed
243
244
        bound_protein_names = set(
            [ppi_complex.complex.protein.short_name for ppi_complex in self.get_ppi_bound_complexes()])
245
246
247
248
        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
249
        if partner_str != '':
250
251
            name += ' / ' + partner_str
        return name
252

Hervé  MENAGER's avatar
Hervé MENAGER committed
253

Hervé  MENAGER's avatar
Hervé MENAGER committed
254
class PpiComplex(models.Model):
255
256
    ppi = models.ForeignKey(Ppi, models.CASCADE)
    complex = models.ForeignKey(ProteinDomainComplex, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
257
258
    cc_nb = models.IntegerField(
        'Number of copies of the complex in the PPI', default=1)
Hervé  MENAGER's avatar
Hervé MENAGER committed
259
260
261
262

    class Meta:
        verbose_name_plural = "Ppi complexes"

263
264
265
    def __str__(self):
        return 'PPI {}, Complex {} ({})'.format(self.ppi, self.complex, self.cc_nb)

Hervé  MENAGER's avatar
Hervé MENAGER committed
266

267
class Compound(AutoFillableModel):
268
269
    canonical_smile = models.TextField(
        'Canonical Smile', unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
270
271
272
273
274
    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)
275
    fsp3 = models.DecimalField('Fsp3', max_digits=3, decimal_places=2)
Hervé  MENAGER's avatar
Hervé MENAGER committed
276
277
278
279
280
281
282
283
284
285
    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)
286
287
288
289
    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
290
291
292
293
    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')
294
295
    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
296
297
298
299
300
301
302
303
304
305
306
307
308
309
    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')
310
    nb_n = models.IntegerField('Number of nitrogen atoms')
Hervé  MENAGER's avatar
Hervé MENAGER committed
311
312
313
    nb_o = models.IntegerField('Number of oxygen atoms')
    nb_rings = models.IntegerField('Number of rings')
    nb_rotatable_bonds = models.IntegerField('Number of rotatable bonds')
314
315
    inchi = models.TextField('InChi')
    inchikey = models.TextField('InChiKey')
Hervé  MENAGER's avatar
Hervé MENAGER committed
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
    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)
341

342
343
344
    class Meta:
       ordering = ['id']

345
346
347
    def compute_drugbank_compound_similarity(self):
        """ compute Tanimoto similarity to existing DrugBank compounds """
        self.save()
348
        # fingerprints to compute drugbank similarities are in settings module, default FP2
349
        fingerprinter = FingerPrinter(getattr(settings, "DRUGBANK_FINGERPRINTS", "FP2"))
350
351
352
353
354
355
356
357
358
        #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
359
360
361
    @property
    def biblio_refs(self):
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
362
        return all RefCompoundBiblio related to this compound
Hervé  MENAGER's avatar
Hervé MENAGER committed
363
364
        """
        return RefCompoundBiblio.objects.filter(compound=self)
365

366
367
368
369
370
371
372
373
374
375
376
377
378
    @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

379
    @property
Hervé  MENAGER's avatar
Hervé MENAGER committed
380
    def compound_action_ligand_ids(self):
381
382
383
        """
        return all PDB codes of the corresponding compound actions
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
384
        ligand_ids = set()
385
        for ca in self.compoundaction_set.all():
Hervé  MENAGER's avatar
Hervé MENAGER committed
386
387
            ligand_ids.add(ca.ligand_id)
        return ligand_ids
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
440
441
    @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

442
443
    @property
    def best_pXC50_activity(self):
444
        return self.compoundactivityresult_set.aggregate(Max('activity'))['activity__max']
445
446
447
448
449
450

    @property
    def best_pXC50_compound_activity_result(self):
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
451
        return self.compoundactivityresult_set.filter(activity=best_pXC50_activity)[0]
452

453
454
455
456
457
    @property
    def le(self):
        """
        LE: Ligand Efficiency
        """
458
459
460
461
        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
462
463
464
465
466
467

    @property
    def lle(self):
        """
        LLE: Lipophilic Efficiency
        """
468
469
470
        best_pXC50_activity = self.best_pXC50_activity
        if best_pXC50_activity is None:
            return None
Hervé  MENAGER's avatar
Hervé MENAGER committed
471
        return float(best_pXC50_activity - self.a_log_p)
472

473
    @property
474
    def best_pXC50_activity_ppi_name(self):
475
        """
476
        Name of the PPI corresponding to the best PXC50 activity
477
        """
478
479
480
        best_activity_car = self.best_pXC50_compound_activity_result
        if best_activity_car is None:
            return None
481
482
483
        ppi_name = best_activity_car.test_activity_description.ppi.name
        return ppi_name

484
485
486
487
488
489
490
491
492
493
494
    @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

495
496
497
498
499
500
501
502
503
504
505
506
507
508
    @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()

509
510
511
512
513
514
    @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()]))
515
516
517
518
519
    
    @property
    def sorted_similar_drugbank_compounds(self):
        return self.drugbankcompoundtanimoto_set.order_by('-tanimoto')
        
520
521
522
    def autofill(self):
        # compute InChi and InChiKey
        self.inchi = smi2inchi(self.canonical_smile)
523
        self.inchikey = smi2inchikey(self.canonical_smile)
524
        self.compute_drugbank_compound_similarity()
525

526
527
528
    def __str__(self):
        return 'Compound #{}'.format(self.id)

529
530
531
class CompoundTanimoto(models.Model):
    canonical_smiles = models.TextField(
        'Canonical Smile')
532
    fingerprint = models.TextField('Fingerprint')
533
534
535
536
537
538
    compound = models.ForeignKey(Compound, models.CASCADE)
    tanimoto = models.DecimalField(
        'Tanimoto value', max_digits=5, decimal_places=4)

    class Meta:
        unique_together = (
539
            ('canonical_smiles', 'fingerprint', 'compound'))
540

541
def create_tanimoto(smiles_query, fingerprint):
542
    if CompoundTanimoto.objects.filter(canonical_smiles=smiles_query, fingerprint=fingerprint).count()==0:
543
        smiles_dict = {c.id:c.canonical_smile for c in Compound.objects.all()}
544
        fingerprinter = FingerPrinter(fingerprint)
545
546
547
548
549
        #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():
550
            cts.append(CompoundTanimoto(canonical_smiles=smiles_query, fingerprint=fingerprint, compound=Compound.objects.get(id=id_), tanimoto=tanimoto_dict[id_]))
551
        CompoundTanimoto.objects.bulk_create(cts)
552

553
class PcaBiplotData(models.Model):
554
555
    pca_biplot_data = models.TextField(
        'PCA biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
556

557

558
class LeLleBiplotData(models.Model):
559
560
    le_lle_biplot_data = models.TextField(
        'LE-LLE biplot JSON data', blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
561

562

563
564
class CellLine(models.Model):
    name = models.CharField('Name', max_length=50, unique=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
565

566
567
568
    def __str__(self):
        return self.name

Hervé  MENAGER's avatar
Hervé MENAGER committed
569

570
571
572
573
574
575
576
577
578
579
class TestActivityDescription(models.Model):
    TEST_TYPES = (
        ('BIOCH', 'Biochemical assay'),
        ('CELL', 'Cellular assay')
    )
    TEST_MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
580
581
582
583
    PROTEIN_BOUND_CONSTRUCTS = (
        ('F', 'Full length'),
        ('U', 'Unspecified')
    )
584
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
585
586
    protein_domain_bound_complex = models.ForeignKey(
        ProteinDomainBoundComplex, models.CASCADE)
587
    ppi = models.ForeignKey(Ppi, models.CASCADE, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
588
    test_name = models.CharField('Test name', max_length=100)
589
    is_primary = models.BooleanField('Is primary')
Hervé  MENAGER's avatar
Hervé MENAGER committed
590
591
    protein_bound_construct = models.CharField(
        'Protein bound construct', max_length=5, choices=PROTEIN_BOUND_CONSTRUCTS, blank=True, null=True)
592
    test_type = models.CharField('Test type', max_length=5, choices=TEST_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
593
594
595
596
    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
597
598
    cell_line = models.ForeignKey(
        CellLine, models.CASCADE, blank=True, null=True)
599

Hervé  MENAGER's avatar
Hervé MENAGER committed
600
    def get_complexes(self):
601
602
603
604
        """
        get the complexes tested for this PPI
        depends on the modulation type
        """
Hervé  MENAGER's avatar
Hervé MENAGER committed
605
        if self.test_modulation_type == 'I':
606
            return self.ppi.ppicomplex_set.all()
607
608
        else:
            return self.ppi.get_ppi_bound_complexes()
609

610
611
612
    @property
    def protein_domain_partner_complex(self):
        for ppic in self.ppi.ppicomplex_set.all():
Hervé  MENAGER's avatar
Hervé MENAGER committed
613
            if hasattr(ppic.complex, 'proteindomainpartnercomplex'):
614
615
616
                return ppic.complex.proteindomainpartnercomplex
        return None

Hervé  MENAGER's avatar
Hervé MENAGER committed
617

618
class CompoundActivityResult(models.Model):
619
620
621
622
623
    MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
Hervé  MENAGER's avatar
Hervé MENAGER committed
624
    ACTIVITY_TYPES = (
Hervé  MENAGER's avatar
Hervé MENAGER committed
625
626
627
628
        ('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
629
    )
630
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
631
632
    test_activity_description = models.ForeignKey(
        TestActivityDescription, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
633
634
635
636
    activity_type = models.CharField(
        'Activity type', max_length=5, choices=ACTIVITY_TYPES)
    activity = models.DecimalField(
        'Activity', max_digits=12, decimal_places=10)
637
    inhibition_percentage = models.DecimalField(
638
        'Inhibition percentage', max_digits=3, decimal_places=0, null=True, blank=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
639
640
    modulation_type = models.CharField(
        'Modulation type', max_length=1, choices=MODULATION_TYPES)
Hervé  MENAGER's avatar
Hervé MENAGER committed
641
642

    class Meta:
Hervé  MENAGER's avatar
Hervé MENAGER committed
643
644
645
        unique_together = (
            ('compound', 'test_activity_description', 'activity_type'),)

646
647
    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
648

649
650
651
    def is_best(self):
        return self.compound.best_pXC50_compound_activity_result.id == self.id

Hervé  MENAGER's avatar
Hervé MENAGER committed
652

653
class TestCytotoxDescription(models.Model):
654
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
655
    test_name = models.CharField('Cytotoxicity test name', max_length=100)
656
    cell_line = models.ForeignKey(CellLine, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
657
658
659
    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
660

661
class CompoundCytotoxicityResult(models.Model):
662
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
663
664
    test_cytotoxicity_description = models.ForeignKey(
        TestCytotoxDescription, models.CASCADE)
665
    toxicity = models.BooleanField('Toxicity', default=False)
Hervé  MENAGER's avatar
Hervé MENAGER committed
666
667

    class Meta:
668
        unique_together = (('compound', 'test_cytotoxicity_description'),)
669

670
671
672
    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
673

674
class TestPKDescription(models.Model):
675
676
677
678
679
680
    ADMINISTRATION_MODES = (
        ('IV', ''),
        ('PO', ''),
        ('IP', ''),
        ('SL', 'SL')
    )
681
    biblio = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
682
    test_name = models.CharField('Pharmacokinetic test name', max_length=100)
683
    organism = models.ForeignKey(Taxonomy, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
684
685
    administration_mode = models.CharField(
        'Administration mode', max_length=2, choices=ADMINISTRATION_MODES, blank=True, null=True)
686
687
    concentration = models.DecimalField(
        'Concentration in mg/l', max_digits=7, decimal_places=3, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
688
    dose = models.DecimalField(
689
        'Dose in mg/kg', max_digits=9, decimal_places=4, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
690
691
692
    dose_interval = models.IntegerField(
        'Dose interval, in hours', blank=True, null=True)

693
694

class CompoundPKResult(models.Model):
695
696
    compound = models.ForeignKey(Compound, models.CASCADE)
    test_pk_description = models.ForeignKey(TestPKDescription, models.CASCADE)
697
    tolerated = models.NullBooleanField('Tolerated', null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
698
699
700
701
    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)
702
    c_max = models.DecimalField(
Hervé  MENAGER's avatar
Hervé MENAGER committed
703
704
705
706
707
708
709
        '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
710
711

    class Meta:
712
        unique_together = (('compound', 'test_pk_description'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
713

714
715
716
    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
717

718
class CompoundAction(models.Model):
719
720
    ACTIVATION_MODES = (
        ('O', 'Orthosteric'),
721
722
        ('A', 'Allosteric'),
        ('U', 'Unspecified')
723
    )
724
    compound = models.ForeignKey(Compound, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
725
726
    activation_mode = models.CharField(
        'Activation mode', max_length=1, choices=ACTIVATION_MODES)
727
    ppi = models.ForeignKey(Ppi, models.CASCADE)
728
    ligand_id = models.CharField('PDB Ligand ID', max_length=3, blank=True, null=True)
Hervé  MENAGER's avatar
Hervé MENAGER committed
729
730
    nb_copy_compounds = models.IntegerField(
        'Number of copies for the compound')
Hervé  MENAGER's avatar
Hervé MENAGER committed
731
732

    class Meta:
733
        unique_together = (('ppi', 'compound', 'activation_mode', 'ligand_id'),)
Hervé  MENAGER's avatar
Hervé MENAGER committed
734
735

    def get_complexes(self):
736
737
738
739
        """
        get the complexes involved in the compound action
        which are always the bound complexes
        """
740
        return self.ppi.get_ppi_bound_complexes()
Hervé  MENAGER's avatar
Hervé MENAGER committed
741

742
743
    def __str__(self):
        return 'Action of {} on {}'.format(self.compound, self.ppi)
Hervé  MENAGER's avatar
Hervé MENAGER committed
744

745
class RefCompoundBiblio(models.Model):
746
747
    compound = models.ForeignKey(Compound, models.CASCADE)
    bibliography = models.ForeignKey(Bibliography, models.CASCADE)
Hervé  MENAGER's avatar
Hervé MENAGER committed
748
749
    compound_name = models.CharField(
        'Compound name in the publication', max_length=50)
Hervé  MENAGER's avatar
Hervé MENAGER committed
750
751

    class Meta:
752
        unique_together = (('compound', 'bibliography'),)
753

754
class DrugBankCompound(models.Model):
755
756
757
758
759
    id = models.TextField(
        'Drugbank ID', unique=True, primary_key=True)
    common_name = models.TextField('Common name')
    canonical_smiles = models.TextField(
        'Canonical SMILES')
760
761
762
763
764
765

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)