models.py

from __future__ import unicode_literals

from django.db import models

from .ws import get_pubmed_info, get_epo_info, get_uniprot_info, get_taxonomy_info, get_go_info, get_pfam_info

class Bibliography(models.Model):
    """
    Bibliography data table
    """
    SOURCES = (
        ('PM', 'PubMed article'),
        ('PT', 'Patent')
    )
    source = models.CharField('Bibliographic type', max_length=2, choices=SOURCES)
    id_source = models.CharField('Bibliographic ID', max_length=25)
    title = models.CharField('Title', max_length=300)
    journal_name = models.CharField('Journal name', max_length=50, null=True)
    authors_list = models.CharField('Authors list', max_length=500)
    biblio_year = models.PositiveSmallIntegerField('Year')
    cytotox = models.BooleanField('Cytotoxicity data', default=False)
    in_silico = models.BooleanField('in silico study performed', default=False)
    in_vitro = models.BooleanField('in vitro study performed', default=False)
    in_vivo = models.BooleanField('in vivo study performed', default=False)
    in_cellulo = models.BooleanField('in cellulo study performed', default=False)
    pharmacokinetic = models.BooleanField('pharmacokinetic study performed', default=False)
    xray = models.BooleanField('contains xray data', default=False)

    def save(self, *args, **kwargs):
        if self.source == 'PM':
            info = get_pubmed_info(self.id_source)
        else:
            info = get_epo_info(self.id_source)
        self.title = info['title']
        self.journal_name = info['journal_name']
        self.authors_list = info['authors_list']
        self.biblio_year = info['biblio_year']
        super(Bibliography, self).save(*args, **kwargs)

    class Meta:
        verbose_name_plural = "bibliographies"



class Taxonomy(models.Model):
    taxonomy_id = models.DecimalField('NCBI TaxID', unique=True, max_digits=9, decimal_places=0)
    name = models.CharField('Organism name', max_length=200)

    def save(self, *args, **kwargs):
        info = get_taxonomy_info(self.taxonomy_id)
        self.name = info['scientific_name']
        super(Taxonomy, self).save(*args, **kwargs)

    def __str__(self):
        return self.name

    class Meta:
        verbose_name_plural = "taxonomies"

class MolecularFunction(models.Model):
    go_id = models.CharField('Gene Ontology ID', unique=True, max_length=10) # GO term id format: 'GO:0000000' 
    description = models.CharField('description', max_length=500)

    def save(self, *args, **kwargs):
        info = get_go_info(self.go_id)
        self.description = info['label']
        super(MolecularFunction, self).save(*args, **kwargs)

    def __str__(self):
        return self.description

class Protein(models.Model):
    uniprot_id = models.CharField('Uniprot ID', unique=True, max_length=10)
    recommended_name_long = models.CharField('Uniprot Recommended Name (long)', max_length=75)
    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)
    organism = models.ForeignKey('Taxonomy')
    molecular_functions = models.ManyToManyField(MolecularFunction)

    def save(self, *args, **kwargs):
        info = get_uniprot_info(self.uniprot_id)
        self.recommended_name_long = info['recommended_name']
        self.gene_name = info['gene']
        self.entry_name = info['entry_name']
        try:
            taxonomy = Taxonomy.objects.get(taxonomy_id=info['organism'])
        except Taxonomy.DoesNotExist:
            taxonomy = Taxonomy()
            taxonomy.taxonomy_id = info['organism']
            taxonomy.save()
        self.organism = taxonomy
        super(Protein, self).save(*args, **kwargs)
        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
                mol_function.save()
            self.molecular_functions.add(mol_function)

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

class Domain(models.Model):
    pfam_acc = models.CharField('Pfam Accession', max_length=10, unique=True) 
    pfam_id = models.CharField('Pfam Family Identifier', max_length=20) 
    pfam_description = models.CharField('Pfam Description', max_length=100)
    domain_family = models.CharField('Domain family', max_length=25)  #TODO: what is this field? check database contents

    def save(self, *args, **kwargs):
        info = get_pfam_info(self.pfam_acc)
        self.pfam_id = info['id']
        self.pfam_description = info['description']
        super(Domain, self).save(*args, **kwargs)

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

class ProteinDomainComplex(models.Model):
    protein = models.ForeignKey('Protein')
    domain = models.ForeignKey('Domain')
    ppc_copy_nb = models.IntegerField('Number of copies of the protein in the complex')
    
    class Meta:
        verbose_name_plural = "complexes"

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

class ProteinDomainBoundComplex(ProteinDomainComplex):
    ppp_copy_nb_per_p = models.IntegerField('Number of copies of the protein in the pocket')
    pockets_nb = models.IntegerField('Total number of pockets in the complex')
    class Meta:
        verbose_name_plural = "bound complexes"
    
class ProteinDomainPartnerComplex(ProteinDomainComplex):
    class Meta:
        verbose_name_plural = "partner complexes"

class Symmetry(models.Model):
    code = models.CharField('Symmetry code', max_length=2)
    description = models.CharField('Description', max_length=300)

    class Meta:
        verbose_name_plural = "symmetries"

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


class Disease(models.Model):
    name = models.CharField('Disease', max_length=30, unique=True) # is there any database/nomenclature for diseases?

    def __str__(self):
        return self.name

class Ppi(models.Model):
    pdb_id = models.CharField('PDB ID', max_length=4, null=True)
    symmetry = models.ForeignKey(Symmetry)
    diseases = models.ManyToManyField(Disease)

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

    def get_ppi_complexes(self):
        """
        return all ppi complexes belonging to this ppi
        """
        return PpiComplex.objects.filter(ppi=self)

    def get_ppi_bound_complexes(self):
        """
        return bound ppi complexes belonging to this ppi
        """
        #this is the less efficient query ever seen, FIXME
        return PpiComplex.objects.filter(ppi=self, complex__in=ProteinDomainBoundComplex.objects.all())


class PpiComplex(models.Model):
    ppi = models.ForeignKey(Ppi)
    complex = models.ForeignKey(ProteinDomainComplex)
    cc_nb = models.IntegerField('Number of copies of the complex in the PPI', default=1)

    class Meta:
        verbose_name_plural = "Ppi complexes"

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


class Compound(models.Model):
    canonical_smile = models.CharField('Canonical Smile', unique=True, max_length=500)
    #TODO index this table on canonical_smile
    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)  
    fsp3 = models.DecimalField('Fsp3', max_digits=3, decimal_places=2)
    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)
    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')
    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') 
    nb_benzene_like_rings = models.IntegerField('Number of benzene-like rings')
    nb_bonds = models.IntegerField('Number of bonds')
    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')  
    nb_n = models.IntegerField('Number of nitrogen atoms')
    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', unique=True, 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', unique=True, max_length=30, blank=True, null=True)  
    iupac_name = models.CharField('IUPAC name', unique=True, max_length=255, blank=True, null=True)  
    mddr_compound = models.ForeignKey('MDDRCompoundImport', blank=True, null=True)  


class MDDRCompoundImport(models.Model):
    mddr_compound_id = models.IntegerField('MDDR compound ID')  
    mddr_name = models.CharField('MDDR name', max_length=40)  
    dvpmt_phase = models.CharField('Development phase', max_length=20)  
    canonical_smile = models.CharField('Canonical Smile', max_length=500, unique=True, blank=True, null=True)  
    #TODO index this table on canonical_smile
    db_import_date = models.DecimalField('MDDR release year/month', max_digits=6, decimal_places=0)  

    class Meta:
        # over multiple releases of the MDDR database, the same compound can evolve in its development phase
        # the same compound can have different names and development phases in the same MDDR release
        unique_together = (('mddr_compound_id', 'mddr_name', 'dvpmt_phase'),)
        verbose_name_plural = "MDDR compound imports"


class MDDRCompoundActivityClass(models.Model):
    mddr_compound = models.ForeignKey(MDDRCompoundImport)  
    activity_class = models.CharField('Activity Class', max_length=100)  

    class Meta:
        unique_together = (('mddr_compound', 'activity_class'),)
        verbose_name_plural = "MDDR compound activity classes"


class MDDRSimilarity(models.Model):
    canonical_smile_ippidb = models.CharField('Canonical Smile for IPPIDB compound', max_length=500, unique=True, blank=True, null=True)
    canonical_smile_mddr = models.CharField('Canonical Smile for MDDR Compound', max_length=500, unique=True, blank=True, null=True)
    tanimoto = models.DecimalField('Tanimoto', max_digits=6, decimal_places=5)  

    class Meta:
        unique_together = (('canonical_smile_ippidb', 'canonical_smile_mddr'),)
        verbose_name_plural = "MDDR similarities"

class CellLine(models.Model):
    name = models.CharField('Name', max_length=50, unique=True)

class TestActivityDescription(models.Model):
    TEST_TYPES = (
        ('BIOCH', 'Biochemical assay'),
        ('CELL', 'Cellular assay')
    )
    TEST_MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
    biblio = models.ForeignKey(Bibliography)
    ppi = models.ForeignKey(Ppi, blank=True, null=True)  
    test_name = models.CharField('Test name', max_length=100)  
    test_type = models.CharField('Test type', max_length=5, choices=TEST_TYPES)
    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')  
    cell_line = models.ForeignKey(CellLine)

    def get_complexes(self):
        """
        get the complexes tested for this PPI
        depends on the modulation type
        """
        if self.test_modulation_type=='I':
            return self.ppi.get_ppi_complexes()
        else:
            return self.ppi.get_ppi_bound_complexes()

class CompoundActivityResult(models.Model):
    MODULATION_TYPES = (
        ('B', 'Binding'),
        ('I', 'Inhibition'),
        ('S', 'Stabilization')
    )
    ACTIVITY_TYPES = (
        ('pIC50','pIC50 (half maximal inhibitory concentration, -log10)'),
        ('pEC50','pEC50 (half maximal effective concentration, -log10)'),
        ('pKd','pKd (dissociation constant, -log10)'),
        ('pKi','pKi (inhibition constant, -log10)'),
    )
    compound = models.ForeignKey(Compound)  
    test_activity_description = models.ForeignKey(TestActivityDescription)  
    activity_type = models.CharField('Activity type', max_length=5, choices=ACTIVITY_TYPES)
    activity = models.DecimalField('Activity', max_digits=12, decimal_places=10)  
    modulation_type = models.CharField('Modulation type', max_length=1, choices=MODULATION_TYPES)  

    class Meta:
        unique_together = (('compound', 'test_activity_description', 'activity_type'),)

class TestCytotoxDescription(models.Model):
    biblio = models.ForeignKey(Bibliography)  
    test_name = models.CharField('Cytotoxicity test name', max_length=100)  
    cell_line = models.ForeignKey(CellLine)
    compound_concentration = models.DecimalField('Compound concentration in μM', max_digits=7, decimal_places=3, blank=True, null=True)  

class CompoundCytotoxicityResult(models.Model):
    compound = models.ForeignKey(Compound)  
    test_cytotoxicity_description = models.ForeignKey(TestCytotoxDescription)
    toxicity = models.BooleanField('Toxicity', default=False)

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

class AdministrationMode(models.Model):
    name = models.CharField('Administration mode', max_length=20, blank=True, null=True)

class TestPKDescription(models.Model):
    biblio = models.ForeignKey(Bibliography)  
    test_name = models.CharField('Pharmacokinetic test name', max_length=100)  
    organism = models.ForeignKey(Taxonomy)
    administration_mode = models.ForeignKey(AdministrationMode, blank=True, null=True)  
    dose = models.DecimalField('Dose in mg/kg', max_digits=7, decimal_places=4, blank=True, null=True)  
    dose_interval = models.IntegerField('Dose interval, in hours', blank=True, null=True)  

class CompoundPKResult(models.Model):
    compound = models.ForeignKey(Compound)  
    test_pk_description = models.ForeignKey(TestPKDescription)
    tolerated = models.NullBooleanField('Tolerated', null=True)
    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)  
    cmax = models.DecimalField('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)  

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


class CompoundAction(models.Model):
    ACTIVATION_MODES = (
        ('O', 'Orthosteric'),
        ('A', 'Allosteric')
    )
    compound = models.ForeignKey(Compound)  
    activation_mode = models.CharField('Activation mode', max_length=1, choices=ACTIVATION_MODES)  
    ppi = models.ForeignKey(Ppi)
    pdb_id = models.CharField('PDB ID', max_length=4)
    nb_copy_compounds = models.IntegerField('Number of copies for the compound')  

    class Meta:
        unique_together = (('ppi', 'compound', 'activation_mode', 'pdb_id'),)

    def get_complexes(self):
        """
        get the complexes involved in the compound action
        which are always the bound complexes
        """
        return ppi.get_ppi_bound_complexes()

class RefCompoundBiblio(models.Model):
    compound = models.ForeignKey(Compound)  
    bibliography = models.ForeignKey(Bibliography)  
    compound_name = models.CharField('Compound name in the publication', max_length=50)  

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