def parse_ecoli_dip(session):
with open('Ecoli/DIP.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
ids_A = row['ID interactor A'].split('|')
ids_B = row['ID interactor B'].split('|')
refseq_A, uniprotkb_A, refseq_B, uniprotkb_B = '', '', '', ''
for id in ids_A:
fields = id.split(':')
if fields[0] == 'refseq':
refseq_A = fields[1]
elif fields[0] == 'uniprotkb':
uniprotkb_A = fields[1]
for id in ids_B:
fields = id.split(':')
if fields[0] == 'refseq':
refseq_B = fields[1]
elif fields[0] == 'uniprotkb':
uniprotkb_B = fields[1]
orthologs_A, orthologs_B = [], []
if uniprotkb_A != '':
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == uniprotkb_A).all()
if (len(orthologs_A) == 0) & (refseq_A != ''):
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_refseq == refseq_A).all()
if uniprotkb_B != '':
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == uniprotkb_B).all()
if (len(orthologs_B) == 0) & (refseq_B != ''):
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_refseq == refseq_B).all()
for ortholog_A in orthologs_A:
for ortholog_B in orthologs_B:
if (ortholog_A is not None) and (ortholog_B is not None):
if ortholog_A.strain_protein == ortholog_B.strain_protein:
interactors.append(
[[ortholog_A.protein, ortholog_A.ortholog_id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
for interactor_pair in interactors:
is_new = 0
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
is_new = 1
interaction = Interaction(
strain=interactor_pair[0][0].strain,
interactors=[
interactor_pair[0][0], interactor_pair[1][0]
],
type='p-p',
ortholog_derived='fe')
session.add(interaction), session.commit()
detections, pmids, types, list = [], [], [], []
if row['Interaction detection method(s)'] != '-':
detections = row['Interaction detection method(s)'].split(
'|')
list.append(detections)
if row['Publication Identifier(s)'] != '-':
pmids = row['Publication Identifier(s)'].split('|')
list.append(pmids)
if row['Interaction type(s)'] != '-':
types = row['Interaction type(s)'].split('|')
list.append(types)
interactor_a, interactor_b = '', ''
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
for num in range(0, len(list[0])):
type = types[num].split('(')[1][:-1]
pmid = pmids[num * 2].split('pubmed:')[1]
detection = detections[num].split('(')[1][:-1]
# there are more than one pmid sometimes
reference = InteractionReference(
interaction_id=interaction.id,
detection_method=detection,
pmid=pmid,
source_db=row['Source database(s)'].split('(')[1][:-1],
interactor_a=interactor_a,
interactor_b=interactor_b)
session.add(reference)
if is_new:
if interaction.is_experimental is None:
if is_experimental_psimi(
row['Interaction detection method(s)'].
split('MI:')[1][:4]):
interaction.is_experimental = 1
else:
interaction.is_experimental = 0
elif is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'DIP').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='DIP')
session.add(source)
session.commit()
print(session.query(Interaction).count())
def parse_ecoli_mentha(session):
with open('Ecoli/PSICQUIC/mentha.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
if (row['#ID(s) interactor A'] == '-') | (row['ID(s) interactor B']
== '-'):
continue
interactors = []
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot ==
row['#ID(s) interactor A'].split(':')[1]).all()
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot ==
row['ID(s) interactor B'].split(':')[1]).all()
for ortholog_A in orthologs_A:
for ortholog_B in orthologs_B:
if (ortholog_A is not None) and (ortholog_B is not None):
if ortholog_A.strain_protein == ortholog_B.strain_protein:
interactors.append(
[[ortholog_A.protein, ortholog_A.ortholog_id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
for interactor_pair in interactors:
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
interaction = Interaction(
strain=interactor_pair[0][0].strain,
interactors=[
interactor_pair[0][0], interactor_pair[1][0]
],
type=(interactor_pair[0][0].type + '-' +
interactor_pair[1][0]),
ortholog_derived='fe')
#ask about marking ecoli ortholog interactions as experimental!!
if 'MI:' in row['Interaction detection method(s)']:
#iterate through all methods
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
session.add(interaction), session.commit()
interactor_a, interactor_b = '', ''
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
reference = InteractionReference(
interaction_id=interaction.id,
psimi_detection=row['Interaction detection method(s)'].
split('MI:')[1][:4],
detection_method=row['Interaction detection method(s)'].
split('(')[1][:-1],
pmid=row['Publication Identifier(s)'].split('pubmed:')[1],
psimi_type=row['Interaction type(s)'].split('MI:')[1][:4],
interaction_type=row['Interaction type(s)'].split(
'(')[1][:-1],
psimi_db=row['Source database(s)'].split('MI:')[1][:4],
source_db=row['Source database(s)'].split('(')[1][:-1],
confidence_score=row['Confidence value(s)'])
session.add(reference)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'mentha').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='mentha')
session.add(source)
session.commit()
print(session.query(Interaction).count())
def parse_mpidb(session):
with open('PAO1/PSICQUIC/MPIDB.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
if (row['Taxid interactor A'].split('|')[0] != 'taxid:208964(pseae)') |\
(row['Taxid interactor B'].split('|')[0] != 'taxid:208964(pseae)'):
continue
A_id = row['#ID(s) interactor A'].split(':')[1]
B_id = row['ID(s) interactor B'].split(':')[1]
if session.query(Interactor).filter(
Interactor.id == A_id).first() is not None:
interactors.append(
session.query(Interactor).filter(
Interactor.id == A_id).one())
elif session.query(Protein).filter(
Protein.uniprotkb == A_id).first() is not None:
interactors.append(
session.query(Protein).filter(
Protein.uniprotkb == A_id).one())
if session.query(Interactor).filter(
Interactor.id == B_id).first() is not None:
interactors.append(
session.query(Interactor).filter(
Interactor.id == B_id).one())
elif session.query(Protein).filter(
Protein.uniprotkb == B_id).first() is not None:
interactors.append(
session.query(Protein).filter(
Protein.uniprotkb == B_id).one())
if len(interactors) != 2: continue
homogenous = (interactors[0] == interactors[1])
interaction = session.query(Interaction).filter(
(Interaction.interactors.contains(interactors[0])),
(Interaction.interactors.contains(interactors[1])),
(Interaction.homogenous == homogenous)).first()
if interaction is None:
type = interactors[0].type + '-' + interactors[1].type
interaction = Interaction(strain='PAO1',
type=type,
homogenous=homogenous,
interactors=interactors)
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
else:
interaction.is_experimental = 0
session.add(interaction), session.commit()
else:
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
reference = InteractionReference(
interaction_id=interaction.id,
detection_method=row['Interaction detection method(s)'].split(
'(')[1][:-1],
author_ln=row['Publication 1st author(s)'].split(' ')[0],
pub_date=row['Publication 1st author(s)'].split('(')[1][:-1],
pmid=row['Publication Identifier(s)'].split('pubmed:')[1][:8],
confidence=row['Confidence value(s)'],
interaction_type=row['Interaction type(s)'].split('(')[1][:-1],
source_db=row['Source database(s)'])
session.add(reference)
for xref in row['Interaction identifier(s)'].split('|'):
xref_field = xref.split(':')
xref = session.query(InteractionXref).filter(
InteractionXref.accession == xref_field[1],
InteractionXref.interaction_id == interaction.id).first()
if xref is None:
xref = InteractionXref(interaction_id=interaction.id,
accession=xref_field[1],
data_source=xref_field[0])
session.add(xref)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'MPIDB').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='MPIDB')
session.add(source)
session.commit()
print(session.query(Interaction).count())
def parse_ecoli_uniprot(session):
with open('Ecoli/PSICQUIC/UniProt.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
orthologs_B = []
id_B = row['ID(s) interactor B'].split(':')
if id_B[0] == 'uniprotkb':
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == id_B[1]).all()
if len(orthologs_B) == 0: continue
orthologs_A = []
metabolite = None
id_A = row['#ID(s) interactor A'].split(':')
if id_A[0] == 'uniprotkb':
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == id_A[1]).all()
elif id_A[0] == 'chebi':
metabolite = session.query(Metabolite).filter(
Metabolite.chebi == id_A[1]).first()
if metabolite is None:
metabolite = Metabolite(id=id_A[1], chebi=id_A[1])
session.add(metabolite), session.commit()
for ortholog_A in orthologs_A:
for ortholog_B in orthologs_B:
if (ortholog_A is not None) and (ortholog_B is not None):
if ortholog_A.strain_protein == ortholog_B.strain_protein:
interactors.append(
[[ortholog_A.protein, ortholog_A.ortholog_id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
if metabolite is not None:
for ortholog_B in orthologs_B:
interactors.append(
[[metabolite, metabolite.id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
for interactor_pair in interactors:
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
interaction = Interaction(
strain=interactor_pair[0][0].strain,
interactors=[
interactor_pair[0][0], interactor_pair[1][0]
],
type=(interactor_pair[0][0].type + '-' +
interactor_pair[1][0]),
ortholog_derived='fe')
if 'MI:' in row['Interaction detection method(s)']:
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
session.add(interaction), session.commit()
interactor_a, interactor_b = None, None
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
reference = InteractionReference(
interaction_id=interaction.id,
psimi_detection=row['Interaction detection method(s)'].
split('MI:')[1][:4],
detection_method=row['Interaction detection method(s)'].
split('(')[1][:-1],
author_ln=row['Publication 1st author(s)'].split(' ')[0],
pub_date=row['Publication 1st author(s)'].split(
'(')[1][:-1],
pmid=row['Publication Identifier(s)'].split(
'pubmed:')[1].split('|')[0],
psimi_type=row['Interaction type(s)'].split('MI:')[1][:4],
interaction_type=row['Interaction type(s)'].split(
'(')[1][:-1],
psimi_db=row['Source database(s)'].split('MI:')[1][:4],
source_db=row['Source database(s)'].split('(')[1][:-1],
confidence_score=row['Confidence value(s)'],
interactor_a_id=interactor_a,
interactor_b_id=interactor_b)
session.add(reference)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'UniProt').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='UniProt')
session.add(source)
session.commit()
print(session.query(Interaction).count())
def parse_mentha(file, strain, taxid, session):
with open(file) as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
if ((row['Taxid interactor A'].split('|')[0] != taxid) |
(row['Taxid interactor B'].split('|')[0] != taxid)): continue
A_id = row['#ID(s) interactor A'].split(':')[1]
B_id = row['ID(s) interactor B'].split(':')[1]
if session.query(Interactor).filter(Interactor.id == A_id).first() is not None:
interactors.append(session.query(Interactor).filter(Interactor.id == A_id).one())
elif session.query(Protein).filter(Protein.uniprotkb == A_id).first() is not None:
interactors.append(session.query(Protein).filter(Protein.uniprotkb == A_id).one())
if session.query(Interactor).filter(Interactor.id == B_id).first() is not None:
interactors.append(session.query(Interactor).filter(Interactor.id == B_id).one())
elif session.query(Protein).filter(Protein.uniprotkb == B_id).first() is not None:
interactors.append(session.query(Protein).filter(Protein.uniprotkb == B_id).one())
if len(interactors) != 2: continue
homogenous = (interactors[0] == interactors[1])
interaction = session.query(Interaction).filter(Interaction.interactors.contains(interactors[0]),
Interaction.interactors.contains(interactors[1]),
Interaction.homogenous == homogenous).first()
if interaction is None:
type=(interactors[0].type + '-' + interactors[1].type)
interaction = Interaction(strain=strain, type=type, homogenous=homogenous, interactors=interactors)
if is_experimental_psimi(row['Interaction detection method(s)'].split('MI:')[1][:4]):
interaction.is_experimental = 1
else:
interaction.is_experimental = 0
session.add(interaction), session.commit()
else:
if is_experimental_psimi(row['Interaction detection method(s)'].split('MI:')[1][:4]):
interaction.is_experimental = 1
reference = InteractionReference(interaction_id=interaction.id,
detection_method=row['Interaction detection method(s)'].split('(')[1][:-1],
pmid=row['Publication Identifier(s)'].split('pubmed:')[1][:8],
interaction_type=row['Interaction type(s)'].split('(')[1][:-1],
source_db=row['Source database(s)'].split('(')[1][:-1],
confidence_score=row['Confidence value(s)'])
session.add(reference)
xref_field = row['Interaction identifier(s)'].split(':')
xref = session.query(InteractionXref).filter(InteractionXref.accession == xref_field[1],
InteractionXref.interaction_id == interaction.id).first()
if xref is None:
xref = InteractionXref(interaction_id=interaction.id, accession=xref_field[1],
data_source=xref_field[0])
session.add(xref)
source = session.query(InteractionSource).filter(InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'mentha').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id, data_source='mentha')
session.add(source)
print(session.query(Interaction).count())
def parse_irefindex(file, strain, taxid, session):
with open(file) as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
if ((row['Taxid interactor A'].split('|')[0] != taxid) |
(row['Taxid interactor B'].split('|')[0] != taxid)): continue
A_id = row['#ID(s) interactor A'].split(':')
B_id = row['ID(s) interactor B'].split(':')
if A_id[0] == 'uniprotkb':
if session.query(Interactor).filter(Interactor.id == A_id[1]).first() is not None:
interactors.append(session.query(Interactor).filter(Interactor.id == A_id[1]).one())
elif session.query(Protein).filter(Protein.uniprotkb == A_id[1]).first() is not None:
interactors.append(session.query(Protein).filter(Protein.uniprotkb == A_id[1]).one())
elif A_id[0] == 'refseq':
if session.query(Protein).filter(Protein.ncbi_acc == A_id[1]).first() is not None:
interactors.append(session.query(Protein).filter(Protein.ncbi_acc == A_id[1]).one())
if B_id[0] == 'uniprotkb':
if session.query(Interactor).filter(Interactor.id == B_id[1]).first() is not None:
interactors.append(session.query(Interactor).filter(Interactor.id == B_id[1]).one())
elif session.query(Protein).filter(Protein.uniprotkb == B_id[1]).first() is not None:
interactors.append(session.query(Protein).filter(Protein.uniprotkb == B_id[1]).one())
elif B_id[0] == 'refseq':
if session.query(Protein).filter(Protein.ncbi_acc == B_id[1]).first() is not None:
interactors.append(session.query(Protein).filter(Protein.ncbi_acc == B_id[1]).one())
if len(interactors) != 2: continue
homogenous = (interactors[0] == interactors[1])
interaction = session.query(Interaction).filter((Interaction.interactors.contains(interactors[0])),
(Interaction.interactors.contains(interactors[1])),
(Interaction.homogenous == homogenous)).first()
if interaction is None:
type = interactors[0].type + '-' + interactors[1].type
interaction = Interaction(strain=strain, type=type, homogenous=homogenous, interactors=interactors)
if row['Interaction detection method(s)'] != '-':
if is_experimental_psimi(row['Interaction detection method(s)'].split('MI:')[1][:4]):
interaction.is_experimental = 1
else:
interaction.is_experimental = 0
else:
if is_experimental_psimi(row['Interaction detection method(s)'].split('MI:')[1][:4]):
interaction.is_experimental = 1
elif (row['Interaction detection method(s)'] == '-') and (interaction.is_experimental == 0):
interaction.is_experimental = None
author, date, type= None, None, None
pmids, detections = [None], [None]
if row['Interaction detection method(s)'] != '-':
del detections[0]
for method in row['Interaction detection method(s)'].split('|'):
detections.append(method.split('(')[1][:-1])
if (row['Interaction type(s)'] != '-'):
type = row['Interaction type(s)'].split('(')[1][:-1]
if (row['Publication 1st author(s)'] != '-'):
author = row['Publication 1st author(s)'].split('-')[0][0].upper() + \
row['Publication 1st author(s)'].split('-')[0][1:]
date = row['Publication 1st author(s)'].split('-')[1]
if (row['Publication Identifier(s)'] != '-'):
del pmids[0]
for pmid in row['Publication Identifier(s)'].split('|'):
pmids.append(pmid.split('pubmed:')[1][:8])
for pmid in pmids:
for detection in detections:
reference = InteractionReference(interaction_id=interaction.id,
detection_method=detection,
author_ln=author,
pub_date=date,
pmid=pmid,
interaction_type=type,
source_db=row['Source database(s)'].split('(')[1][:-1],
confidence_score=row['Confidence value(s)'])
session.add(reference)
for xref in row['Interaction identifier(s)'].split('|'):
xref_field = xref.split(':')
xref = session.query(InteractionXref).filter(InteractionXref.accession == xref_field[1],
InteractionXref.interaction_id == interaction.id).first()
if xref is None:
xref = InteractionXref(interaction_id=interaction.id, accession=xref_field[1],
data_source=xref_field[0])
session.add(xref)
source = session.query(InteractionSource).filter(InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'iRefIndex').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id, data_source='iRefIndex')
session.add(source)
print(session.query(Interaction).count())
def parse_ecoli_bindingdb(session):
with open('Data/Ecoli/PSICQUIC/BindingDB.txt') as csvfile:
reader = csv.DictReader(csvfile)
# iterate through each interaction
for row in reader:
uniprot_protein = None
# check if interactor B has uniprot ID
if 'uniprotkb' in row['ID(s) interactor B']:
uniprot_protein = row['ID(s) interactor B'].split(
'uniprotkb:')[1].split('|')[0]
if uniprot_protein is None: continue
orthologs = []
for ecoli_ortholog in session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == uniprot_protein).all():
if ecoli_ortholog is not None:
orthologs.append(
[ecoli_ortholog.protein, ecoli_ortholog.ortholog_id])
if len(orthologs) == 0: continue
ids_metabolite = row['#ID(s) interactor A'].split('|')
chebi_metabolite, pubchem_metabolite = None, None
# check if interactor A has ChEBI id
for id in ids_metabolite:
if id.split(':')[0] == 'chebi':
chebi_metabolite = id.split(':')[1][1:-1]
metabolite = None
# if interactor A has ChEBI id, query for matching metabolite
if chebi_metabolite is not None:
metabolite = session.query(Metabolite).filter(
Metabolite.chebi == chebi_metabolite).first()
# if unable to identify metabolite based on ChEBI id, try using pubchem id
if metabolite is None:
alt_ids_metabolite = row['Alt. ID(s) interactor A'].split('|')
for id in alt_ids_metabolite:
if id.split(':')[0] == 'pubchem':
pubchem_metabolite = id.split(':')[1]
metabolite = session.query(Metabolite).filter(
Metabolite.id == pubchem_metabolite).first()
# if unable to find interactor A in database, create new metabolite
if metabolite is None:
metabolite = Metabolite(id=pubchem_metabolite,
pubchem=pubchem_metabolite,
chebi=chebi_metabolite)
session.add(metabolite), session.commit()
for interactor in orthologs:
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor[0]),
Interaction.interactors.contains(metabolite)).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
interaction = Interaction(
strain=interactor.strain,
interactors=[metabolite, interactor[0]],
type='p-m',
ortholog_derived='fe')
# should ortholog interactions be marked as experimental?
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
session.add(interaction), session.commit()
interactor_a, interactor_b = '', ''
if interaction.interactors[0] == metabolite:
interactor_a = metabolite.id
interactor_b = interactor[1]
else:
interactor_b = metabolite.id
interactor_a = interactor[1]
author, date, pmid = None, None, None
if row['Publication 1st author(s)'] != '-':
author = row['Publication 1st author(s)'].split(' ')[0]
date = row['Publication 1st author(s)'].split('(')[1][:-1]
if 'pubmed:' in row['Publication Identifier(s)']:
pmid = row['Publication Identifier(s)'].split(
'pubmed:')[1][:8]
reference = InteractionReference(
interaction_id=interaction.id,
detection_method=row['Interaction detection method(s)'].
split('(')[1][:-1],
author_ln=author,
pmid=pmid,
pub_date=date,
interaction_type=row['Interaction type(s)'].split(
'(')[1][:-1],
source_db=row['Source database(s)'].split('(')[1][:-1],
confidence=row['Confidence value(s)'].split('(')[0],
interactor_a=interactor_a,
interactor_b=interactor_b)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'BindingDB').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='BindingDB')
session.add(source)
session.add(reference)
session.commit()
def parse_ecoli_ebi_goa_nonintact(session):
with open('Ecoli/PSICQUIC/EBI-GOA-nonIntAct.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
interactors = []
uniprot_A, uniprot_B = None, None
if 'uniprotkb:' in row['#ID(s) interactor A']:
uniprot_A = row['#ID(s) interactor A'].split('uniprotkb:')[1]
if 'uniprotkb:' in row['ID(s) interactor B']:
uniprot_B = row['ID(s) interactor B'].split('uniprotkb:')[1]
if (uniprot_A is None) | (uniprot_B is None): continue
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == uniprot_A).all()
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == uniprot_B).all()
for ortholog_A in orthologs_A:
for ortholog_B in orthologs_B:
if (ortholog_A is not None) and (ortholog_B is not None):
if ortholog_A.strain_protein == ortholog_B.strain_protein:
interactors.append(
[[ortholog_A.protein, ortholog_A.ortholog_id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
for interactor_pair in interactors:
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
interaction = Interaction(
strain=interactor_pair[0][0].strain,
interactors=interactor_pair,
type='p-p',
ortholog_derived='fe')
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
session.add(interaction), session.commit()
interactor_a, interactor_b = '', ''
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
reference = InteractionReference(
interaction_id=interaction.id,
detection_method=row['Interaction detection method(s)'].
split('(')[1][:-1],
author_ln=row['Publication 1st author(s)'].split(' ')[0],
pub_date=row['Publication 1st author(s)'].split('(')[1],
pmid=row['Publication Identifier(s)'].split('pubmed:')[1],
interaction_type=row['Interaction type(s)'].split(
'(')[1][:-1],
source_db=row['Source database(s)'].split('(')[1][:-1],
interactor_a_id=row['#ID(s) interactor A'].split(':')[1],
interactor_b_id=row['ID(s) interactor B'].split(':')[1])
session.add(reference)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source ==
'EBI-GOA non-IntAct').first()
if source is None:
source = InteractionSource(
interaction_id=interaction.id,
data_source='EBI-GOA non-IntAct')
session.add(source)
session.commit()
print(session.query(Interaction).count())
def parse_ecoli_irefindex(session):
with open('Ecoli/PSICQUIC/iRefIndex.txt') as csvfile:
reader = csv.DictReader(csvfile, delimiter='\t')
for row in reader:
if (row['#ID(s) interactor A'] == '-') | (row['ID(s) interactor B']
== '-'):
continue
interactors = []
orthologs_A = []
id_A = row['#ID(s) interactor A'].split(':')
if id_A[0] == 'uniprotkb':
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == id_A[1]).all()
elif id_A[0] == 'refseq':
orthologs_A = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_refseq == id_A[1]).all()
if len(orthologs_A) == 0: continue
orthologs_B = []
id_B = row['ID(s) interactor B'].split(':')
if id_B[0] == 'uniprotkb':
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_uniprot == id_B[1]).all()
elif id_B[0] == 'refseq':
orthologs_B = session.query(OrthologEcoli).filter(
OrthologEcoli.ortholog_refseq == id_B[1]).all()
for ortholog_A in orthologs_A:
for ortholog_B in orthologs_B:
if (ortholog_A is not None) and (ortholog_B is not None):
if ortholog_A.strain_protein == ortholog_B.strain_protein:
interactors.append(
[[ortholog_A.protein, ortholog_A.ortholog_id],
[ortholog_B.protein, ortholog_B.ortholog_id]])
for interactor_pair in interactors:
homogenous = (interactor_pair[0][0] == interactor_pair[1][0])
interaction = session.query(Interaction).filter(
Interaction.interactors.contains(interactor_pair[0][0]),
Interaction.interactors.contains(interactor_pair[1][0]),
Interaction.homogenous == homogenous).first()
if interaction is not None:
if interaction.ortholog_derived is None:
interaction.ortholog_derived = 'cfe'
elif 'fe' not in interaction.ortholog_derived:
interaction.ortholog_derived += ', cfe'
session.commit()
else:
interaction = Interaction(
strain=interactor_pair[0][0].strain,
interactors=[
interactor_pair[0][0], interactor_pair[1][0]
],
type=(interactor_pair[0][0].type + '-' +
interactor_pair[1][0]),
ortholog_derived='fe')
if 'MI:' in row['Interaction detection method(s)']:
#iterate through all methods
if is_experimental_psimi(
row['Interaction detection method(s)'].split(
'MI:')[1][:4]):
interaction.is_experimental = 1
session.add(interaction), session.commit()
interactor_a, interactor_b = '', ''
if interaction.interactors[0] == interactor_pair[0][0]:
interactor_a = interactor_pair[0][1]
interactor_b = interactor_pair[1][1]
else:
interactor_b = interactor_pair[0][1]
interactor_a = interactor_pair[1][1]
author, date, psimi_type, type = None, None, None, None
confidences, psimi_detections, detections, pmids = [None], [
None
], [None], [None]
if row['Publication 1st author(s)'] != '-':
author = row['Publication 1st author(s)'].split(' ')[0]
date = row['Publication 1st author(s)'].split('(')[1][:-1]
if row['Interaction type(s)'] != '-':
type = row['Interaction type(s)'].split('(')[1][:-1]
if 'MI' in row['Interaction type(s)']:
psimi_type = row['Interaction type(s)'].split(
'MI:')[1][:4]
if row['Publication Identifier(s)'] != '-':
del pmids[0]
for pmid in row['Publication Identifier(s)'].split('|'):
pmids.append(pmid.split(':')[1])
if row['Interaction detection method(s)'] != '-':
del detections[0]
del psimi_detections[0]
for detection in row['Publication Identifier(s)'].split(
'|'):
detections.append(detection.split('(')[1][:-1])
psimi_detections.append(detection.split('MI:')[1][:4])
for pmid in pmids:
for confidence in confidences:
for (detection,
psimi_detection) in zip(detections,
psimi_detections):
reference = InteractionReference(
interaction_id=interaction.id,
psimi_detection=psimi_detection,
detection_method=detection,
author_ln=author,
date=date,
psimi_type=psimi_type,
interaction_type=type,
psimi_db=row['Source database(s)'].split(
'MI')[1][:4],
source_db=row['Source database(s)'].split(
'(')[1][:-1],
confidence=confidence,
interactor_a=interactor_a,
interactor_b=interactor_b)
session.add(reference)
source = session.query(InteractionSource).filter(
InteractionSource.interaction_id == interaction.id,
InteractionSource.data_source == 'iRefIndex').first()
if source is None:
source = InteractionSource(interaction_id=interaction.id,
data_source='iRefIndex')
session.add(source)
session.commit()
print(session.query(Interaction).count())
