def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="signorConfig.yml",
data_type='databases')
directory = os.path.join(databases_directory, "SIGNOR")
builder_utils.checkDirectory(directory)
url = config['url']
modifications = config['modifications']
amino_acids = config['amino_acids']
accronyms = config['accronyms']
entities_header = config['entities_header']
relationships_headers = config['rel_headers']
entities = set()
relationships = defaultdict(set)
filename = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
entities, relationships = parse_substrates(filename, modifications,
accronyms, amino_acids)
return entities, relationships, entities_header, relationships_headers
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="smpdbConfig.yml",
data_type='databases')
urls = config['smpdb_urls']
entities = set()
relationships = defaultdict(set)
entities_header = config['pathway_header']
relationships_headers = config['relationships_header']
directory = os.path.join(databases_directory, "SMPDB")
builder_utils.checkDirectory(directory)
for dataset in urls:
url = urls[dataset]
file_name = url.split('/')[-1]
if download:
builder_utils.downloadDB(url, directory)
zipped_file = os.path.join(directory, file_name)
with zipfile.ZipFile(zipped_file) as rf:
if dataset == "pathway":
entities = parsePathways(config, rf)
elif dataset == "protein":
relationships.update(parsePathwayProteinRelationships(rf))
elif dataset == "metabolite":
relationships.update(
parsePathwayMetaboliteDrugRelationships(rf))
builder_utils.remove_directory(directory)
return entities, relationships, entities_header, relationships_headers
def parser(databases_directory, drug_source, download=True):
config = builder_utils.get_config(config_name="siderConfig.yml", data_type='databases')
url = config['SIDER_url']
header = config['header']
output_file = 'sider_has_side_effect.tsv'
drugmapping = mp.getSTRINGMapping(source=drug_source, download=download, db="STITCH")
phenotypemapping = mp.getMappingFromOntology(ontology="Phenotype", source=config['SIDER_source'])
relationships = set()
directory = os.path.join(databases_directory, "SIDER")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
associations = gzip.open(fileName, 'r')
for line in associations:
data = line.decode('utf-8').rstrip("\r\n").split("\t")
drug = re.sub(r'CID\d', 'CIDm', data[0])
se = data[2]
evidence_from = str(data[3])
#freq = data[4]
#lower_bound = data[5]
#upper_bound = data[6]
if se.lower() in phenotypemapping and drug in drugmapping:
for d in drugmapping[drug]:
p = phenotypemapping[se.lower()]
relationships.add((d, p, "HAS_SIDE_EFFECT", "SIDER", se, evidence_from))
associations.close()
return (relationships, header, output_file, drugmapping, phenotypemapping)
def parserIndications(databases_directory, drugMapping, phenotypeMapping, download=True):
config = builder_utils.get_config(config_name="siderConfig.yml", data_type='databases')
url = config['SIDER_indications']
header = config['indications_header']
output_file = 'sider_is_indicated_for.tsv'
relationships = set()
directory = os.path.join(databases_directory, "SIDER")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
associations = gzip.open(fileName, 'r')
for line in associations:
data = line.decode('utf-8').rstrip("\r\n").split("\t")
drug = re.sub(r'CID\d', 'CIDm', data[0])
se = data[1]
evidence = data[2]
if se.lower() in phenotypeMapping and drug in drugMapping:
for d in drugMapping[drug]:
p = phenotypeMapping[se.lower()]
relationships.add((d, p, "IS_INDICATED_FOR", evidence, "SIDER", se))
associations.close()
builder_utils.remove_directory(directory)
return (relationships, header, output_file)
def parser(projectId):
data = {}
config = builder_utils.get_config(config_name="wes.yml",
data_type='experiments')
directory = '../../../data/experiments/PROJECTID/wes/'
if 'directory' in config:
directory = config['directory']
directory = directory.replace('PROJECTID', projectId)
wes_data = parseWESDataset(projectId, config, directory)
if wes_data is not None:
somatic_mutations = pd.DataFrame()
for sample in wes_data:
entities, variantRows, sampleRows, geneRows, chrRows = extractWESRelationships(
wes_data[sample], config)
data[('somatic_mutation_known_variant', 'w')] = variantRows
data[('somatic_mutation_sample', 'w')] = sampleRows
data[('somatic_mutation_gene', 'w')] = geneRows
data[('somatic_mutation_chromosome', 'w')] = chrRows
if somatic_mutations.empty:
somatic_mutations = entities
else:
new = set(entities.index).difference(
set(somatic_mutations.index))
somatic_mutations = somatic_mutations.append(
entities.loc[new, :], ignore_index=False)
somatic_mutations = somatic_mutations.reset_index()
data[('somatic_mutation', 'w')] = somatic_mutations
return data
def parser(databases_directory,
import_directory,
download=True,
updated_on=None):
config = builder_utils.get_config(config_name="uniprotConfig.yml",
data_type='databases')
relationships_header = config['relationships_header']
parse_release_notes(databases_directory, config, download)
stats = parse_fasta(databases_directory,
config,
import_directory,
download=download,
updated_on=updated_on)
#Proteins
stats.update(
parse_idmapping_file(databases_directory,
config,
import_directory,
download=download,
updated_on=updated_on))
#Peptides
entities, relationships = parseUniProtPeptides(config, databases_directory,
download)
entities_header = config['peptides_header']
output_file = os.path.join(import_directory, "Peptide.tsv")
stats.update(
print_single_file(entities,
entities_header,
output_file,
"entity",
"Peptide",
is_first=True,
updated_on=updated_on))
stats.update(
print_multiple_relationships_files(relationships,
relationships_header,
import_directory,
is_first=True,
updated_on=updated_on))
#Variants
stats.update(
parseUniProtVariants(config, databases_directory, import_directory,
download))
#Gene ontology annotation
relationships = parseUniProtAnnotations(config, databases_directory,
download)
relationships_header = config['go_header']
stats.update(
print_multiple_relationships_files(relationships,
relationships_header,
import_directory,
is_first=True,
updated_on=updated_on))
directory = os.path.join(databases_directory, "UniProt")
builder_utils.remove_directory(directory)
return stats
def parser(databases_directory, importDirectory, download=True):
config = builder_utils.get_config(config_name="jensenlabConfig.yml",
data_type='databases')
outputfileName = "Publications.tsv"
url = config['db_url']
ifile = config['organisms_file']
organisms = str(config['organisms'])
directory = os.path.join(databases_directory, "Jensenlab")
builder_utils.checkDirectory(os.path.join(directory, "textmining"))
if download:
builder_utils.downloadDB(url.replace("FILE", ifile),
os.path.join(directory, "textmining"))
ifile = os.path.join(directory, os.path.join("textmining", ifile))
valid_pubs = read_valid_pubs(organisms, ifile)
entities, header = parse_PMC_list(config,
os.path.join(directory, "textmining"),
download=download,
valid_pubs=valid_pubs)
num_entities = len(entities)
outputfile = os.path.join(importDirectory, outputfileName)
builder_utils.write_entities(entities, header, outputfile)
entities = None
for qtype in config['db_mentions_types']:
parse_mentions(config, directory, qtype, importDirectory, download)
builder_utils.remove_directory(os.path.join(directory, "textmining"))
return (num_entities, outputfile)
def parser(databases_directory, download=True):
relationships = defaultdict(set)
config = builder_utils.get_config(config_name="disgenetConfig.yml",
data_type='databases')
files = config['disgenet_files']
mapping_files = config['disgenet_mapping_files']
url = config['disgenet_url']
directory = os.path.join(databases_directory, "disgenet")
builder_utils.checkDirectory(directory)
header = config['disgenet_header']
output_file = 'disgenet_associated_with.tsv'
if download:
for f in files:
builder_utils.downloadDB(url + files[f], directory)
for f in mapping_files:
builder_utils.downloadDB(url + mapping_files[f], directory)
proteinMapping = readDisGeNetProteinMapping(config, directory)
diseaseMapping = readDisGeNetDiseaseMapping(config, directory)
for f in files:
first = True
associations = gzip.open(os.path.join(directory, files[f]), 'r')
dtype, atype = f.split('_')
if dtype == 'gene':
idType = "Protein"
scorePos = 9
if dtype == 'variant':
idType = "Transcript"
scorePos = 5
for line in associations:
if first:
first = False
continue
try:
data = line.decode('utf-8').rstrip("\r\n").split("\t")
geneId = str(int(data[0]))
#disease_specificity_index = data[2]
#disease_pleiotropy_index = data[3]
diseaseId = data[4]
score = float(data[scorePos])
pmids = data[13]
source = data[-1]
if geneId in proteinMapping:
for identifier in proteinMapping[geneId]:
if diseaseId in diseaseMapping:
for code in diseaseMapping[diseaseId]:
code = "DOID:" + code
relationships[idType].add(
(identifier, code, "ASSOCIATED_WITH",
score, atype, "DisGeNet: " + source,
pmids))
except UnicodeDecodeError:
continue
associations.close()
builder_utils.remove_directory(directory)
return (relationships, header, output_file)
def parser(projectId, type='clinical'):
data = {}
experiments_directory = ckg_utils.read_ckg_config(
key='experiments_directory')
project_directory = os.path.join(experiments_directory,
'PROJECTID/project/')
clinical_directory = os.path.join(experiments_directory,
'PROJECTID/clinical/')
design_directory = os.path.join(experiments_directory,
'PROJECTID/experimental_design/')
project_directory = project_directory.replace('PROJECTID', projectId)
clinical_directory = clinical_directory.replace('PROJECTID', projectId)
design_directory = design_directory.replace('PROJECTID', projectId)
config = builder_utils.get_config(config_name="clinical.yml",
data_type='experiments')
if type == 'project':
project_dfs = project_parser(projectId, config, project_directory)
data.update(project_dfs)
elif type == 'experimental_design':
design_dfs = experimental_design_parser(projectId, config,
design_directory)
data.update(design_dfs)
elif type == 'clinical':
clinical_dfs = clinical_parser(projectId, config, clinical_directory)
data.update(clinical_dfs)
return data
def parser(databases_directory, download=True):
relationships = defaultdict(set)
directory = os.path.join(databases_directory, "FooDB")
builder_utils.checkDirectory(directory)
config = builder_utils.get_config(config_name="foodbConfig.yml",
data_type='databases')
database_url = config['database_url']
entities_header = config['entities_header']
relationships_headers = config['relationships_headers']
tar_fileName = os.path.join(directory, database_url.split('/')[-1])
if download:
builder_utils.downloadDB(database_url, directory)
contents = {}
food = set()
compounds = {}
try:
tf = tarfile.open(tar_fileName, 'r')
file_content = tf.getnames()
tar_dir = file_content[1]
tf.extractall(path=directory)
tf.close()
for file_name in config['files']:
path = os.path.join(directory, os.path.join(tar_dir, file_name))
with open(path, 'r', encoding="utf-8", errors='replace') as f:
if file_name == "Content.csv":
contents = parseContents(f)
elif file_name == "Food.csv":
food, mapping = parseFood(f)
elif file_name == "Compound.csv":
compounds = parseCompounds(f)
for food_id, compound_id in contents:
if compound_id in compounds:
compound_code = compounds[compound_id].replace(
"HMDB", "HMDB00")
relationships[("food", "has_content"
)].add((food_id, compound_code, "HAS_CONTENT") +
contents[(food_id, compound_id)])
mp.reset_mapping(entity="Food")
with open(os.path.join(directory, "mapping.tsv"),
'w',
encoding='utf-8') as out:
for food_id in mapping:
for alias in mapping[food_id]:
out.write(str(food_id) + "\t" + str(alias) + "\n")
mp.mark_complete_mapping(entity="Food")
except tarfile.ReadError as err:
raise Exception("Error importing database FooDB.\n {}".format(err))
builder_utils.remove_directory(directory)
return food, relationships, entities_header, relationships_headers
def get_configuration(processing_tool, data_type):
configuration = None
if processing_tool is not None:
config = builder_utils.get_config(config_name="proteomics.yml",
data_type='experiments')
if processing_tool in config:
tool_configuration = config[processing_tool]
if data_type in tool_configuration:
configuration = tool_configuration[data_type]
return configuration
def get_dataset_configuration(processing_format, data_type):
config = builder_utils.get_config(config_name="proteomics.yml",
data_type='experiments')
dataset_config = {}
if processing_format in config:
if data_type is not None:
if data_type in config[processing_format]:
dataset_config = config[processing_format][data_type]
else:
dataset_config = config[processing_format]
return dataset_config
def parser(databases_directory):
config = builder_utils.get_config(config_name="drugBankConfig.yml", data_type='databases')
directory = os.path.join(databases_directory, "DrugBank")
builder_utils.checkDirectory(directory)
drugs = extract_drugs(config, directory)
build_DrugBank_dictionary(config, directory, drugs)
relationships = build_relationships_from_DrugBank(config, drugs)
entities, attributes = build_drug_entity(config, drugs)
entities_header = ['ID'] + attributes
relationships_headers = config['relationships_headers']
return (entities, relationships, entities_header, relationships_headers)
def parser(projectId, type='proteomics', directory=None):
data = {}
experiments_directory = ckg_utils.read_ckg_config(
key='experiments_directory')
config = builder_utils.get_config(config_name="proteomics.yml",
data_type='experiments')
if directory is None:
directory = os.path.join(experiments_directory, 'PROJECTID/' + type)
directory = directory.replace('PROJECTID', projectId)
data = parse_from_directory(projectId, directory, config)
return data
def parser(databases_directory, download=True):
result = {}
config = builder_utils.get_config(config_name="jensenlabConfig.yml", data_type='databases')
string_mapping = mp.getSTRINGMapping(download=download)
for qtype in config['db_types']:
relationships = parsePairs(config, databases_directory, qtype, string_mapping)
entity1, entity2 = config['db_types'][qtype]
outputfileName = entity1+"_"+entity2+"_associated_with_integrated.tsv"
header = config['header']
result[qtype] = (relationships, header, outputfileName)
return result
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="gwasCatalogConfig.yml",
data_type='databases')
url = config['GWASCat_url']
entities_header = config['entities_header']
relationships_header = config['relationships_header']
entities = set()
relationships = defaultdict(set)
directory = os.path.join(databases_directory, "GWAScatalog")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
with open(fileName, 'r', encoding="utf-8") as catalog:
for line in catalog:
data = line.rstrip("\r\n").split("\t")
if len(data) > 36:
pubmedid = data[1]
date = data[3]
title = data[6]
sample_size = data[8]
replication_size = data[9]
#chromosome = data[11]
#position = data[12]
#genes_mapped = data[14].split(" - ")
snp_id = data[20].split('-')[0]
freq = data[26]
pval = data[27]
odds_ratio = data[30]
trait = data[34]
exp_factor = data[35]
study = data[36]
entities.add((study, "GWAS_study", title, date, sample_size,
replication_size, trait))
if pubmedid != "":
relationships["published_in_publication"].add(
(study, pubmedid, "PUBLISHED_IN", "GWAS Catalog"))
if snp_id != "":
relationships["variant_found_in_gwas"].add(
(re.sub(r"^\W+|\W+$", "",
snp_id), study, "VARIANT_FOUND_IN_GWAS", freq,
pval, odds_ratio, trait, "GWAS Catalog"))
if exp_factor != "":
exp_factor = exp_factor.split('/')[-1].replace('_', ':')
relationships["studies_trait"].add(
(study, exp_factor, "STUDIES_TRAIT", "GWAS Catalog"))
builder_utils.remove_directory(directory)
return (entities, relationships, entities_header, relationships_header)
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="hmdbConfig.yml", data_type='databases')
directory = os.path.join(databases_directory, "HMDB")
builder_utils.checkDirectory(directory)
metabolites = extract_metabolites(config, directory, download)
mapping = mp.getMappingFromOntology(ontology="Disease", source=config['HMDB_DO_source'])
mapping.update(mp.getMappingFromOntology(ontology="Tissue", source=None))
entities, attributes = build_metabolite_entity(config, directory, metabolites)
relationships = build_relationships_from_HMDB(config, metabolites, mapping)
entities_header = ['ID'] + attributes
relationships_header = config['relationships_header']
#builder_utils.remove_directory(directory)
return (entities, relationships, entities_header, relationships_header)
def get_STRING_mapping_url(db="STRING"):
"""
Get the url for downloading the mapping file from either STRING or STITCH
:param str db: Which database to get the url from: STRING or STITCH
:return: url where to download the mapping file
"""
url = None
config = builder_utils.get_config(config_name="stringConfig.yml",
data_type='databases')
if db.upper() == "STRING":
url = config['STRING_mapping_url']
elif db.upper() == "STITCH":
url = config['STITCH_mapping_url']
return url
def parser(databases_dir, download=True):
config = builder_utils.get_config(config_name="goaConfig.yml",
data_type='databases')
url = config['url']
rel_header = config['header']
protein_mapping = mp.getMappingForEntity(entity="Protein")
valid_proteins = list(set(protein_mapping.values))
directory = os.path.join(databases_dir, "GOA")
builder_utils.checkDirectory(directory)
file_name = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
annotations = parse_annotations_with_pandas(file_name, valid_proteins)
builder_utils.remove_directory(directory)
return annotations, rel_header
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="reactomeConfig.yml",
data_type='databases')
urls = config['reactome_urls']
entities = set()
relationships = defaultdict(set)
entities_header = config['pathway_header']
relationships_headers = config['relationships_header']
directory = os.path.join(databases_directory, "Reactome")
builder_utils.checkDirectory(directory)
metabolite_mapping = mp.getMappingForEntity("Metabolite")
#drug_mapping = mp.getMappingForEntity("Drug")
for dataset in urls:
url = urls[dataset]
file_name = url.split('/')[-1]
if download:
builder_utils.downloadDB(url, directory)
f = os.path.join(directory, file_name)
with open(f, 'r') as rf:
if dataset == "pathway":
entities = parsePathways(config, databases_directory, rf)
elif dataset == "hierarchy":
relationships[("pathway",
"has_parent")] = parsePathwayHierarchy(rf)
elif dataset == "protein":
relationships[(
dataset,
"annotated_to_pathway")] = parsePathwayRelationships(
config, rf)
elif dataset == "metabolite":
relationships[(
dataset,
"annotated_to_pathway")] = parsePathwayRelationships(
config, rf, metabolite_mapping)
#elif dataset == "drug":
#relationships[(dataset, "annotated_to_pathway")] = set()
builder_utils.remove_directory(directory)
return entities, relationships, entities_header, relationships_headers
def parser(databases_directory, download=True):
relationships = set()
config = builder_utils.get_config(config_name="mutationDsConfig.yml",
data_type='databases')
header = config['header']
output_file_name = "mutation_curated_affects_interaction_with.tsv"
regex = r":(\w+)\("
url = config['mutations_url']
directory = os.path.join(databases_directory, "MutationDs")
builder_utils.checkDirectory(directory)
file_name = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
with open(file_name, 'r') as mf:
first = True
for line in mf:
if first:
first = False
continue
data = line.rstrip("\r\n").split("\t")
if len(data) > 12:
internal_id = data[0]
pvariant = '_'.join(data[1].split(':'))
effect = data[5]
organism = data[10]
interaction = data[11]
evidence = data[12]
if organism.startswith("9606"):
matches = re.finditer(regex, interaction)
for matchNum, match in enumerate(matches, start=1):
interactor = match.group(1)
relationships.add((pvariant, interactor,
"CURATED_AFFECTS_INTERACTION_WITH",
effect, interaction, evidence,
internal_id, "Intact-MutationDs"))
builder_utils.remove_directory(directory)
return (relationships, header, output_file_name)
def parser(databases_directory, download=True):
config = builder_utils.get_config(
config_name="drugGeneInteractionDBConfig.yml", data_type='databases')
url = config['DGIdb_url']
header = config['header']
output_file = "dgidb_targets.tsv"
drugmapping = mp.getMappingForEntity("Drug")
relationships = set()
directory = os.path.join(databases_directory, "DGIdb")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
with open(fileName, 'r', encoding='utf-8') as associations:
first = True
for line in associations:
if first:
first = False
continue
data = line.rstrip("\r\n").split("\t")
gene = data[0]
source = data[3]
interactionType = data[4] if data[4] != '' else 'unknown'
drug = data[8].lower()
if drug == "":
drug = data[7]
if drug == "" and data[6] != "":
drug = data[6]
else:
continue
if gene != "":
if drug in drugmapping:
drug = drugmapping[drug]
relationships.add((drug, gene, "TARGETS", "NA", "NA", "NA",
interactionType, "DGIdb: " + source))
builder_utils.remove_directory(directory)
return (relationships, header, output_file)
def parser(databases_directory):
directory = os.path.join(databases_directory, "PhosphoSitePlus")
builder_utils.checkDirectory(directory)
config = builder_utils.get_config(config_name="pspConfig.yml",
data_type='databases')
modifications = config['modifications']
annotation_files = config['annotation_files']
entities_header = config['entities_header']
relationships_headers = config['rel_headers']
entities = set()
relationships = defaultdict(set)
for site_file in config['site_files']:
file_name = os.path.join(directory, site_file)
with gzip.open(file_name, 'r') as f:
sites, site_relationships = parseSites(f, modifications)
entities.update(sites)
for r in site_relationships:
relationships[r].update(site_relationships[r])
for er in annotation_files:
entity, relationship_type = er.split('-')
file_name = os.path.join(directory, annotation_files[er])
with gzip.open(file_name, 'r') as f:
if entity == "disease":
mapping = mp.getMappingFromOntology(ontology="Disease",
source=None)
relationships[(entity, relationship_type)].update(
parseDiseaseAnnotations(f, modifications, mapping))
elif entity == "biological_process":
mapping = mp.getMappingFromOntology(ontology="Gene_ontology",
source=None)
relationships[(entity, relationship_type)].update(
parseRegulationAnnotations(f, modifications, mapping))
elif entity == "substrate":
relationships[(entity,
relationship_type)] = parseKinaseSubstrates(
f, modifications)
return entities, relationships, entities_header, relationships_headers
def parser(databases_directory, download=True):
directory = os.path.join(databases_directory, "ExposomeExplorer")
builder_utils.checkDirectory(directory)
config = builder_utils.get_config(config_name="exposomeConfig.yml", data_type='databases')
database_urls = config['database_urls']
relationships_header = config['relationships_header']
mapping = mp.getMappingForEntity("Food")
correlations = {}
for url in database_urls:
zipped_fileName = os.path.join(directory, url.split('/')[-1])
file_name = '.'.join(url.split('/')[-1].split('.')[0:2])
if download:
builder_utils.downloadDB(url, directory)
with zipfile.ZipFile(zipped_fileName) as z:
if file_name == "biomarkers.csv":
biomarkers = parseBiomarkersFile(z, file_name)
elif file_name == "correlations.csv":
correlations = parseCorrelationsFile(z, file_name, biomarkers, mapping)
builder_utils.remove_directory(directory)
return correlations, relationships_header
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="hgncConfig.yml",
data_type='databases')
url = config['hgnc_url']
entities = set()
directory = os.path.join(databases_directory, "HGNC")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
taxid = 9606
entities_header = config['header']
if download:
builder_utils.downloadDB(url, directory)
with open(fileName, 'r', encoding="utf-8") as df:
first = True
for line in df:
if first:
first = False
continue
data = line.rstrip("\r\n").split("\t")
geneSymbol = data[1]
geneName = data[2]
status = data[5]
geneFamily = data[12]
synonyms = data[18:23]
transcript = data[23]
if status != "Approved":
continue
entities.add((geneSymbol, "Gene", geneName, geneFamily,
",".join(synonyms), taxid))
#relationships.add((geneSymbol, transcript, "TRANSCRIBED_INTO"))
builder_utils.remove_directory(directory)
return entities, entities_header
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="pathwayCommonsConfig.yml", data_type='databases')
url = config['pathwayCommons_pathways_url']
entities = set()
relationships = set()
directory = os.path.join(databases_directory, "PathwayCommons")
builder_utils.checkDirectory(directory)
fileName = url.split('/')[-1]
entities_header = config['pathways_header']
relationships_header = config['relationships_header']
if download:
builder_utils.downloadDB(url, directory)
f = os.path.join(directory, fileName)
associations = gzip.open(f, 'r')
for line in associations:
data = line.decode('utf-8').rstrip("\r\n").split("\t")
linkout = data[0]
code = data[0].split("/")[-1]
ptw_dict = dict([item.split(": ")[0], ":".join(item.split(": ")[1:])] for item in data[1].split("; "))
proteins = data[2:]
if "organism" in ptw_dict and ptw_dict["organism"] == "9606":
name = ptw_dict["name"]
source = ptw_dict["datasource"]
else:
continue
entities.add((code, "Pathway", name, name, ptw_dict["organism"], source, linkout))
for protein in proteins:
relationships.add((protein, code, "ANNOTATED_IN_PATHWAY", linkout, "PathwayCommons: "+source))
associations.close()
builder_utils.remove_directory(directory)
return (entities, relationships, entities_header, relationships_header)
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="hpaConfig.yml",
data_type='databases')
url = config['hpa_pathology_url']
disease_mapping = mp.getMappingFromOntology(ontology="Disease",
source=None)
protein_mapping = mp.getMultipleMappingForEntity("Protein")
directory = os.path.join(databases_directory, "HPA")
builder_utils.checkDirectory(directory)
compressed_fileName = os.path.join(directory, url.split('/')[-1])
file_name = '.'.join(url.split('/')[-1].split('.')[0:2])
relationships_headers = config['relationships_headers']
if download:
builder_utils.downloadDB(url, directory)
with zipfile.ZipFile(compressed_fileName) as z:
if file_name == "pathology.tsv":
pathology = parsePathologyFile(config, z, file_name,
protein_mapping, disease_mapping)
builder_utils.remove_directory(directory)
return (pathology, relationships_headers)
def parser(databases_directory, download=True):
variant_regex = r"(\D\d+\D)$"
regex = r"(chr\d+)\:g\.(\d+)(\w)>(\w)"
config = builder_utils.get_config(
config_name="cancerGenomeInterpreterConfig.yml", data_type='databases')
url = config['cancerBiomarkers_url']
entities_header = config['entities_header']
relationships_headers = config['relationships_headers']
amino_acids = config['amino_acids']
mapping = mp.getMappingFromOntology(ontology="Disease", source=None)
drugmapping = mp.getMappingForEntity("Drug")
protein_mapping = mp.getMultipleMappingForEntity("Protein")
fileName = config['cancerBiomarkers_variant_file']
relationships = defaultdict(set)
entities = set()
directory = os.path.join(databases_directory, "CancerGenomeInterpreter")
builder_utils.checkDirectory(directory)
zipFile = os.path.join(directory, url.split('/')[-1])
if download:
builder_utils.downloadDB(url, directory)
with zipfile.ZipFile(zipFile) as z:
if fileName in z.namelist():
with z.open(fileName, 'r') as responses:
first = True
for line in responses:
if first:
first = False
continue
data = line.decode('utf-8').rstrip("\r\n").split("\t")
gene_variant = data[0].split(':')
if len(gene_variant) < 2:
continue
gene = gene_variant[0]
variants = gene_variant[1].split(',')
#alterationType = data[1]
response = data[3]
drugs = data[10].split(';')
#status = data[11].split(';')
evidence = data[12]
tumors = data[16].split(';')
publications = data[17].split(';')
identifier = data[21]
prot_variant = data[22]
matches = re.match(regex, identifier)
alternative_names = [identifier]
if matches is not None:
cpra = matches.groups()
chromosome, position, reference, alternative = cpra
variant = chromosome + ":g." + position + reference + ">" + alternative
if prot_variant != "":
prot_variant = prot_variant.split(':')[1]
alternative_names.append(prot_variant)
valid_variants = []
if gene in protein_mapping:
for protein in protein_mapping[gene]:
for variant in variants:
match = re.search(variant_regex, variant)
if match:
if variant[0] in amino_acids and variant[
-1] in amino_acids:
valid_variant = protein + '_p.' + amino_acids[
variant[0]] + ''.join(
variant[1:-1]) + amino_acids[
variant[-1]]
valid_variants.append(valid_variant)
entities.add(
(valid_variant,
"Clinically_relevant_variant",
",".join(alternative_names),
chromosome, position, reference,
alternative, "", "", "CGI"))
relationships[
"known_variant_is_clinically_relevant"].add(
(valid_variant, valid_variant,
"KNOWN_VARIANT_IS_CLINICALLY_RELEVANT",
"CGI"))
for drug in drugs:
if drug.lower() in drugmapping:
drug = drugmapping[drug.lower()]
elif drug.split(" ")[0].lower() in drugmapping:
drug = drugmapping[drug.split(" ")[0].lower()]
elif " ".join(
drug.split(" ")[1:]).lower() in drugmapping:
drug = drugmapping[" ".join(
drug.split(" ")[1:]).lower()]
relationships["targets"].add(
(drug, gene, "CURATED_TARGETS", evidence, response,
",".join(tumors), "curated", "CGI"))
for valid_variant in valid_variants:
relationships[
"targets_clinically_relevant_variant"].add(
(drug, valid_variant,
"TARGETS_CLINICALLY_RELEVANT_VARIANT",
evidence, response, "".join(tumors),
"curated", "CGI"))
for tumor in tumors:
if tumor.lower() in mapping:
tumor = mapping[tumor.lower()]
for valid_variant in valid_variants:
relationships["associated_with"].add(
(valid_variant, tumor, "ASSOCIATED_WITH",
"curated", "curated", "CGI",
len(publications)))
builder_utils.remove_directory(directory)
return (entities, relationships, entities_header, relationships_headers)
def parser(databases_directory, download=True):
config = builder_utils.get_config(config_name="refseqConfig.yml",
data_type='databases')
url = config['refseq_url']
ftp_dir = config['refseq_ftp_dir']
entities = defaultdict(set)
relationships = defaultdict(set)
directory = os.path.join(databases_directory, "RefSeq")
builder_utils.checkDirectory(directory)
fileName = os.path.join(directory, url.split('/')[-1])
headers = config['headerEntities']
taxid = 9606
if download:
file_dir = builder_utils.list_ftp_directory(ftp_dir)[0]
new_file = file_dir.split('/')[-1] + "_feature_table.txt.gz"
url = ftp_dir + file_dir.split('/')[-1] + "/" + new_file
builder_utils.downloadDB(url, directory)
fileName = os.path.join(directory, new_file)
if os.path.isfile(fileName):
df = builder_utils.read_gzipped_file(fileName)
first = True
for line in df:
if first:
first = False
continue
data = line.rstrip("\r\n").split("\t")
tclass = data[1]
assembly = data[2]
chrom = data[5]
geneAcc = data[6]
start = data[7]
end = data[8]
strand = data[9]
protAcc = data[10]
name = data[13]
symbol = data[14]
if protAcc != "":
entities["Transcript"].add(
(protAcc, "Transcript", name, tclass, assembly, taxid))
if chrom != "":
entities["Chromosome"].add(
(chrom, "Chromosome", chrom, taxid))
relationships["LOCATED_IN"].add(
(protAcc, chrom, "LOCATED_IN", start, end, strand,
"RefSeq"))
if symbol != "":
relationships["TRANSCRIBED_INTO"].add(
(symbol, protAcc, "TRANSCRIBED_INTO", "RefSeq"))
elif geneAcc != "":
entities["Transcript"].add(
(geneAcc, "Transcript", name, tclass, assembly, taxid))
if chrom != "":
entities["Chromosome"].add(
(chrom, "Chromosome", chrom, taxid))
relationships["LOCATED_IN"].add(
(protAcc, chrom, "LOCATED_IN", start, end, strand,
"RefSeq"))
df.close()
builder_utils.remove_directory(directory)
return (entities, relationships, headers)
def parser(databases_directory,
import_directory,
download=True,
updated_on=None):
config = builder_utils.get_config(config_name="pfamConfig.yml",
data_type='databases')
entity_header = config['entity_header']
relationship_headers = config['relationship_headers']
directory = os.path.join(databases_directory, 'Pfam')
builder_utils.checkDirectory(directory)
protein_mapping = mp.getMappingForEntity(entity="Protein")
valid_proteins = list(set(protein_mapping.values()))
ftp_url = config['ftp_url']
filename = config['full_uniprot_file']
# url = config['test']
if not os.path.exists(os.path.join(directory, filename)):
if download:
builder_utils.downloadDB(ftp_url + filename, directory)
stats = set()
if os.path.exists(os.path.join(directory, filename)):
fhandler = builder_utils.read_gzipped_file(
os.path.join(directory, filename))
identifier = None
description = []
lines = []
missed = 0
entities = set()
relationships = defaultdict(set)
is_first = True
i = 0
read_lines = 0
num_entities = 0
num_relationships = {}
try:
for line in fhandler:
i += 1
read_lines += 1
if line.startswith("# STOCKHOLM"):
if identifier is not None:
entities.add((identifier, 'Functional_region', name,
" ".join(description), "PFam"))
if len(entities) == 100:
print_files(entities,
entity_header,
outputfile=os.path.join(
import_directory,
'Functional_region.tsv'),
is_first=is_first)
num_entities += len(entities)
if 'mentioned_in_publication' in relationships:
print_files(
relationships['mentioned_in_publication'],
relationship_headers[
'mentioned_in_publication'],
outputfile=os.path.join(
import_directory,
'Functional_region_mentioned_in_publication.tsv'
),
is_first=is_first)
if 'mentioned_in_publication' not in num_relationships:
num_relationships[
'mentioned_in_publication'] = 0
num_relationships[
'mentioned_in_publication'] += len(
relationships[
'mentioned_in_publication'])
if 'found_in_protein' in relationships:
print_files(
relationships['found_in_protein'],
relationship_headers['found_in_protein'],
outputfile=os.path.join(
import_directory,
'Functional_region_found_in_protein.tsv'
),
is_first=is_first,
filter_for=('END_ID', valid_proteins))
if 'found_in_protein' not in num_relationships:
num_relationships['found_in_protein'] = 0
num_relationships['found_in_protein'] += len(
relationships['found_in_protein'])
entities = set()
relationships = defaultdict(set)
is_first = False
identifier = None
description = []
elif line.startswith("#=GF"):
data = line.rstrip('\r\n').split()
if 'AC' in data:
identifier = data[2].split('.')[0]
elif 'DE' in data:
name = " ".join(data[2:])
elif 'RM' in data:
relationships['mentioned_in_publication'].add(
(identifier, data[2], "MENTIONED_IN_PUBLICATION",
"PFam"))
elif 'CC' in data:
description.append(" ".join(data[2:]))
elif not line.startswith('//'):
data = line.rstrip('\r\n').split()
protein, positions = data[0].split('/')
protein = protein.replace('.', '-')
start, end = positions.split('-')
sequence = data[1]
relationships['found_in_protein'].add(
(identifier, protein, "FOUND_IN_PROTEIN", start, end,
sequence, "PFam"))
if protein.split('-')[0] != protein:
relationships['found_in_protein'].add(
(identifier, protein.split('-')[0],
"FOUND_IN_PROTEIN", start, end, sequence, "PFam"))
except UnicodeDecodeError:
lines.append(i)
missed += 1
fhandler.close()
if len(entities) > 0:
print_files(entities,
entity_header,
outputfile=os.path.join(import_directory,
'Functional_region.tsv'),
is_first=is_first)
num_entities += len(entities)
print_files(relationships['mentioned_in_publication'],
relationship_headers['mentioned_in_publication'],
outputfile=os.path.join(
import_directory,
'Functional_region_mentioned_in_publication.tsv'),
is_first=is_first)
num_relationships['mentioned_in_publication'] += len(
relationships['mentioned_in_publication'])
print_files(relationships['found_in_protein'],
relationship_headers['found_in_protein'],
outputfile=os.path.join(
import_directory,
'Functional_region_found_in_protein.tsv'),
is_first=is_first)
num_relationships['found_in_protein'] += len(
relationships['found_in_protein'])
stats.add(
builder_utils.buildStats(num_entities, "entity",
"Functional_region", "Pfam",
'Functional_region.tsv', updated_on))
for rel in num_relationships:
stats.add(
builder_utils.buildStats(num_relationships[rel],
"relationship", rel.upper(), "Pfam",
'Functional_region_' + rel + '.tsv',
updated_on))
builder_utils.remove_directory(directory)
return stats
