class ZFINTestCase(SourceTestCase):
def setUp(self):
self.source = ZFIN('rdf_graph', True)
self.source.settestonly(True)
self._setDirToSource()
return
def tearDown(self):
self.source = None
return
@unittest.skip(
'Will eventually write test to check if phenotype sextuples' +
'are mapped to ZP ids')
def test_allZPAvailable(self):
"""
This test will identify if there are
any missing ZP terms in the mapping file
:return:
"""
# TODO add this test to check if all phenotype sextuples
# are mapped to ZP ids
return
class ZFINTestCase(SourceTestCase):
def setUp(self):
self.source = ZFIN('rdf_graph', True)
self.source.settestonly(True)
self._setDirToSource()
return
def tearDown(self):
self.source = None
return
@unittest.skip(
'Will eventually write test to check if phenotype sextuples' +
'are mapped to ZP ids')
def test_allZPAvailable(self):
"""
This test will identify if there are
any missing ZP terms in the mapping file
:return:
"""
# TODO add this test to check if all phenotype sextuples
# are mapped to ZP ids
return
def parse(self, limit=None):
zfin_parser = ZFIN(self.graph_type, self.are_bnodes_skized)
model = Model(self.graph)
zp_file = '/'.join((self.rawdir, self.files['zpmap']['file']))
g2p_file = '/'.join((self.rawdir, self.files['g2p_clean']['file']))
zfin_parser.zp_map = zfin_parser._load_zp_mappings(zp_file)
with open(g2p_file, 'r', encoding="utf8") as csvfile:
filereader = csv.reader(csvfile, delimiter='\t', quotechar='\"')
for row in filereader:
(internal_id, symbol, gene_id, subterm1_id, subterm1_label,
pc_rel_id, pc_rel_label, superterm1_id, superterm1_label,
quality_id, quality_name, modifier, subterm2_id,
subterm2_label, pc_rel2_id, pc_rel2_id, superterm2_id,
superterm2_label, fish_id, fish_label, start_stage, end_stage,
environment, pub_id, figure_id, unknown_field) = row
zp_id = zfin_parser._map_sextuple_to_phenotype(
superterm1_id, subterm1_id, quality_id, superterm2_id,
subterm2_id, modifier)
gene_curie = "ZFIN:{0}".format(gene_id)
model.makeLeader(gene_curie)
pub_curie = "ZFIN:{0}".format(pub_id)
if zp_id:
assoc = G2PAssoc(self.graph, self.name, gene_curie, zp_id)
if pub_id:
reference = Reference(self.graph, pub_curie,
Reference.ref_types['document'])
reference.addRefToGraph()
assoc.add_source(pub_curie)
assoc.add_evidence('ECO:0000059')
assoc.add_association_to_graph()
def parse(self, limit=None):
zfin_parser = ZFIN(self.graph_type, self.are_bnodes_skized)
model = Model(self.graph)
zp_file = '/'.join((self.rawdir, self.files['zpmap']['file']))
g2p_file = '/'.join((self.rawdir, self.files['g2p_clean']['file']))
zfin_parser.zp_map = zfin_parser._load_zp_mappings(zp_file)
with open(g2p_file, 'r', encoding="utf8") as csvfile:
filereader = csv.reader(csvfile, delimiter='\t', quotechar='\"')
for row in filereader:
(internal_id, symbol, gene_id, subterm1_id, subterm1_label,
pc_rel_id, pc_rel_label, superterm1_id, superterm1_label,
quality_id, quality_name, modifier, subterm2_id,
subterm2_label, pc_rel2_id, pc_rel2_id, superterm2_id,
superterm2_label, fish_id, fish_label, start_stage, end_stage,
environment, pub_id, figure_id, unknown_field) = row
zp_id = zfin_parser._map_sextuple_to_phenotype(
superterm1_id, subterm1_id, quality_id, superterm2_id,
subterm2_id, modifier)
gene_curie = "ZFIN:{0}".format(gene_id)
model.makeLeader(gene_curie)
pub_curie = "ZFIN:{0}".format(pub_id)
if zp_id:
assoc = G2PAssoc(self.graph, self.name, gene_curie, zp_id)
if pub_id:
reference = Reference(self.graph, pub_curie,
Reference.ref_types['document'])
reference.addRefToGraph()
assoc.add_source(pub_curie)
assoc.add_evidence('ECO:0000059')
assoc.add_association_to_graph()
def __init__(self,
graph_type,
are_bnodes_skolemized,
data_release_version=None,
tax_ids=None):
super().__init__(
graph_type=graph_type,
are_bnodes_skized=are_bnodes_skolemized,
data_release_version=data_release_version,
name='go',
ingest_title='Gene Ontology',
ingest_url='http://www.geneontology.org',
ingest_logo='source-geneontology.png',
license_url=None,
data_rights='http://geneontology.org/page/use-and-license'
# file_handle=None
)
self.test_ids = []
# note: dipper-etl defaults tax_ids to '9606'
# note: sorting tax_ids for stable digest
if tax_ids is not None and [] != set(tax_ids).difference(['9606']):
LOG.info('Have %s given as taxon to ingest', str(tax_ids))
self.tax_ids = sorted([str(x) for x in tax_ids])
nottax = set(tax_ids) - set(self.files.keys())
if nottax:
LOG.error('Cant process taxon number(s):\t%s', str(nottax))
self.tax_ids = list(set(self.tax_ids) - nottax)
else:
self.tax_ids = sorted(['9606', '10090', '7955'])
LOG.info("Filtering to the following taxa: %s", self.tax_ids)
# moving this from process_gaf() to avoid repeating this for each
# file to be processed.
if '7955' in self.tax_ids:
self.zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
if '6239' in self.tax_ids:
self.wbase = WormBase(self.graph_type, self.are_bnodes_skized)
if 'gene' not in self.all_test_ids:
LOG.warning("not configured with gene test ids.")
else:
self.test_ids = self.all_test_ids['gene']
# build the id map for mapping uniprot ids to genes ... ONCE
self.uniprot_entrez_id_map = self.get_uniprot_entrez_id_map()
# gaf evidence code mapping is built in parse(), after the file is fetched.
self.gaf_eco = {}
def process_gaf(self, file, limit, id_map=None, eco_map=None):
if self.test_mode:
graph = self.testgraph
else:
graph = self.graph
model = Model(graph)
geno = Genotype(graph)
LOG.info("Processing Gene Associations from %s", file)
line_counter = 0
uniprot_hit = 0
uniprot_miss = 0
if '7955' in self.tax_ids:
zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
if '6239' in self.tax_ids:
wbase = WormBase(self.graph_type, self.are_bnodes_skized)
with gzip.open(file, 'rb') as csvfile:
filereader = csv.reader(io.TextIOWrapper(csvfile, newline=""),
delimiter='\t',
quotechar='\"')
for row in filereader:
line_counter += 1
# comments start with exclamation
if re.match(r'!', ''.join(row)):
continue
if len(row) > 17 or len(row) < 15:
LOG.warning(
"Wrong number of columns %i, expected 15 or 17\n%s",
len(row), row)
continue
if 17 > len(row) >= 15:
row += [""] * (17 - len(row))
(dbase, gene_num, gene_symbol, qualifier, go_id, ref,
eco_symbol, with_or_from, aspect, gene_name, gene_synonym,
object_type, taxon, date, assigned_by, annotation_extension,
gene_product_form_id) = row
# test for required fields
if (dbase == '' or gene_num == '' or gene_symbol == ''
or go_id == '' or ref == '' or eco_symbol == ''
or aspect == '' or object_type == '' or taxon == ''
or date == '' or assigned_by == ''):
LOG.error(
"Missing required part of annotation on row %d:\n" +
'\t'.join(row), line_counter)
continue
# deal with qualifier NOT, contributes_to, colocalizes_with
if re.search(r'NOT', qualifier):
continue
if dbase in self.localtt:
dbase = self.localtt[dbase]
uniprotid = None
gene_id = None
if dbase == 'UniProtKB':
if id_map is not None and gene_num in id_map:
gene_id = id_map[gene_num]
uniprotid = ':'.join((dbase, gene_num))
(dbase, gene_num) = gene_id.split(':')
uniprot_hit += 1
else:
# LOG.warning(
# "UniProt id %s is without a 1:1 mapping to entrez/ensembl",
# gene_num)
uniprot_miss += 1
continue
else:
gene_num = gene_num.split(':')[-1] # last
gene_id = ':'.join((dbase, gene_num))
if self.test_mode and not (re.match(r'NCBIGene', gene_id)
and int(gene_num) in self.test_ids):
continue
model.addClassToGraph(gene_id, gene_symbol)
if gene_name != '':
model.addDescription(gene_id, gene_name)
if gene_synonym != '':
for syn in re.split(r'\|', gene_synonym):
model.addSynonym(gene_id, syn.strip())
if re.search(r'\|', taxon):
# TODO add annotations with >1 taxon
LOG.info(">1 taxon (%s) on line %d. skipping", taxon,
line_counter)
else:
tax_id = re.sub(r'taxon:', 'NCBITaxon:', taxon)
geno.addTaxon(tax_id, gene_id)
assoc = Assoc(graph, self.name)
assoc.set_subject(gene_id)
assoc.set_object(go_id)
try:
eco_id = eco_map[eco_symbol]
assoc.add_evidence(eco_id)
except KeyError:
LOG.error("Evidence code (%s) not mapped", eco_symbol)
refs = re.split(r'\|', ref)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0] # sidestep 'MGI:MGI:'
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
refg = Reference(graph, ref)
if prefix == 'PMID':
ref_type = self.globaltt['journal article']
refg.setType(ref_type)
refg.addRefToGraph()
assoc.add_source(ref)
# TODO add the source of the annotations from assigned by?
rel = self.resolve(aspect, mandatory=False)
if rel is not None and aspect == rel:
if aspect == 'F' and re.search(r'contributes_to',
qualifier):
assoc.set_relationship(self.globaltt['contributes to'])
else:
LOG.error(
"Aspect: %s with qualifier: %s is not recognized",
aspect, qualifier)
elif rel is not None:
assoc.set_relationship(rel)
assoc.add_association_to_graph()
else:
LOG.warning("No predicate for association \n%s\n",
str(assoc))
if uniprotid is not None:
assoc.set_description('Mapped from ' + uniprotid)
# object_type should be one of:
# protein_complex; protein; transcript; ncRNA; rRNA; tRNA;
# snRNA; snoRNA; any subtype of ncRNA in the Sequence Ontology.
# If the precise product type is unknown,
# gene_product should be used
#######################################################################
# Derive G2P Associations from IMP annotations
# in version 2.1 Pipe will indicate 'OR'
# and Comma will indicate 'AND'.
# in version 2.0, multiple values are separated by pipes
# where the pipe has been used to mean 'AND'
if eco_symbol == 'IMP' and with_or_from != '':
withitems = re.split(r'\|', with_or_from)
phenotypeid = go_id + 'PHENOTYPE'
# create phenotype associations
for i in withitems:
if i == '' or re.match(
r'(UniProtKB|WBPhenotype|InterPro|HGNC)', i):
LOG.warning(
"Don't know what having a uniprot id " +
"in the 'with' column means of %s", uniprotid)
continue
i = re.sub(r'MGI\:MGI\:', 'MGI:', i)
i = re.sub(r'WB:', 'WormBase:', i)
# for worms and fish, they might give a RNAi or MORPH
# in these cases make a reagent-targeted gene
if re.search('MRPHLNO|CRISPR|TALEN', i):
targeted_gene_id = zfin.make_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(
i, gene_id, targeted_gene_id)
# TODO PYLINT why is this needed?
# Redefinition of assoc type from
# dipper.models.assoc.Association.Assoc to
# dipper.models.assoc.G2PAssoc.G2PAssoc
assoc = G2PAssoc(graph, self.name,
targeted_gene_id, phenotypeid)
elif re.search(r'WBRNAi', i):
targeted_gene_id = wbase.make_reagent_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(
i, gene_id, targeted_gene_id)
assoc = G2PAssoc(graph, self.name,
targeted_gene_id, phenotypeid)
else:
assoc = G2PAssoc(graph, self.name, i, phenotypeid)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0]
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
assoc.add_source(ref)
# experimental phenotypic evidence
assoc.add_evidence(self.globaltt[
'experimental phenotypic evidence'])
assoc.add_association_to_graph()
# TODO should the G2PAssoc be
# the evidence for the GO assoc?
if not self.test_mode and limit is not None and line_counter > limit:
break
uniprot_tot = (uniprot_hit + uniprot_miss)
uniprot_per = 0.0
if uniprot_tot != 0:
uniprot_per = 100.0 * uniprot_hit / uniprot_tot
LOG.info(
"Uniprot: %.2f%% of %i benefited from the 1/4 day id mapping download",
uniprot_per, uniprot_tot)
return
def setUp(self):
self.source = ZFIN('rdf_graph', True)
self.source.settestonly(True)
self._setDirToSource()
return
class GeneOntology(Source):
"""
This is the parser for the
[Gene Ontology Annotations](http://www.geneontology.org),
from which we process gene-process/function/subcellular
location associations.
We generate the GO graph to include the following information:
* genes
* gene-process
* gene-function
* gene-location
We process only a subset of the organisms:
Status: IN PROGRESS / INCOMPLETE
"""
gaf_columns = [ # GAF2.1 files contain the following columns:
'DB',
'DB_Object_ID',
'DB_Object_Symbol',
'Qualifier',
'GO_ID',
'DB:Reference',
'Evidence Code',
'With (or) From',
'Aspect',
'DB_Object_Name',
'DB_Object_Synonym',
'DB_Object_Type',
'Taxon and Interacting taxon',
'Date',
'Assigned_By',
'Annotation_Extension',
'Gene_Product_Form_ID'
]
files = {
'9615': { # Canis lupus familiaris
'file': 'goa_dog.gaf.gz',
'url': GOGA + '/goa_dog.gaf.gz',
'columnns': gaf_columns
},
'7227': { # Drosophila melanogaster
'file': 'fb.gaf.gz',
'url': GOGA + '/fb.gaf.gz',
'columnns': gaf_columns
},
'7955': { # Danio rerio
'file': 'zfin.gaf.gz',
'url': GOGA + '/zfin.gaf.gz',
'columnns': gaf_columns
},
'10090': { # Mus musculus
'file': 'mgi.gaf.gz',
'url': GOGA + '/mgi.gaf.gz',
'columnns': gaf_columns
},
'10116': { # Rattus norvegicus
'file': 'rgd.gaf.gz',
'url': GOGA + '/rgd.gaf.gz',
'columnns': gaf_columns
},
'6239': { # Caenorhabditis elegans
'file': 'wb.gaf.gz',
'url': GOGA + '/wb.gaf.gz',
'columnns': gaf_columns
},
'9823': { # Sus scrofa
'file': 'goa_pig.gaf.gz',
'url': GOGA + '/goa_pig.gaf.gz',
'columnns': gaf_columns
},
'9031': { # Gallus gallus
'file': 'goa_chicken.gaf.gz',
'url': GOGA + '/goa_chicken.gaf.gz',
'columnns': gaf_columns
},
'9606': { # H**o sapiens
'file': 'goa_human.gaf.gz',
'url': GOGA + '/goa_human.gaf.gz',
'columnns': gaf_columns
},
'9913': { # Bos taurus
'file': 'goa_cow.gaf.gz',
'url': GOGA + '/goa_cow.gaf.gz',
'columnns': gaf_columns
},
'559292': { # Saccharomyces cerevisiae 4932
'file': 'sgd.gaf.gz',
'url': GOGA + '/sgd.gaf.gz',
'columnns': gaf_columns
},
'4896': { # Schizosaccharomyces pombe (yeast)
'file': 'pombase.gaf.gz',
'url': GOGA + '/pombase.gaf.gz',
'columnns': gaf_columns
},
'5782': { # Dictyostelium (slime mold genus)
'file': 'dictibase.gaf.gz',
'url': GOGA + '/dictybase.gaf.gz',
'columnns': gaf_columns
},
'5052': { # Aspergillus (fungi) http://www.aspergillusgenome.org/
'file': 'aspgd.gaf.gz',
'url': GOGA + '/aspgd.gaf.gz',
'columnns': gaf_columns
},
# consider this after most others - should this be part of GO?
# 'multispecies': {
# 'file': 'gene_association.goa_uniprot.gz',
# 'url': FTPEBI + 'GO/goa/UNIPROT/gene_association.goa_uniprot.gz'},
'go-references': {
'file': 'GO.references',
# Quoth the header of this file: "This file is DEPRECATED.
# Please see go-refs.json relative to this location"
# (http://current.geneontology.org/metadata/go-refs.json)
'url': 'http://www.geneontology.org/doc/GO.references'
},
'id-map': { # 5GB mapping file takes 6 hours to DL ... maps UniProt to Ensembl
'file': 'idmapping_selected.tab.gz',
'url': FTPEBI + UPCRKB + 'idmapping/idmapping_selected.tab.gz',
# ftp://ftp.uniprot.org
# /pub/databases/uniprot/current_release/knowledgebase/idmapping/README
'columns': [
'UniProtKB-AC',
'UniProtKB-ID',
'GeneID (EntrezGene)',
'RefSeq',
'GI',
'PDB',
'GO',
'UniRef100',
'UniRef90',
'UniRef50',
'UniParc',
'PIR',
'NCBI-taxon',
'MIM',
'UniGene',
'PubMed',
'EMBL',
'EMBL-CDS',
'Ensembl',
'Ensembl_TRS',
'Ensembl_PRO',
'Additional PubMed'
]
}
}
# consider moving the go-ref and id-map above to here in map_files
map_files = {
'eco_map': 'http://purl.obolibrary.org/obo/eco/gaf-eco-mapping.txt',
}
def __init__(self,
graph_type,
are_bnodes_skolemized,
data_release_version=None,
tax_ids=None):
super().__init__(
graph_type=graph_type,
are_bnodes_skized=are_bnodes_skolemized,
data_release_version=data_release_version,
name='go',
ingest_title='Gene Ontology',
ingest_url='http://www.geneontology.org',
ingest_logo='source-geneontology.png',
license_url=None,
data_rights='http://geneontology.org/page/use-and-license'
# file_handle=None
)
self.test_ids = []
# note: dipper-etl defaults tax_ids to '9606'
# note: sorting tax_ids for stable digest
if tax_ids is not None and [] != set(tax_ids).difference(['9606']):
LOG.info('Have %s given as taxon to ingest', str(tax_ids))
self.tax_ids = sorted([str(x) for x in tax_ids])
nottax = set(tax_ids) - set(self.files.keys())
if nottax:
LOG.error('Cant process taxon number(s):\t%s', str(nottax))
self.tax_ids = list(set(self.tax_ids) - nottax)
else:
self.tax_ids = sorted(['9606', '10090', '7955'])
LOG.info("Filtering to the following taxa: %s", self.tax_ids)
# moving this from process_gaf() to avoid repeating this for each
# file to be processed.
if '7955' in self.tax_ids:
self.zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
if '6239' in self.tax_ids:
self.wbase = WormBase(self.graph_type, self.are_bnodes_skized)
if 'gene' not in self.all_test_ids:
LOG.warning("not configured with gene test ids.")
else:
self.test_ids = self.all_test_ids['gene']
# build the id map for mapping uniprot ids to genes ... ONCE
self.uniprot_entrez_id_map = self.get_uniprot_entrez_id_map()
self.eco_map = self.get_eco_map(self.map_files['eco_map'])
def fetch(self, is_dl_forced=False):
self.get_files(is_dl_forced)
def parse(self, limit=None):
if limit is not None:
LOG.info("Only parsing first %s rows of each file", limit)
LOG.info("Parsing files...")
if self.test_only:
self.test_mode = True
for txid_num in list(set(self.files).intersection(self.tax_ids)):
gaffile = '/'.join((self.rawdir, self.files[txid_num]['file']))
self.process_gaf(gaffile, limit, self.uniprot_entrez_id_map, self.eco_map)
LOG.info("Finished parsing.")
def process_gaf(self, gaffile, limit, id_map=None, eco_map=None):
if self.test_mode:
graph = self.testgraph
else:
graph = self.graph
model = Model(graph)
geno = Genotype(graph)
LOG.info("Processing Gene Associations from %s", gaffile)
uniprot_hit = 0
uniprot_miss = 0
col = self.gaf_columns
with gzip.open(gaffile, 'rb') as csvfile:
reader = csv.reader(
io.TextIOWrapper(csvfile, newline=""), delimiter='\t', quotechar='\"')
for row in reader:
# comments start with exclamation
if row[0][0] == '!':
continue
if len(row) != len(col):
LOG.error(
"Wrong number of columns %i, expected ... got:\n\t%s",
len(col), row)
exit(1)
dbase = row[col.index('DB')].strip()
gene_num = row[col.index('DB_Object_ID')].strip()
gene_symbol = row[col.index('DB_Object_Symbol')].strip()
qualifier = row[col.index('Qualifier')]
go_id = row[col.index('GO_ID')].strip()
ref = row[col.index('DB:Reference')].strip()
eco_symbol = row[col.index('Evidence Code')].strip()
with_or_from = row[col.index('With (or) From')]
aspect = row[col.index('Aspect')].strip()
gene_name = row[col.index('DB_Object_Name')]
gene_synonym = row[col.index('DB_Object_Synonym')]
# object_type = row[col.index('DB_Object_Type')].strip()
taxon = row[col.index('Taxon and Interacting taxon')].strip()
# date = row[col.index('Date')].strip()
# assigned_by = row[col.index('Assigned_By')].strip()
# annotation_extension = row[col.index('Annotation_Extension')]
# gene_product_form_id = row[col.index('Gene_Product_Form_ID')]
# test for required fields
if '' in [row[:10], row[12]]:
LOG.error(
"Missing required part of annotation on row %i:\n%s",
reader.line_num, str(row[:-4]))
continue
# (Don't) deal with qualifier NOT, contributes_to, colocalizes_with
if re.search(r'NOT', qualifier):
continue
if dbase in self.localtt:
dbase = self.localtt[dbase]
uniprotid = None
gene_id = None
if dbase == 'UniProtKB':
if id_map is not None and gene_num in id_map:
gene_id = id_map[gene_num]
uniprotid = ':'.join((dbase, gene_num))
(dbase, gene_num) = gene_id.split(':')
uniprot_hit += 1
else:
# LOG.warning(
# "UniProt id %s is without a 1:1 mapping to entrez/ensembl",
# gene_num)
uniprot_miss += 1
continue
else:
gene_num = gene_num.split(':')[-1] # last
gene_id = ':'.join((dbase, gene_num))
if self.test_mode and gene_id[:9] != 'NCBIGene:' and\
gene_num not in self.test_ids:
continue
model.addClassToGraph(gene_id, gene_symbol)
if gene_name != '':
model.addDescription(gene_id, gene_name)
if gene_synonym != '':
for syn in re.split(r'\|', gene_synonym):
syn = syn.strip()
if syn[:10] == 'UniProtKB:':
model.addTriple(
gene_id, self.globaltt['has gene product'], syn)
elif re.fullmatch(graph.curie_regexp, syn) is not None:
LOG.warning(
'possible curie "%s" as a literal synomym for %s',
syn, gene_id)
model.addSynonym(gene_id, syn)
else:
model.addSynonym(gene_id, syn)
for txid in taxon.split('|'):
tax_curie = re.sub(r'taxon:', 'NCBITaxon:', txid)
geno.addTaxon(tax_curie, gene_id)
assoc = Assoc(graph, self.name)
assoc.set_subject(gene_id)
assoc.set_object(go_id)
try:
eco_id = eco_map[eco_symbol]
assoc.add_evidence(eco_id)
except KeyError:
LOG.error("Evidence code (%s) not mapped", eco_symbol)
refs = re.split(r'\|', ref)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0] # sidestep 'MGI:MGI:'
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
refg = Reference(graph, ref)
if prefix == 'PMID':
ref_type = self.globaltt['journal article']
refg.setType(ref_type)
refg.addRefToGraph()
assoc.add_source(ref)
# TODO add the source of the annotations from assigned by?
rel = self.resolve(aspect, mandatory=False)
if rel is not None and aspect == rel:
if aspect == 'F' and re.search(r'contributes_to', qualifier):
assoc.set_relationship(self.globaltt['contributes to'])
else:
LOG.error(
"Aspect: %s with qualifier: %s is not recognized",
aspect, qualifier)
elif rel is not None:
assoc.set_relationship(rel)
assoc.add_association_to_graph()
else:
LOG.warning("No predicate for association \n%s\n", str(assoc))
if uniprotid is not None:
assoc.set_description('Mapped from ' + uniprotid)
# object_type should be one of:
# protein_complex; protein; transcript; ncRNA; rRNA; tRNA;
# snRNA; snoRNA; any subtype of ncRNA in the Sequence Ontology.
# If the precise product type is unknown,
# gene_product should be used
########################################################################
# Derive G2P Associations from IMP annotations
# in version 2.1 Pipe will indicate 'OR'
# and Comma will indicate 'AND'.
# in version 2.0, multiple values are separated by pipes
# where the pipe has been used to mean 'AND'
if eco_symbol == 'IMP' and with_or_from != '':
withitems = with_or_from.split('|')
phenotypeid = go_id + 'PHENOTYPE'
# create phenotype associations
for itm in withitems:
if itm == '' or re.match(
r'(UniProtKB|WBPhenotype|InterPro|HGNC)', itm):
LOG.warning(
"Skipping %s from or with %s", uniprotid, itm)
continue
itm = re.sub(r'MGI\:MGI\:', 'MGI:', itm)
itm = re.sub(r'WB:', 'WormBase:', itm)
# for worms and fish, they might give a RNAi or MORPH
# in these cases make a reagent-targeted gene
if re.search('MRPHLNO|CRISPR|TALEN', itm):
targeted_gene_id = self.zfin.make_targeted_gene_id(
gene_id, itm)
geno.addReagentTargetedGene(itm, gene_id, targeted_gene_id)
# TODO PYLINT why is this needed?
# Redefinition of assoc type from
# dipper.models.assoc.Association.Assoc to
# dipper.models.assoc.G2PAssoc.G2PAssoc
assoc = G2PAssoc(
graph, self.name, targeted_gene_id, phenotypeid)
elif re.search(r'WBRNAi', itm):
targeted_gene_id = self.wbase.make_reagent_targeted_gene_id(
gene_id, itm)
geno.addReagentTargetedGene(itm, gene_id, targeted_gene_id)
assoc = G2PAssoc(
graph, self.name, targeted_gene_id, phenotypeid)
else:
assoc = G2PAssoc(graph, self.name, itm, phenotypeid)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0]
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
assoc.add_source(ref)
# experimental phenotypic evidence
assoc.add_evidence(
self.globaltt['experimental phenotypic evidence'])
assoc.add_association_to_graph()
# TODO should the G2PAssoc be the evidence for the GO assoc?
if not self.test_mode and limit is not None and \
reader.line_num > limit:
break
uniprot_tot = (uniprot_hit + uniprot_miss)
uniprot_per = 0.0
if uniprot_tot != 0:
uniprot_per = 100.0 * uniprot_hit / uniprot_tot
LOG.info(
"Uniprot: %.2f%% of %i benefited from the 1/4 day id mapping download",
uniprot_per, uniprot_tot)
def get_uniprot_entrez_id_map(self):
src_key = 'id-map'
taxon_digest = GraphUtils.digest_id(str(self.tax_ids))
id_map = {}
smallfile = '/'.join((self.rawdir, 'id_map_' + taxon_digest + '.yaml'))
bigfile = '/'.join((self.rawdir, self.files[src_key]['file']))
# if processed smallfile exists and is newer than bigfile then use it instesd
if os.path.isfile(smallfile) and \
os.path.getctime(smallfile) > os.path.getctime(bigfile):
LOG.info("Using the cheap mapping file %s", smallfile)
with open(smallfile, 'r') as yamlreader:
id_map = yaml.safe_load(yamlreader)
else:
LOG.info(
"Expensive Mapping from Uniprot IDs to Entrez/ENSEMBL gene ids for %s",
self.tax_ids)
self.fetch_from_url(self.files[src_key]['url'], bigfile)
col = self.files[src_key]['columns']
ummapped_uniprot = 0
with gzip.open(bigfile, 'rb') as csvfile:
csv.field_size_limit(sys.maxsize)
reader = csv.reader( # warning this file is over 10GB unzipped
io.TextIOWrapper(csvfile, newline=""),
delimiter='\t', quotechar='\"')
for row in reader:
uniprotkb_ac = row[col.index('UniProtKB-AC')].strip()
# uniprotkb_id = row[col.index('UniProtKB-ID')]
geneid = row[col.index('GeneID (EntrezGene)')].strip()
# refseq = row[col.index('RefSeq')]
# gi = row[col.index('GI')]
# pdb = row[col.index('PDB')]
# go = row[col.index('GO')]
# uniref100 = row[col.index('UniRef100')]
# unifref90 = row[col.index('UniRef90')]
# uniref50 = row[col.index('UniRef50')]
# uniparc = row[col.index('UniParc')]
# pir = row[col.index('PIR')]
ncbitaxon = row[col.index('NCBI-taxon')].strip()
# mim = row[col.index('MIM')]
# unigene = row[col.index('UniGene')]
# pubmed = row[col.index('PubMed')]
# embl = row[col.index('EMBL')]
# embl_cds = row[col.index('EMBL-CDS')]
ensembl = row[col.index('Ensembl')].strip()
# ensembl_trs = row[col.index('Ensembl_TRS')]
# ensembl_pro = row[col.index('Ensembl_PRO')]
# other_pubmed = row[col.index('Additional PubMed')]
if ncbitaxon not in self.tax_ids:
continue
# neither empty nor a list
if geneid != '' and ';' not in geneid:
id_map[uniprotkb_ac] = 'NCBIGene:' + geneid
elif ensembl != '' and ';' not in ensembl:
id_map[uniprotkb_ac] = 'ENSEMBL:' + ensembl
else:
ummapped_uniprot += 1
LOG.info("Writing id_map out as %s", smallfile)
with open(smallfile, 'w') as yamlwriter:
yaml.dump(id_map, yamlwriter)
LOG.warning('Did not find 1:1 gene IDs for %i uniprots', ummapped_uniprot)
LOG.info(
"Acquired %i 1:1 uniprot to [entrez|ensembl] mappings", len(id_map.keys()))
return id_map
def getTestSuite(self):
import unittest
from tests.test_geneontology import GeneOntologyTestCase
test_suite = unittest.TestLoader().loadTestsFromTestCase(GeneOntologyTestCase)
return test_suite
def setUp(self):
self.source = ZFIN()
self.source.settestonly(True)
self.source.setnobnodes(True)
self._setDirToSource()
return
def setUp(self):
self.source = ZFIN('rdf_graph', True)
self.source.settestonly(True)
self._setDirToSource()
class ZFINTestCase(SourceTestCase):
def setUp(self):
self.source = ZFIN('rdf_graph', True)
self.source.settestonly(True)
self._setDirToSource()
def tearDown(self):
self.source = None
def test_mapping_of_phenotypes_to_zp_ids(self):
"""
test that code correctly uses zp_map to map phenotypes to zp ids
:return:
"""
mapping_file = "./tests/resources/zfin/zp-mapping-test-map.txt"
pheno_file = "./tests/resources/zfin/zp-mapping-test-phenotype.txt"
self.source.zp_map = self.source._load_zp_mappings(mapping_file)
pheno_dat = open(pheno_file).read().split('\t')
(fish_num, fish_name, start_stage_id, start_stage_name, end_stage_id,
end_stage_name, subterm1_id, subterm1_name, postcomp1_rel_id,
postcomp1_rel_name, superterm1_id, superterm1_name, quality_id,
quality_name, modifier, subterm2_id, subterm2_name, postcomp2_rel_id,
postcomp2_rel_name, superterm2_id, superterm2_name, pub_id, env_id) = \
pheno_dat
self.assertEqual(
self.source._map_octuple_to_phenotype(
subterm1_id, postcomp1_rel_id, superterm1_id, quality_id,
subterm2_id, postcomp2_rel_id, superterm2_id, "abnormal"),
'ZP:0022140')
def test_load_zp_mappings(self):
"""
test correct loading of zp mappings file and construction of zp_map
"""
if self.source is not None:
try:
zp_map = self.source._load_zp_mappings(
"./tests/resources/zfin/zp-mapping-test.txt")
self.assertIsInstance(
zp_map, dict, "_load_zp_mappings() didn't return dict!")
self.assertTrue(
len(zp_map) == 1,
"_load_zp_mappings() didn't return exactly one thing!")
self.assertDictEqual(
zp_map, {
'MONARCH:b308a8f1c67793a56d16': {
'post_composed_relationship_id_1': 'BFO:0000050',
'post_composed_relationship_id_2': 'BFO:0000050',
'quality_id': 'PATO:0001453',
'subterm1_id': 'ZFA:0009114',
'subterm2_id': 'GO:0005927',
'superterm1_id': 'ZFA:0001056',
'superterm2_id': 'ZFA:0001056',
'zp_id': 'ZP:0002959',
'modifier': 'PATO:0000460'
}
},
"_load_zp_mappings() " + "didn't return what I expected!")
except Exception as t_except:
LOGGER.error(t_except)
def test_make_zpkey(self):
"""
test that _make_zpkey returns correct id
"""
if self.source is not None:
try:
dummy_args = list(map(str, list(range(1,
9)))) # 1 - 8 as strings
expected_key = self.source.make_id("_".join(dummy_args))
self.assertEqual(self.source._make_zpkey(*dummy_args),
expected_key)
self.assertEqual(
self.source._make_zpkey(['0'] * 8),
self.source._make_zpkey([''] * 8),
"_make_zpkey() doesn't seem to be replacing empty " +
"strings with zeros before making key," +
"this might cause zp_map lookup issues")
except Exception as t_except:
LOGGER.error(t_except)
def test_genotype_labels(self):
"""
test that genotype label is set correctly after parse()
"""
if self.source is not None:
test_resource_dir = "../../tests/resources/zfin/"
self.source.files['fish_components']['file'] = test_resource_dir + \
"genotype-label-test-fish_components_fish.txt"
self.source.files['backgrounds']['file'] = test_resource_dir + \
"genotype-label-test-genotype_backgrounds.txt"
self.source.files['geno']['file'] = test_resource_dir + \
"genotype-label-test-genotype_features.txt"
self.source.parse()
this_iri = URIRef("http://zfin.org/ZDB-GENO-070228-3")
expect_genotype_label = "shha<sup>tbx392/tbx392</sup> (AB)"
self.assertEqual(str(self.source.testgraph.label(this_iri, None)),
expect_genotype_label)
def process_gaf(self, file, limit, id_map=None, eco_map=None):
if self.test_mode:
graph = self.testgraph
else:
graph = self.graph
model = Model(graph)
geno = Genotype(graph)
LOG.info("Processing Gene Associations from %s", file)
line_counter = 0
uniprot_hit = 0
uniprot_miss = 0
if 7955 in self.tax_ids:
zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
if 6239 in self.tax_ids:
wbase = WormBase(self.graph_type, self.are_bnodes_skized)
with gzip.open(file, 'rb') as csvfile:
filereader = csv.reader(
io.TextIOWrapper(csvfile, newline=""), delimiter='\t', quotechar='\"')
for row in filereader:
line_counter += 1
# comments start with exclamation
if re.match(r'!', ''.join(row)):
continue
if len(row) > 17 or len(row) < 15:
LOG.warning(
"Wrong number of columns %i, expected 15 or 17\n%s",
len(row), row)
continue
if 17 > len(row) >= 15:
row += [""] * (17 - len(row))
(dbase,
gene_num,
gene_symbol,
qualifier,
go_id,
ref,
eco_symbol,
with_or_from,
aspect,
gene_name,
gene_synonym,
object_type,
taxon,
date,
assigned_by,
annotation_extension,
gene_product_form_id) = row
# test for required fields
if (dbase == '' or gene_num == '' or gene_symbol == '' or
go_id == '' or ref == '' or eco_symbol == '' or
aspect == '' or object_type == '' or taxon == '' or
date == '' or assigned_by == ''):
LOG.error(
"Missing required part of annotation on row %d:\n"+'\t'
.join(row), line_counter)
continue
# deal with qualifier NOT, contributes_to, colocalizes_with
if re.search(r'NOT', qualifier):
continue
if dbase in self.localtt:
dbase = self.localtt[dbase]
uniprotid = None
gene_id = None
if dbase == 'UniProtKB':
if id_map is not None and gene_num in id_map:
gene_id = id_map[gene_num]
uniprotid = ':'.join((dbase, gene_num))
(dbase, gene_num) = gene_id.split(':')
uniprot_hit += 1
else:
# LOG.warning(
# "UniProt id %s is without a 1:1 mapping to entrez/ensembl",
# gene_num)
uniprot_miss += 1
continue
else:
gene_num = gene_num.split(':')[-1] # last
gene_id = ':'.join((dbase, gene_num))
if self.test_mode and not(
re.match(r'NCBIGene', gene_id) and
int(gene_num) in self.test_ids):
continue
model.addClassToGraph(gene_id, gene_symbol)
if gene_name != '':
model.addDescription(gene_id, gene_name)
if gene_synonym != '':
for syn in re.split(r'\|', gene_synonym):
model.addSynonym(gene_id, syn.strip())
if re.search(r'\|', taxon):
# TODO add annotations with >1 taxon
LOG.info(
">1 taxon (%s) on line %d. skipping", taxon, line_counter)
else:
tax_id = re.sub(r'taxon:', 'NCBITaxon:', taxon)
geno.addTaxon(tax_id, gene_id)
assoc = Assoc(graph, self.name)
assoc.set_subject(gene_id)
assoc.set_object(go_id)
try:
eco_id = eco_map[eco_symbol]
assoc.add_evidence(eco_id)
except KeyError:
LOG.error("Evidence code (%s) not mapped", eco_symbol)
refs = re.split(r'\|', ref)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0] # sidestep 'MGI:MGI:'
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
refg = Reference(graph, ref)
if prefix == 'PMID':
ref_type = self.globaltt['journal article']
refg.setType(ref_type)
refg.addRefToGraph()
assoc.add_source(ref)
# TODO add the source of the annotations from assigned by?
rel = self.resolve(aspect, mandatory=False)
if rel is not None and aspect == rel:
if aspect == 'F' and re.search(r'contributes_to', qualifier):
assoc.set_relationship(self.globaltt['contributes to'])
else:
LOG.error(
"Aspect: %s with qualifier: %s is not recognized",
aspect, qualifier)
elif rel is not None:
assoc.set_relationship(rel)
assoc.add_association_to_graph()
else:
LOG.warning("No predicate for association \n%s\n", str(assoc))
if uniprotid is not None:
assoc.set_description('Mapped from ' + uniprotid)
# object_type should be one of:
# protein_complex; protein; transcript; ncRNA; rRNA; tRNA;
# snRNA; snoRNA; any subtype of ncRNA in the Sequence Ontology.
# If the precise product type is unknown,
# gene_product should be used
#######################################################################
# Derive G2P Associations from IMP annotations
# in version 2.1 Pipe will indicate 'OR'
# and Comma will indicate 'AND'.
# in version 2.0, multiple values are separated by pipes
# where the pipe has been used to mean 'AND'
if eco_symbol == 'IMP' and with_or_from != '':
withitems = re.split(r'\|', with_or_from)
phenotypeid = go_id+'PHENOTYPE'
# create phenotype associations
for i in withitems:
if i == '' or re.match(
r'(UniProtKB|WBPhenotype|InterPro|HGNC)', i):
LOG.warning(
"Don't know what having a uniprot id " +
"in the 'with' column means of %s", uniprotid)
continue
i = re.sub(r'MGI\:MGI\:', 'MGI:', i)
i = re.sub(r'WB:', 'WormBase:', i)
# for worms and fish, they might give a RNAi or MORPH
# in these cases make a reagent-targeted gene
if re.search('MRPHLNO|CRISPR|TALEN', i):
targeted_gene_id = zfin.make_targeted_gene_id(gene_id, i)
geno.addReagentTargetedGene(i, gene_id, targeted_gene_id)
# TODO PYLINT why is this needed?
# Redefinition of assoc type from
# dipper.models.assoc.Association.Assoc to
# dipper.models.assoc.G2PAssoc.G2PAssoc
assoc = G2PAssoc(
graph, self.name, targeted_gene_id, phenotypeid)
elif re.search(r'WBRNAi', i):
targeted_gene_id = wbase.make_reagent_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(i, gene_id, targeted_gene_id)
assoc = G2PAssoc(
graph, self.name, targeted_gene_id, phenotypeid)
else:
assoc = G2PAssoc(graph, self.name, i, phenotypeid)
for ref in refs:
ref = ref.strip()
if ref != '':
prefix = ref.split(':')[0]
if prefix in self.localtt:
prefix = self.localtt[prefix]
ref = ':'.join((prefix, ref.split(':')[-1]))
assoc.add_source(ref)
# experimental phenotypic evidence
assoc.add_evidence(
self.globaltt['experimental phenotypic evidence'])
assoc.add_association_to_graph()
# TODO should the G2PAssoc be
# the evidence for the GO assoc?
if not self.test_mode and limit is not None and line_counter > limit:
break
uniprot_tot = (uniprot_hit + uniprot_miss)
uniprot_per = 0.0
if uniprot_tot != 0:
uniprot_per = 100.0 * uniprot_hit / uniprot_tot
LOG.info(
"Uniprot: %f.2%% of %i benifited from the 1/4 day id mapping download",
uniprot_per, uniprot_tot)
return
def process_gaf(self, file, limit, id_map=None):
if self.testMode:
g = self.testgraph
else:
g = self.graph
model = Model(g)
geno = Genotype(g)
logger.info("Processing Gene Associations from %s", file)
line_counter = 0
if 7955 in self.tax_ids:
zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
elif 6239 in self.tax_ids:
wbase = WormBase(self.graph_type, self.are_bnodes_skized)
with gzip.open(file, 'rb') as csvfile:
filereader = csv.reader(io.TextIOWrapper(csvfile, newline=""),
delimiter='\t', quotechar='\"')
for row in filereader:
line_counter += 1
# comments start with exclamation
if re.match(r'!', ''.join(row)):
continue
(db, gene_num, gene_symbol, qualifier, go_id, ref, eco_symbol,
with_or_from, aspect, gene_name, gene_synonym, object_type,
taxon, date, assigned_by, annotation_extension,
gene_product_form_id) = row
# test for required fields
if (db == '' or gene_num == '' or gene_symbol == '' or
go_id == '' or ref == '' or eco_symbol == '' or
aspect == '' or object_type == '' or taxon == '' or
date == '' or assigned_by == ''):
logger.error(
"Missing required part of annotation " +
"on row %d:\n"+'\t'.join(row),
line_counter)
continue
# deal with qualifier NOT, contributes_to, colocalizes_with
if re.search(r'NOT', qualifier):
continue
db = self.clean_db_prefix(db)
uniprotid = None
gene_id = None
if db == 'UniProtKB':
mapped_ids = id_map.get(gene_num)
if id_map is not None and mapped_ids is not None:
if len(mapped_ids) == 1:
gene_id = mapped_ids[0]
uniprotid = ':'.join((db, gene_num))
gene_num = re.sub(r'\w+\:', '', gene_id)
elif len(mapped_ids) > 1:
# logger.warning(
# "Skipping gene id mapped for >1 gene %s -> %s",
# gene_num, str(mapped_ids))
continue
else:
continue
elif db == 'MGI':
gene_num = re.sub(r'MGI:', '', gene_num)
gene_id = ':'.join((db, gene_num))
gene_id = re.sub(r'MGI\:MGI\:', 'MGI:', gene_id)
else:
gene_id = ':'.join((db, gene_num))
if self.testMode \
and not(
re.match(r'NCBIGene', gene_id) and
int(gene_num) in self.test_ids):
continue
model.addClassToGraph(gene_id, gene_symbol)
if gene_name != '':
model.addDescription(gene_id, gene_name)
if gene_synonym != '':
for s in re.split(r'\|', gene_synonym):
model.addSynonym(gene_id, s.strip())
if re.search(r'\|', taxon):
# TODO add annotations with >1 taxon
logger.info(">1 taxon (%s) on line %d. skipping", taxon,
line_counter)
else:
tax_id = re.sub(r'taxon:', 'NCBITaxon:', taxon)
geno.addTaxon(tax_id, gene_id)
assoc = Assoc(g, self.name)
assoc.set_subject(gene_id)
assoc.set_object(go_id)
eco_id = self.map_go_evidence_code_to_eco(eco_symbol)
if eco_id is not None:
assoc.add_evidence(eco_id)
refs = re.split(r'\|', ref)
for r in refs:
r = r.strip()
if r != '':
prefix = re.split(r':', r)[0]
r = re.sub(prefix, self.clean_db_prefix(prefix), r)
r = re.sub(r'MGI\:MGI\:', 'MGI:', r)
ref = Reference(g, r)
if re.match(r'PMID', r):
ref_type = Reference.ref_types['journal_article']
ref.setType(ref_type)
ref.addRefToGraph()
assoc.add_source(r)
# TODO add the source of the annotations from assigned by?
aspect_rel_map = {
'P': model.object_properties['involved_in'], # involved in
'F': model.object_properties['enables'], # enables
'C': model.object_properties['part_of'] # part of
}
if aspect not in aspect_rel_map:
logger.error("Aspect not recognized: %s", aspect)
rel = aspect_rel_map.get(aspect)
if aspect == 'F' and re.search(r'contributes_to', qualifier):
rel = model.object_properties['contributes_to']
assoc.set_relationship(rel)
if uniprotid is not None:
assoc.set_description('Mapped from '+uniprotid)
# object_type should be one of:
# protein_complex; protein; transcript; ncRNA; rRNA; tRNA;
# snRNA; snoRNA; any subtype of ncRNA in the Sequence Ontology.
# If the precise product type is unknown,
# gene_product should be used
assoc.add_association_to_graph()
# Derive G2P Associations from IMP annotations
# in version 2.1 Pipe will indicate 'OR'
# and Comma will indicate 'AND'.
# in version 2.0, multiple values are separated by pipes
# where the pipe has been used to mean 'AND'
if eco_symbol == 'IMP' and with_or_from != '':
withitems = re.split(r'\|', with_or_from)
phenotypeid = go_id+'PHENOTYPE'
# create phenotype associations
for i in withitems:
if i == '' or \
re.match(
r'(UniProtKB|WBPhenotype|InterPro|HGNC)',
i):
logger.warning(
"Don't know what having a uniprot id " +
"in the 'with' column means of %s",
uniprotid)
continue
i = re.sub(r'MGI\:MGI\:', 'MGI:', i)
i = re.sub(r'WB:', 'WormBase:', i)
# for worms and fish, they might give a RNAi or MORPH
# in these cases make a reagent-targeted gene
if re.search('MRPHLNO|CRISPR|TALEN', i):
targeted_gene_id = zfin.make_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(i, gene_id,
targeted_gene_id)
# TODO PYLINT why is this:
# Redefinition of assoc type from
# dipper.models.assoc.Association.Assoc to
# dipper.models.assoc.G2PAssoc.G2PAssoc
assoc = G2PAssoc(g, self.name, targeted_gene_id,
phenotypeid)
elif re.search(r'WBRNAi', i):
targeted_gene_id = \
wbase.make_reagent_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(
i, gene_id, targeted_gene_id)
assoc = G2PAssoc(
g, self.name, targeted_gene_id, phenotypeid)
else:
assoc = G2PAssoc(g, self.name, i, phenotypeid)
for r in refs:
r = r.strip()
if r != '':
prefix = re.split(r':', r)[0]
r = re.sub(
prefix, self.clean_db_prefix(prefix), r)
r = re.sub(r'MGI\:MGI\:', 'MGI:', r)
assoc.add_source(r)
# experimental phenotypic evidence
assoc.add_evidence("ECO:0000059")
assoc.add_association_to_graph()
# TODO should the G2PAssoc be
# the evidence for the GO assoc?
if not self.testMode and \
limit is not None and line_counter > limit:
break
return
def process_gaf(self, file, limit, id_map=None):
if self.testMode:
g = self.testgraph
else:
g = self.graph
model = Model(g)
geno = Genotype(g)
logger.info("Processing Gene Associations from %s", file)
line_counter = 0
if 7955 in self.tax_ids:
zfin = ZFIN(self.graph_type, self.are_bnodes_skized)
elif 6239 in self.tax_ids:
wbase = WormBase(self.graph_type, self.are_bnodes_skized)
with gzip.open(file, 'rb') as csvfile:
filereader = csv.reader(io.TextIOWrapper(csvfile, newline=""),
delimiter='\t',
quotechar='\"')
for row in filereader:
line_counter += 1
# comments start with exclamation
if re.match(r'!', ''.join(row)):
continue
(db, gene_num, gene_symbol, qualifier, go_id, ref, eco_symbol,
with_or_from, aspect, gene_name, gene_synonym, object_type,
taxon, date, assigned_by, annotation_extension,
gene_product_form_id) = row
# test for required fields
if (db == '' or gene_num == '' or gene_symbol == ''
or go_id == '' or ref == '' or eco_symbol == ''
or aspect == '' or object_type == '' or taxon == ''
or date == '' or assigned_by == ''):
logger.error(
"Missing required part of annotation " +
"on row %d:\n" + '\t'.join(row), line_counter)
continue
# deal with qualifier NOT, contributes_to, colocalizes_with
if re.search(r'NOT', qualifier):
continue
db = self.clean_db_prefix(db)
uniprotid = None
gene_id = None
if db == 'UniProtKB':
mapped_ids = id_map.get(gene_num)
if id_map is not None and mapped_ids is not None:
if len(mapped_ids) == 1:
gene_id = mapped_ids[0]
uniprotid = ':'.join((db, gene_num))
gene_num = re.sub(r'\w+\:', '', gene_id)
elif len(mapped_ids) > 1:
# logger.warning(
# "Skipping gene id mapped for >1 gene %s -> %s",
# gene_num, str(mapped_ids))
continue
else:
continue
elif db == 'MGI':
gene_num = re.sub(r'MGI:', '', gene_num)
gene_id = ':'.join((db, gene_num))
gene_id = re.sub(r'MGI\:MGI\:', 'MGI:', gene_id)
else:
gene_id = ':'.join((db, gene_num))
if self.testMode \
and not(
re.match(r'NCBIGene', gene_id) and
int(gene_num) in self.test_ids):
continue
model.addClassToGraph(gene_id, gene_symbol)
if gene_name != '':
model.addDescription(gene_id, gene_name)
if gene_synonym != '':
for s in re.split(r'\|', gene_synonym):
model.addSynonym(gene_id, s.strip())
if re.search(r'\|', taxon):
# TODO add annotations with >1 taxon
logger.info(">1 taxon (%s) on line %d. skipping", taxon,
line_counter)
else:
tax_id = re.sub(r'taxon:', 'NCBITaxon:', taxon)
geno.addTaxon(tax_id, gene_id)
assoc = Assoc(g, self.name)
assoc.set_subject(gene_id)
assoc.set_object(go_id)
eco_id = self.map_go_evidence_code_to_eco(eco_symbol)
if eco_id is not None:
assoc.add_evidence(eco_id)
refs = re.split(r'\|', ref)
for r in refs:
r = r.strip()
if r != '':
prefix = re.split(r':', r)[0]
r = re.sub(prefix, self.clean_db_prefix(prefix), r)
r = re.sub(r'MGI\:MGI\:', 'MGI:', r)
ref = Reference(g, r)
if re.match(r'PMID', r):
ref_type = Reference.ref_types['journal_article']
ref.setType(ref_type)
ref.addRefToGraph()
assoc.add_source(r)
# TODO add the source of the annotations from assigned by?
aspect_rel_map = {
'P': model.object_properties['involved_in'], # involved in
'F': model.object_properties['enables'], # enables
'C': model.object_properties['part_of'] # part of
}
if aspect not in aspect_rel_map:
logger.error("Aspect not recognized: %s", aspect)
rel = aspect_rel_map.get(aspect)
if aspect == 'F' and re.search(r'contributes_to', qualifier):
rel = model.object_properties['contributes_to']
assoc.set_relationship(rel)
if uniprotid is not None:
assoc.set_description('Mapped from ' + uniprotid)
# object_type should be one of:
# protein_complex; protein; transcript; ncRNA; rRNA; tRNA;
# snRNA; snoRNA; any subtype of ncRNA in the Sequence Ontology.
# If the precise product type is unknown,
# gene_product should be used
assoc.add_association_to_graph()
# Derive G2P Associations from IMP annotations
# in version 2.1 Pipe will indicate 'OR'
# and Comma will indicate 'AND'.
# in version 2.0, multiple values are separated by pipes
# where the pipe has been used to mean 'AND'
if eco_symbol == 'IMP' and with_or_from != '':
withitems = re.split(r'\|', with_or_from)
phenotypeid = go_id + 'PHENOTYPE'
# create phenotype associations
for i in withitems:
if i == '' or \
re.match(
r'(UniProtKB|WBPhenotype|InterPro|HGNC)',
i):
logger.warning(
"Don't know what having a uniprot id " +
"in the 'with' column means of %s", uniprotid)
continue
i = re.sub(r'MGI\:MGI\:', 'MGI:', i)
i = re.sub(r'WB:', 'WormBase:', i)
# for worms and fish, they might give a RNAi or MORPH
# in these cases make a reagent-targeted gene
if re.search('MRPHLNO|CRISPR|TALEN', i):
targeted_gene_id = zfin.make_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(
i, gene_id, targeted_gene_id)
# TODO PYLINT why is this:
# Redefinition of assoc type from
# dipper.models.assoc.Association.Assoc to
# dipper.models.assoc.G2PAssoc.G2PAssoc
assoc = G2PAssoc(g, self.name, targeted_gene_id,
phenotypeid)
elif re.search(r'WBRNAi', i):
targeted_gene_id = \
wbase.make_reagent_targeted_gene_id(
gene_id, i)
geno.addReagentTargetedGene(
i, gene_id, targeted_gene_id)
assoc = G2PAssoc(g, self.name, targeted_gene_id,
phenotypeid)
else:
assoc = G2PAssoc(g, self.name, i, phenotypeid)
for r in refs:
r = r.strip()
if r != '':
prefix = re.split(r':', r)[0]
r = re.sub(prefix,
self.clean_db_prefix(prefix), r)
r = re.sub(r'MGI\:MGI\:', 'MGI:', r)
assoc.add_source(r)
# experimental phenotypic evidence
assoc.add_evidence("ECO:0000059")
assoc.add_association_to_graph()
# TODO should the G2PAssoc be
# the evidence for the GO assoc?
if not self.testMode and \
limit is not None and line_counter > limit:
break
return
def parse(self, limit=None):
zfin_parser = ZFIN(self.graph_type, self.are_bnodes_skized)
model = Model(self.graph)
src_key = 'zpmap' # keep same-as zfin.files[key]
zfin_parser.zp_map = zfin_parser._load_zp_mappings(src_key)
src_key = 'g2p_clean'
raw = '/'.join((self.rawdir, self.files[src_key]['file']))
LOG.info("Processing clean Geno to Pheno from file: %s", raw)
col = self.files[src_key]['columns']
collen = len(col)
with open(raw, 'r', encoding="utf8") as csvfile:
reader = csv.reader(csvfile, delimiter='\t', quotechar='\"')
for row in reader:
if len(row) != collen:
LOG.warning('Row: %i has unexpected format',
reader.line_num)
# internal_id = row[col.index('ID')]
# symbol = row[col.index('Gene Symbol')]
gene_id = row[col.index('Gene ID')]
subterm1_id = row[col.index(
'Affected Structure or Process 1 subterm ID')]
# subterm1_label = row[col.index(
# 'Affected Structure or Process 1 subterm Name')]
pc_rel_id = row[col.index(
'Post-composed Relationship ID')].strip()
# pc_rel_label = row[col.index('Post-composed Relationship Name')]
superterm1_id = row[col.index(
'Affected Structure or Process 1 superterm ID')].strip()
# superterm1_label = row[col.index(
# 'Affected Structure or Process 1 superterm Name')]
quality_id = row[col.index('Phenotype Keyword ID')].strip()
# quality_name = row[col.index('Phenotype Keyword Name')]
modifier = row[col.index('Phenotype Tag')].strip()
subterm2_id = row[col.index(
'Affected Structure or Process 2 subterm ID')].strip()
# subterm2_label = row[col.index(
# 'Affected Structure or Process 2 subterm name')]
pc_rel2_id = row[col.index(
'Post-composed Relationship (rel) ID')]
# pc_rel2_label = row[col.index(
# 'Post-composed Relationship (rel) Name')]
superterm2_id = row[col.index(
'Affected Structure or Process 2 superterm ID')].strip()
# superterm2_label = row[col.index(
# 'Affected Structure or Process 2 superterm name')]
# fish_id = row[col.index('Fish ID')]
# fish_label = row[col.index('Fish Display Name')]
start_stage = row[col.index('Start Stage ID')]
# end_stage = row[col.index('End Stage ID')]
# environment = row[col.index('Fish Environment ID')]
pub_id = row[col.index('Publication ID')].strip()
# figure_id = row[col.index('Figure ID')]
if modifier != 'abnormal':
LOG.warning(
"skipping phenotype with modifier %s != abnormal ",
modifier)
continue
zp_id = zfin_parser._map_octuple_to_phenotype(
subterm1_id, pc_rel_id, superterm1_id, quality_id,
subterm2_id, pc_rel2_id, superterm2_id, modifier)
gene_curie = "ZFIN:{0}".format(gene_id)
model.makeLeader(gene_curie)
pub_curie = "ZFIN:{0}".format(pub_id)
if zp_id:
assoc = G2PAssoc(self.graph, self.name, gene_curie, zp_id)
if pub_id:
reference = Reference(self.graph, pub_curie,
self.globaltt['document'])
reference.addRefToGraph()
assoc.add_source(pub_curie)
assoc.add_evidence(
self.globaltt['experimental phenotypic evidence'])
assoc.add_association_to_graph()