def convert_gff_to_bed(input_stream, output_stream, attribute_to_use=None):
"""
Convert GFF lines from input_stream to
BED output_stream.
"""
for line in input_stream:
if line.startswith("#"):
# Skip GFF comments
continue
gff_fields = line.strip().split("\t")
chrom = gff_fields[0]
rec_source = gff_fields[1]
rec_type = gff_fields[2]
start = int(gff_fields[3])
end = int(gff_fields[4])
strand = gff_fields[6]
if start > end:
# Flip coordinates if start > end
start, end = end, start
# Convert start to be 0-based
start = start - 1
# Convert coordinates back to strings
start, end = str(start), str(end)
rec_attributes = gff_fields[-1]
name = rec_type
score = "1"
if rec_attributes != ".":
# If there's an ID= attribute present,
# use its value as the BED entry's name
attributes = utils.parse_attributes(rec_attributes)
if attribute_to_use is not None:
if attribute_to_use in attributes:
# Use gene_id as ID if found
name = attributes[attribute_to_use]
else:
print "WARNING: %s attribute not found" \
%(attribute_to_use)
name = attributes["ID"]
elif "ID" in attributes:
name = attributes["ID"]
bed_line = make_bed_line(chrom, start, end, name, score, strand)
output_stream.write("%s\n" % (bed_line))
def convert_gff_to_bed(input_stream, output_stream, attribute_to_use=None):
"""
Convert GFF lines from input_stream to
BED output_stream.
"""
for line in input_stream:
if line.startswith("#"):
# Skip GFF comments
continue
gff_fields = line.strip().split("\t")
chrom = gff_fields[0]
rec_source = gff_fields[1]
rec_type = gff_fields[2]
start = int(gff_fields[3])
end = int(gff_fields[4])
strand = gff_fields[6]
if start > end:
# Flip coordinates if start > end
start, end = end, start
# Convert start to be 0-based
start = start - 1
# Convert coordinates back to strings
start, end = str(start), str(end)
rec_attributes = gff_fields[-1]
name = rec_type
score = "1"
if rec_attributes != ".":
# If there's an ID= attribute present,
# use its value as the BED entry's name
attributes = utils.parse_attributes(rec_attributes)
if attribute_to_use is not None:
if attribute_to_use in attributes:
# Use gene_id as ID if found
name = attributes[attribute_to_use]
else:
print "WARNING: %s attribute not found" % (attribute_to_use)
name = attributes["ID"]
elif "ID" in attributes:
name = attributes["ID"]
bed_line = make_bed_line(chrom, start, end, name, score, strand)
output_stream.write("%s\n" % (bed_line))
def annotate_gff_with_genes(args):
"""
Annotate GFF with genes table.
"""
gff_fname = utils.pathify(args.gff_filename)
if not os.path.isfile(gff_fname):
raise Exception, "Cannot find %s" % (gff_fname)
table_fname = utils.pathify(args.table_filename)
if not os.path.isfile(table_fname):
raise Exception, "Cannot find %s" % (table_fname)
table_bed = get_table_as_bedtool(table_fname)
# Get BedTool for events, containing only the gene entries
all_events_bed = pybedtools.BedTool(gff_fname)
event_genes = \
all_events_bed.filter(lambda entry: entry.fields[2] == "gene")
print "Determining overlap between events and genes..."
# Intersect event genes with gene txStart/txEnd
intersected_bed = \
event_genes.intersect(table_bed, wb=True, s=True, f=1)
# Map event genes to their IDs
#
# event_gene1 -> refseq -> value
# -> ensgene -> value
# event_gene2 -> refseq ->
# ...
event_genes_to_info = \
defaultdict(lambda: defaultdict(list))
for entry in intersected_bed:
event_gene_attrs = utils.parse_attributes(entry.fields[8])
event_gene_str = event_gene_attrs["ID"]
gene_info_field = entry.fields[-1]
# Strip semicolon of ID attributes
if gene_info_field.endswith(";"):
gene_info_field = gene_info_field[0:-1]
# Convert attributes into dictionary
gene_info = utils.parse_attributes(gene_info_field)
ensgene_id = gene_info["ensg_id"]
refseq_id = gene_info["refseq_id"]
gene_symbol = gene_info["gsymbol"]
# Skip null entries
if not is_null_id(ensgene_id):
event_genes_to_info[event_gene_str]["ensg_id"].append(ensgene_id)
if not is_null_id(refseq_id):
event_genes_to_info[event_gene_str]["refseq_id"].append(refseq_id)
if not is_null_id(gene_symbol):
event_genes_to_info[event_gene_str]["gsymbol"].append(gene_symbol)
# Incorporate the gene information into the GFF and output it
# it using gffutils
print "Loading events into GFF database..."
events_db = gffutils.create_db(gff_fname, ":memory:", verbose=False)
output_fname = gff_fname
events_out = gffwriter.GFFWriter(output_fname, in_place=True)
print " - Outputting annotated GFF to: %s" % (output_fname)
def new_recs():
for gene_recs in list(events_db.iter_by_parent_childs()):
gene_rec = gene_recs[0]
event_id = gene_rec.id
# Use existing IDs if present
if "ensgene_id" in gene_rec.attributes:
ensgene_id = gene_rec.attributes["ensg_id"][0]
else:
ensgene_id = "NA"
if "refseq_id" in gene_rec.attributes:
refseq_id = gene_rec.attributes["refseq_id"][0]
else:
refseq_id = "NA"
if "gene_symbol" in gene_rec.attributes:
gene_symbol = gene_rec.attributes["gsymbol"][0]
else:
gene_symbol = "NA"
if event_id in event_genes_to_info:
event_info = event_genes_to_info[event_id]
ensgene_ids = \
utils.unique_list(event_info["ensg_id"])
if len(ensgene_ids) > 0 and ensgene_ids[0] != "NA":
ensgene_id = ",".join(ensgene_ids)
refseq_ids = \
utils.unique_list(event_info["refseq_id"])
if len(refseq_ids) > 0 and refseq_ids[0] != "NA":
refseq_id = ",".join(refseq_ids)
gene_symbols = \
utils.unique_list(event_info["gsymbol"])
if len(gene_symbols) > 0 and gene_symbols[0] != "NA":
gene_symbol = ",".join(gene_symbols)
gene_rec.attributes["ensg_id"] = [ensgene_id]
gene_rec.attributes["refseq_id"] = [refseq_id]
gene_rec.attributes["gsymbol"] = [gene_symbol]
# Yield all the gene's records
for g in gene_recs:
yield g
t1 = time.time()
print "Creating annotated GFF database..."
annotated_db = gffutils.create_db(new_recs(), ":memory:", verbose=False)
t2 = time.time()
print "Creation took %.2f secs" % (t2 - t1)
# Write to file
print "Writing annotated GFF to file..."
for gene_rec in annotated_db.all_features(featuretype="gene"):
events_out.write_gene_recs(annotated_db, gene_rec.id)
events_out.close()
def annotate_gff_with_genes(args):
"""
Annotate GFF with genes table.
"""
gff_fname = utils.pathify(args.gff_filename)
if not os.path.isfile(gff_fname):
raise Exception, "Cannot find %s" %(gff_fname)
table_fname = utils.pathify(args.table_filename)
if not os.path.isfile(table_fname):
raise Exception, "Cannot find %s" %(table_fname)
table_bed = get_table_as_bedtool(table_fname)
# Get BedTool for events, containing only the gene entries
all_events_bed = pybedtools.BedTool(gff_fname)
event_genes = \
all_events_bed.filter(lambda entry: entry.fields[2] == "gene")
print "Determining overlap between events and genes..."
# Intersect event genes with gene txStart/txEnd
intersected_bed = \
event_genes.intersect(table_bed, wb=True, s=True, f=1)
# Map event genes to their IDs
#
# event_gene1 -> refseq -> value
# -> ensgene -> value
# event_gene2 -> refseq ->
# ...
event_genes_to_info = \
defaultdict(lambda: defaultdict(list))
for entry in intersected_bed:
event_gene_attrs = utils.parse_attributes(entry.fields[8])
event_gene_str = event_gene_attrs["ID"]
gene_info_field = entry.fields[-1]
# Strip semicolon of ID attributes
if gene_info_field.endswith(";"):
gene_info_field = gene_info_field[0:-1]
# Convert attributes into dictionary
gene_info = utils.parse_attributes(gene_info_field)
ensgene_id = gene_info["ensg_id"]
refseq_id = gene_info["refseq_id"]
gene_symbol = gene_info["gsymbol"]
# Skip null entries
if not is_null_id(ensgene_id):
event_genes_to_info[event_gene_str]["ensg_id"].append(ensgene_id)
if not is_null_id(refseq_id):
event_genes_to_info[event_gene_str]["refseq_id"].append(refseq_id)
if not is_null_id(gene_symbol):
event_genes_to_info[event_gene_str]["gsymbol"].append(gene_symbol)
# Incorporate the gene information into the GFF and output it
# it using gffutils
print "Loading events into GFF database..."
events_db = gffutils.create_db(gff_fname, ":memory:",
verbose=False)
output_fname = gff_fname
events_out = gffwriter.GFFWriter(output_fname,
in_place=True)
print " - Outputting annotated GFF to: %s" %(output_fname)
def new_recs():
for gene_recs in list(events_db.iter_by_parent_childs()):
gene_rec = gene_recs[0]
event_id = gene_rec.id
# Use existing IDs if present
if "ensgene_id" in gene_rec.attributes:
ensgene_id = gene_rec.attributes["ensg_id"][0]
else:
ensgene_id = "NA"
if "refseq_id" in gene_rec.attributes:
refseq_id = gene_rec.attributes["refseq_id"][0]
else:
refseq_id = "NA"
if "gene_symbol" in gene_rec.attributes:
gene_symbol = gene_rec.attributes["gsymbol"][0]
else:
gene_symbol = "NA"
if event_id in event_genes_to_info:
event_info = event_genes_to_info[event_id]
ensgene_ids = \
utils.unique_list(event_info["ensg_id"])
if len(ensgene_ids) > 0 and ensgene_ids[0] != "NA":
ensgene_id = ",".join(ensgene_ids)
refseq_ids = \
utils.unique_list(event_info["refseq_id"])
if len(refseq_ids) > 0 and refseq_ids[0] != "NA":
refseq_id = ",".join(refseq_ids)
gene_symbols = \
utils.unique_list(event_info["gsymbol"])
if len(gene_symbols) > 0 and gene_symbols[0] != "NA":
gene_symbol = ",".join(gene_symbols)
gene_rec.attributes["ensg_id"] = [ensgene_id]
gene_rec.attributes["refseq_id"] = [refseq_id]
gene_rec.attributes["gsymbol"] = [gene_symbol]
# Yield all the gene's records
for g in gene_recs:
yield g
t1 = time.time()
print "Creating annotated GFF database..."
annotated_db = gffutils.create_db(new_recs(), ":memory:",
verbose=False)
t2 = time.time()
print "Creation took %.2f secs" %(t2 - t1)
# Write to file
print "Writing annotated GFF to file..."
for gene_rec in annotated_db.all_features(featuretype="gene"):
events_out.write_gene_recs(annotated_db, gene_rec.id)
events_out.close()