def rebase(parent, child, interpro=False, protein2dna=False):
child_features = __get_features(child, interpro=interpro)
for rec in GFF.parse(parent):
# TODO, replace with recursion in case it's matched against a
# non-parent feature. We're cheating a bit here right now...
replacement_features = []
for feature in rec.features:
if feature.id in child_features:
new_subfeatures = child_features[feature.id]
# TODO: update starts
fixed_subfeatures = []
for x in new_subfeatures:
# Then update the location of the actual feature
__update_feature_location(x, feature, protein2dna)
if interpro:
for y in ('status', 'Target'):
try:
del x.qualifiers[y]
except:
pass
fixed_subfeatures.append(x)
replacement_features.extend(fixed_subfeatures)
# We do this so we don't include the original set of features that we
# were rebasing against in our result.
rec.features = replacement_features
GFF.write([rec], sys.stdout)
def to_gff(self, filename):
"""
Export to GFF format, saving to the specified filename.
"""
records = []
for fragment in self.__genome.fragments.all():
fragment = fragment.indexed_fragment()
seq = Seq(fragment.sequence)
rec = SeqRecord(seq, "%s" % (fragment.name,))
features = []
for annotation in fragment.annotations():
# FeatureLocation first bp is AfterPosition, so -1
loc = FeatureLocation(annotation.base_first-1, annotation.base_last)
qualifiers = {'name': annotation.feature.name}
feature = SeqFeature(loc,
type=annotation.feature.type,
strand=1,
qualifiers=qualifiers)
features.append(feature)
rec.features = features
records.append(rec)
with open(filename, "w") as out_handle:
GFF.write(records, out_handle, include_fasta=True)
def CpGIslandsToGFF(island_location):
# Output methylation regions (CpG Islands, namely) to a GFF3 compliant file
out_file = os.getcwd() \
+ '/' \
+ os.path.splitext(base)[0] \
+ '.gff'
seq = cur_record.seq
rec = SeqRecord(seq, "ID1")
qualifiers = {"source": "bssimulation", "score": '.', "ID": cur_record.name}
sub_qualifiers = {"source": "bssimulation"}
top_feature = SeqFeature(FeatureLocation(0, len(cur_record)), type="region", strand=0,
qualifiers=qualifiers)
for i in island_location:
begin = int(i[0] - i[1]/2)
end = int(i[0] + i[1]/2)
top_feature.sub_features.append(SeqFeature(FeatureLocation(begin, end),
type="CpG_island",
strand=0,
qualifiers=sub_qualifiers))
rec.features = [top_feature]
with open(out_file, "w") as out_handle:
GFF.write([rec], out_handle)
def to_gff_file(self, file):
"""
Export to GFF format, saving to provided file like object.
"""
records = []
for fragment in self.__genome.fragments.all():
fragment = fragment.indexed_fragment()
seq = Seq(fragment.sequence)
rec = SeqRecord(seq, "%s" % (fragment.name,))
features = []
for annotation in fragment.annotations():
# FeatureLocation first bp is AfterPosition, so -1
loc = FeatureLocation(annotation.base_first - 1, annotation.base_last)
qualifiers = {'name': annotation.feature.name}
strand = annotation.feature.strand
feature = SeqFeature(loc,
type=annotation.feature.type,
strand=0 if strand is None else strand,
qualifiers=qualifiers)
features.append(feature)
rec.features = features
records.append(rec)
GFF.write(records, file, include_fasta=True)
def t_write_from_recs(self):
"""Write out GFF3 from SeqRecord inputs.
"""
seq = Seq("GATCGATCGATCGATCGATC")
rec = SeqRecord(seq, "ID1")
qualifiers = {"source": "prediction", "score": 10.0, "other": ["Some", "annotations"],
"ID": "gene1"}
sub_qualifiers = {"source": "prediction"}
top_feature = SeqFeature(FeatureLocation(0, 20), type="gene", strand=1,
qualifiers=qualifiers)
top_feature.sub_features = [SeqFeature(FeatureLocation(0, 5), type="exon", strand=1,
qualifiers=sub_qualifiers),
SeqFeature(FeatureLocation(15, 20), type="exon", strand=1,
qualifiers=sub_qualifiers)]
rec.features = [top_feature]
out_handle = StringIO.StringIO()
GFF.write([rec], out_handle)
wrote_info = out_handle.getvalue().split("\n")
assert wrote_info[0] == "##gff-version 3"
assert wrote_info[1] == "##sequence-region ID1 1 20"
assert wrote_info[2].split("\t") == ['ID1', 'prediction', 'gene', '1',
'20', '10.0', '+', '.',
'other=Some,annotations;ID=gene1']
assert wrote_info[3].split("\t") == ['ID1', 'prediction', 'exon', '1', '5',
'.', '+', '.', 'Parent=gene1']
def main():
from argparse import ArgumentParser
parser = ArgumentParser("Convert SAM to GFF3 format using BCBio GFF")
parser.add_argument("sam_filename")
parser.add_argument("-i", "--input_fasta", default=None, help="(Optional) input fasta. If given, coverage will be calculated.")
parser.add_argument("-s", "--source", required=True, help="source name (ex: hg38, mm10)")
args = parser.parse_args()
if not args.sam_filename.endswith('.sam'):
print >> sys.stderr, "Only accepts files ending in .sam. Abort!"
sys.exit(-1)
prefix = args.sam_filename[:-4]
output_gff3 = prefix + '.gff3'
q_dict = None
if args.input_fasta is not None:
q_dict = dict((r.id, len(r.seq)) for r in SeqIO.parse(open(args.input_fasta), 'fasta'))
with open(output_gff3, 'w') as f:
recs = [convert_sam_rec_to_gff3_rec(r0, args.source) for r0 in GMAPSAMReader(args.sam_filename, True, query_len_dict=q_dict)]
BCBio_GFF.write(filter(lambda x: x is not None, recs), f)
print >> sys.stderr, "Output written to {0}.".format(output_gff3)
def rebase(parent, child, interpro=False, protein2dna=False):
child_features = __get_features(child, interpro=interpro)
for rec in GFF.parse(parent):
replacement_features = []
for feature in feature_lambda(
rec.features,
feature_test_qual_value,
{
'qualifier': 'ID',
'attribute_list': child_features.keys(),
},
subfeatures=False):
new_subfeatures = child_features[feature.id]
fixed_subfeatures = []
for x in new_subfeatures:
# Then update the location of the actual feature
__update_feature_location(x, feature, protein2dna)
if interpro:
for y in ('status', 'Target'):
try:
del x.qualifiers[y]
except:
pass
fixed_subfeatures.append(x)
replacement_features.extend(fixed_subfeatures)
# We do this so we don't include the original set of features that we
# were rebasing against in our result.
rec.features = replacement_features
rec.annotations = {}
GFF.write([rec], sys.stdout)
def load_gff(gff):
genes = defaultdict(list)
gene_exon_positions = defaultdict(lambda: defaultdict(tuple))
try:
with open(gff) as g:
for line in g:
if line.startswith('#') or 'contig' in line:
continue
feature = GFF(line)
gene_id = get_gene_attribute(feature, "ID")
if feature.featuretype == 'exon':
gene_exon_positions[feature.genename][gene_id] = (feature.start, feature.end)
if feature.featuretype == 'CDS':
for exon in gene_exon_positions[feature.genename]:
e = gene_exon_positions[feature.genename][exon]
if e[0] <= feature.start <= e[1] and e[0] <= feature.end <= e[1]:
gene_id = exon + "_CDS"
if gene_id is None:
print("No gene id for CDS found", feature, end="")
feature.id = gene_id
genes[feature.genename].append(feature)
except IOError:
print("Failed to load GFF file {}".format(gff))
sys.exit()
return genes
def genbank_to_gff(self,
genbank_file):
from Bio import SeqIO
from BCBio import GFF
gff_file = "%s.gff" % (os.path.splitext(genbank_file)[0],)
with open(gff_file, "w") as out_handle:
GFF.write(SeqIO.parse(genbank_file, "genbank"), out_handle, include_fasta=True)
return dict(gff_file=gff_file)
def main(gb_file,include_fasta=None):
out_file = "%s.gff" % os.path.splitext(gb_file)[0]
inc_fasta = False
if include_fasta is not None:
if include_fasta.lower() in ("true","yes","1"):
inc_fasta = True
with open(out_file, "w") as out_handle:
GFF.write(SeqIO.parse(gb_file, "genbank"), out_handle, inc_fasta)
def genbank_to_gff(gb_file):
"""Convert GenBank file to GFF for IGV display.
"""
max_size = 1e4
gff_file = "%s.gff3" % os.path.splitext(gb_file)[0]
if not os.path.exists(gff_file):
with open(gb_file) as in_handle:
with open(gff_file, "w") as out_handle:
gb_iterator = SeqIO.parse(in_handle, "genbank")
GFF.write(_filter_features(gb_iterator, max_size),
out_handle)
def embl2gff(dat, org, gff):
"""
Parse embl file and estract mature miRNA location information.
"""
# extract records
dat_parser = SeqIO.parse(dat, "embl")
# extract organism specific miRNAs
org_mirnas = [mirna for mirna in dat_parser if mirna.name.startswith(org)]
for mirna in org_mirnas:
mirna.id = mirna.name
GFF.write(org_mirnas, gff)
def to_GFF(args):
"""
Convert a GenBank or EMBL file to GFF
Biopython does not natively support GFF
Can be useful for QUAST (Quality Assessment Tool for Genome Assemblies)
:param args: an argparse args list
"""
in_type = args.inFormat.lower()
with open(args.input) as fin, open(args.output, "w") as fout:
GFF.write(SeqIO.parse(fin, in_type), fout)
def t_write_seqrecord(self):
"""Write single SeqRecords.
"""
seq = Seq("GATCGATCGATCGATCGATC")
rec = SeqRecord(seq, "ID1")
qualifiers = {"source": "prediction", "score": 10.0, "other": ["Some", "annotations"],
"ID": "gene1"}
rec.features = [SeqFeature(FeatureLocation(0, 20), type="gene", strand=1,
qualifiers=qualifiers)]
out_handle = StringIO.StringIO()
GFF.write([rec], out_handle, include_fasta=True)
wrote_info = out_handle.getvalue().split("\n")
gff_line = wrote_info[2]
assert gff_line.split("\t")[0] == "ID1"
def t_gff3_to_gff3(self):
"""Read in and write out GFF3 without any loss of information.
"""
recs = SeqIO.to_dict(GFF.parse(self._test_gff_file))
out_handle = StringIO.StringIO()
GFF.write(recs.values(), out_handle)
wrote_handle = StringIO.StringIO(out_handle.getvalue())
recs_two = SeqIO.to_dict(GFF.parse(wrote_handle))
orig_rec = recs.values()[0]
re_rec = recs.values()[0]
assert len(orig_rec.features) == len(re_rec.features)
for i, orig_f in enumerate(orig_rec.features):
assert str(orig_f) == str(re_rec.features[i])
def handle(self, *args, **options):
organism, created = Organism.objects.get_or_create(
common_name=options['organism_name'],
taxon=options['taxon'],
ebi_id=options['ebi_id']
)
for record in SeqIO.parse(options['fasta'], "fasta"):
refseq, created = RefSeq.objects.get_or_create(
name=record.id,
length=len(record.seq),
organism=organism
)
for rec in GFF.parse(options['gff3']):
rs = RefSeq.objects.get(name=rec.id, organism=organism)
for feat in rec.features:
if feat.type != 'gene':
continue
gene, created = Gene.objects.get_or_create(
start=feat.location.start,
end=feat.location.end,
strand=feat.location.strand,
refseq=rs,
db_object_id=feat.id,
db_object_symbol=feat.id
)
def get_gff_dict(gfffile):
"""Creates a dictionary with product information from given gff file.
Returns dictionary. Dictionary key is the contig id, values are products for the contig."""
out_dict = {}
for rec in GFF.parse(gfffile):
# Add features if there are any
if rec.features > 0:
gff_info = None
# Add all features
# Features are separated by ,
# example:
# featuretype;product;product,featuretype;product
# or
# CDS;protein3;protein31,CDS;protein3
for f in rec.features:
if len(f.qualifiers['product']) > 0:
# if gff_info is None, do not add ',' separator
try:
gff_info += ",%s" % ";".join([f.type] + f.qualifiers['product'])
except TypeError:
gff_info = ";".join([f.type] + f.qualifiers['product'])
# Test if there were any features with a product
if gff_info == None:
gff_info = "N/A"
else:
gff_info = "N/A"
out_dict[rec.id] = gff_info
return out_dict
def gene_to_early_exons(gene_name, num_exons):
#Initialize variables
exons = {}
exonCount = 0
maxExons = 0
#Open annotation file
annotation_file = 'crispr_app/Homo_sapiens.GRCh38.84.gtf'
limit_info = dict(
gff_type = ["exon"])
annotation_handle = open(annotation_file)
#Parse through annotated data, searching for matching gene names & exons
strand = ''
for rec in GFF.parse(annotation_handle, limit_info=limit_info, target_lines=1):
feature = rec.features[0]
qualifiers = feature.qualifiers
#Once matching gene is found, determine the exon regions and chromosome
if str(qualifiers['gene_name']).strip('[').strip(']').strip('\'') == gene_name:
chromosome = rec.id
strand = feature.strand
#Get only first version of gene in annotated data
exonNum = str(qualifiers['exon_number']).strip('[').strip(']').strip('\'')
maxExons = max(maxExons, int(exonNum))
exonCount +=1
if exonCount > maxExons:
break
if exonCount > num_exons:
break
exons[exonNum] = [int(feature.location.start), int(feature.location.end), strand]
annotation_handle.close()
return exons, chromosome
def not_t_full_celegans(self):
"""Test the full C elegans chromosome and GFF files.
This is used to test GFF on large files and is not run as a standard
test. You will need to download the files and adjust the paths
to run this.
"""
# read the sequence information
seq_file = os.path.join(self._full_dir, "c_elegans.WS199.dna.fa")
gff_file = os.path.join(self._full_dir, "c_elegans.WS199.gff3")
seq_handle = open(seq_file)
seq_dict = SeqIO.to_dict(SeqIO.parse(seq_handle, "fasta"))
seq_handle.close()
#with open(gff_file) as gff_handle:
# possible_limits = feature_adder.available_limits(gff_handle)
# pprint.pprint(possible_limits)
rnai_types = [('Orfeome', 'PCR_product'),
('GenePair_STS', 'PCR_product'),
('Promoterome', 'PCR_product')]
gene_types = [('Non_coding_transcript', 'gene'),
('Coding_transcript', 'gene'),
('Coding_transcript', 'mRNA'),
('Coding_transcript', 'CDS')]
limit_info = dict(gff_source_type = rnai_types + gene_types)
for rec in GFF.parse(gff_file, seq_dict, limit_info=limit_info):
pass
def gene_positions(genefile, include_chromosome=True, include_strand=True, coding_only=False, ignore_strange_cases=False):
""" Return a gene_ID:(chromosome, strand, start_pos, end_pos) dictionary based on GFF input file.
The positions are 1-based, end-inclusive.
If include_chromosome and/or include_strand is False, the corresponding values are missing from the output tuples.
If coding_only is True, the start/end positions are the start and end of the first and last exon (i.e. excluding the UTRs).
In that case, if a gene doesn't have an mRNA with exons, or has multiple mRNAs, raise an Exception,
unless ignore_strange_cases is True, then just don't include it in the output.
"""
gene_positions = {}
with open(os.path.expanduser(genefile)) as GENEFILE:
# if coding_only is False, only look at genes, not sub-features
genefile_parsing_limits = {'gff_type': ['gene']} if not coding_only else {}
for chromosome_record in GFF.parse(GENEFILE, limit_info=genefile_parsing_limits):
for gene_record in chromosome_record.features:
# BCBio uses 0-based and end-exclusive positions (first-third base is bases 0,1,2, i.e range 0-3) -
# convert to 1-based end-inclusive (so first-third base is bases 1,2,3, i.e. range 1-3)
if include_chromosome: full_pos_info = (chromosome_record.id,)
else: full_pos_info = ()
if include_strand: full_pos_info += (GFF_strands[gene_record.strand],)
if not coding_only:
full_pos_info += get_feature_start_end(gene_record)
else:
try: start_end = get_gene_start_end_excluding_UTRs(gene_record)
except (NoRNAError, MultipleRNAError):
if ignore_strange_cases: continue
else: raise
full_pos_info += start_end
gene_positions[gene_record.id] = full_pos_info
return gene_positions
def read_gff_transcripts(fobj, fname="", min_exons=1, merge=0):
# Setup logging
logger = logging.getLogger('pita')
if merge > 0:
logger.warning("Merging exons not yet implemented for GFF files!")
#limits = dict(gff_type = ["mRNA", "exon"])
smap = {"1":"+",1:"+","-1":"-",-1:"-", None:"+"}
transcripts = []
for rec in GFF.parse(fobj):
chrom = rec.id
for feature in rec.features:
#logger.debug("feature: {0}", feature)
for gene in _gff_type_iterator(feature, ['mRNA', 'transcript', 'inferred_parent']):
#logger.debug("Adding gene: {0}", gene)
exons = []
#logger.debug("subfeatures: {0}", gene.sub_features)
for exon in [f for f in gene.sub_features if f.type == 'exon']:
#link[gene.id] = link.setdefault(gene.id, 0) + 1
start = int(exon.location.start.position)# - 1
end = int(exon.location.end.position)
strand = smap[exon.strand]
exons.append([chrom, start, end, strand])
logger.debug("%s: %s - %s exons", fname, gene.id, len(exons))
if len(exons) >= min_exons:
transcripts.append([gene.id, fname, exons])
return transcripts
def t_key_whitespace(self):
"""Fix keys with problematic whitespace.
"""
tfile = os.path.join(self._test_dir, "spaces.gff3")
for i, line_info in enumerate(GFF.parse_simple(tfile)):
if i > 2:
assert line_info["quals"]["foo"] == ["bar"]
def prepareSample(filter_matrix, gff_path): random.seed() candidate_list = [] handle = open(gff_path, 'r') gene_count = 0 for record in GFF.parse(handle): for feature in record.features: if feature.type == 'gene': locus_tag = feature.qualifiers['locus_tag'][0] isMatch = False gene_count += 1 for key in filter_matrix: if key == locus_tag: isMatch = True break if isMatch == False: candidate_list.append(locus_tag) countToAdd = round(gene_count / 2) - len(filter_matrix) if countToAdd > 0: for i in range(1, countToAdd): list_len = len(candidate_list) list_id = random.randint(0, list_len - 1) locus_str = candidate_list[ list_id ] filter_matrix[locus_str] = (0, 0) candidate_list.remove( locus_str ) handle.close() return(filter_matrix)
def t_ensembl_nested_features(self):
"""Test nesting of features with GFF2 files using transcript_id.
"""
rec_dict = SeqIO.to_dict(GFF.parse(self._ensembl_file))
assert len(rec_dict["I"].features) == 2
t_feature = rec_dict["I"].features[0]
assert len(t_feature.sub_features) == 32
def t_write_fasta(self):
"""Include FASTA records in GFF output.
"""
seq = Seq("GATCGATCGATCGATCGATC")
rec = SeqRecord(seq, "ID1")
qualifiers = {"source": "prediction", "score": 10.0, "other": ["Some", "annotations"],
"ID": "gene1"}
rec.features = [SeqFeature(FeatureLocation(0, 20), type="gene", strand=1,
qualifiers=qualifiers)]
out_handle = StringIO.StringIO()
GFF.write([rec], out_handle, include_fasta=True)
wrote_info = out_handle.getvalue().split("\n")
fasta_parts = wrote_info[3:]
assert fasta_parts[0] == "##FASTA"
assert fasta_parts[1] == ">ID1 <unknown description>"
assert fasta_parts[2] == str(seq)
def t_fasta_directive(self):
"""Parse FASTA sequence information contained in a GFF3 file.
"""
recs = SeqIO.to_dict(GFF.parse(self._gff_file))
assert len(recs) == 1
test_rec = recs['chr17']
assert str(test_rec.seq) == "GATTACAGATTACA"
def doWork( args ):
panel=Panel(fig_width=900, padding = 25, grid=None, xmin=0)
seq_length = 0
for gff in args.gffs:
seqrecord = GFF.parse(gff).next()
if len(seqrecord) > seq_length:
seq_length = len(seqrecord)
#seqrecord = SeqIO.parse(args.infile, "genbank").next()
cds_track = tracks.BaseTrack(sort_by = 'collapse')
for feature in seqrecord.features:
if feature.type == 'CDS':
#print feature.qualifiers['product']
if feature.qualifiers['product'][0] == 'hypothetical protein':
col = '#BDBDBD'
else:
col = '#2B8CBE'
feat = features.GenericSeqFeature(feature, color_by_cm=False,
fc=col )
cds_track.append(feat)
elif feature.type == 'source':
cds_track.append(features.GenericSeqFeature(feature,
color_by_cm=False, alpha=0.0, fc='1.0', ec='1.0'))
else:
cds_track.append(features.GenericSeqFeature(feature,
color_by_cm=False, fc='0.0', ec='0.0'))
panel.add_track(cds_track)
panel.save(args.outfile, xmin=0,xmax=seq_length)
def main(gff_file, fasta_file = None):
# Use splitext to remove the extension of the original input file
out_file = "%s.gb" % os.path.splitext(gff_file)[0]
# Parser will differ slightly if fasta file is given
if os.stat(gff_file) == 0 or ((fasta_file is not None) and os.stat(fasta_file)):
print "ERROR: Empty file provided or cannot stat files"
exit(64);
elif fasta_file is None:
gff_iter = GFF.parse(gff_file) #Parser/generator object
else:
fasta_input = SeqIO.to_dict(SeqIO.parse(fasta_file, "fasta", generic_dna)) # Process fasta file
gff_iter = GFF.parse(gff_file, fasta_input) # Give fasta file to parser
# One line to call all the checking function and to write in genbank format
SeqIO.write(_check_gff(_fix_ncbi_id(gff_iter)), out_file, "genbank")
def shortrna_regions(mirna_gff, star_csv, seq_file):
"""Return miRNA sequences with corresponding guide and star regions.
"""
seq_index = SeqIO.index(seq_file, "fasta")
mirna_seqs = dict()
with open(star_csv) as in_handle:
for name, guide, star in csv.reader(in_handle):
mirna_seqs[name] = (guide.strip(), star.strip())
for rec in GFF.parse(mirna_gff):
cur_seq = str(seq_index[rec.id].seq)
for f in rec.features:
name = f.qualifiers["ID"][0]
start, end = (f.location.nofuzzy_start, f.location.nofuzzy_end)
yield (rec.id, start, end, name)
#guide, star = mirna_seqs.get(name, ("", ""))
for seq_name, guide, star in [(n, g, s) for n, (g, s) in
mirna_seqs.iteritems() if n.startswith(name)]:
for find_seq, ext in [(guide, "guide"), (star, "star")]:
if find_seq:
if f.strand == -1:
find_seq = str(Seq(find_seq).reverse_complement())
region = cur_seq[start:end]
pos = region.find(find_seq)
if pos > -1:
yield (rec.id, start + pos, start + pos + len(find_seq),
"%s_%s" % (seq_name, ext))
else:
print f.strand, name, ext, pos, find_seq, region
raise NotImplementedError
def load_gff(gff):
"""Parses a single GFF file and returns a chromosome-indexed dict for
that file.
Arguments
---------
gff: str
Filepath to GFF
Returns
-------
dict: A dictionary representation of the GFF entries, indexed by
chromosome ID
"""
annotations = {}
if gff.endswith('.gz'):
import gzip
from io import TextIOWrapper
fp = TextIOWrapper(gzip.open(gff))
else:
fp = open(gff)
for entry in GFF.parse(fp):
if len(entry.features) > 0 and entry.features[0].type == 'chromosome':
annotations[entry.id] = entry
fp.close()
return annotations
def gb2gff(infile, outfile):
"""Translate GenBank file to GFF3 file. TODO: the procedure now does not
handle join correctly
Args:
infile (str): input GenBank file
outfile (str): output GFF3 file
Returns:
Number of records written
"""
gb_handle = open(infile, 'r')
gff_handle = open(outfile, 'w')
res = GFF.write(SeqIO.parse(gb_handle, "gb"), gff_handle)
gff_handle.close()
return (res)
def t_wormbase_nested_features(self):
"""Test nesting of features with GFF2 files using Transcript only.
"""
rec_dict = SeqIO.to_dict(GFF.parse(self._wormbase_file))
assert len(rec_dict) == 3
parent_features = [
f for f in rec_dict["I"].features if f.type == "Transcript"
]
assert len(parent_features) == 1
inferred_features = [
f for f in rec_dict["I"].features if f.type == "inferred_parent"
]
assert len(inferred_features) == 0
tfeature = parent_features[0]
assert tfeature.qualifiers["WormPep"][0] == "WP:CE40797"
assert len(tfeature.sub_features) == 46
def working_stuff():
from BCBio import GFF
gff_type = ["gene", "mRNA", "CDS", "exon"]
source_type = zip(["Coding_transcript"] * len(gff_type), gff_type)
filter_type = dict(gff_source_type=source_type, gff_id="I")
gff_handle = open(
"/fml/ag-raetsch/share/databases/genomes/C_elegans/elegans_WS199/annotation/c_elegans.WS199.gff3"
)
element = [e for e in GFF.parse(gff_handle, limit_info=filter_type)]
return element
def t_extra_comma(self):
"""Correctly handle GFF3 files with extra trailing commas.
"""
tfile = os.path.join(self._test_dir, "mouse_extra_comma.gff3")
in_handle = open(tfile)
for rec in GFF.parse(in_handle):
pass
in_handle.close()
tested = False
for sub_top in rec.features[0].sub_features:
for sub in sub_top.sub_features:
if sub.qualifiers.get("Name",
"") == ["CDS:NC_000083.5:LOC100040603"]:
tested = True
assert len(sub.qualifiers["Parent"]) == 1
assert tested, "Did not find sub-feature to test"
def get_features_from_file(handle: IO) -> Dict[str, List[SeqFeature]]:
""" Generates new SeqFeatures from a GFF file.
Arguments:
handle: a file handle/stream with the GFF contents
Returns:
a dictionary mapping record ID to a list of SeqFeatures for that record
"""
try:
gff_records = list(GFF.parse(handle))
except Exception as err:
raise AntismashInputError("could not parse records from GFF3 file") from err
results = {}
for gff_record in gff_records:
features = []
for feature in gff_record.features:
if feature.type == 'CDS':
new_features = [feature]
else:
new_features = check_sub(feature)
if not new_features:
continue
name = feature.id
locus_tag = feature.qualifiers.get("locus_tag")
for qtype in ["gene", "name", "Name"]:
if qtype in feature.qualifiers:
name_tmp = feature.qualifiers[qtype][0]
# Assume name/Name to be sane if they don't contain a space
if " " in name_tmp:
continue
name = name_tmp
break
for i, new_feature in enumerate(new_features):
variant = name
if len(new_features) > 1:
variant = "{0}_{1}".format(name, i)
new_feature.qualifiers['gene'] = [variant]
if locus_tag is not None:
new_feature.qualifiers["locus_tag"] = locus_tag
features.append(new_feature)
results[gff_record.id] = features
return results
def runbarrnap(genome,outfilePath = '.'):
'''
assumes write permission in current working directory
:param genome: path to genome assumes assembly file name ("_genomic.fna.gz")
:param outfile: target file with rRNA sequences detected in fasta format with header
">asmID:ribosomalSubunit:start-end:dir:acc"
:return: True if successfully run, false if not
'''
if os.path.isfile(genome):
fileName = os.path.split(genome)[1]
asmID = fileName.split('.')[0]
outfileName = asmID + '_rRNA.fasta'
try:
tempGenome = asmID + '.tmp'
with gzip.open(genome) as in_file, open(tempGenome,'w') as out_file:
shutil.copyfileobj(in_file,out_file)
except IOError:
tempGenome = asmID + '.tmp'
shutil.copy(genome, tempGenome)
## unzip
command = ['barrnap','--incseq',tempGenome]
try:
barrnapProc = subprocess.Popen(command,stdout=subprocess.PIPE)
handle = GFF.parse(barrnapProc.stdout)
seqs = []
for seq in handle:
for feat in seq.features:
rnaSeq = feat.extract(seq)
rSU = feat.qualifiers['name'][0]
eval = float(feat.qualifiers['score'][0])
start = feat.location.start
end = feat.location.end
dir = feat.location.strand
if 'note' in feat.qualifiers:
partial = 'partial'
else:
partial = 'full'
rnaSeq.name = "{}:{}:{}:{}-{}:{}:{}".format(asmID,rSU,eval,start,end,dir,partial)
rnaSeq.id = "{}:{}:{}:{}-{}:{}:{}".format(asmID,rSU,eval,start,end,dir,partial)
rnaSeq.description = ''
seqs.append(rnaSeq)
with open(os.path.join(outfilePath,outfileName),'w') as outfile:
SeqIO.write(seqs,outfile,'fasta')
except Exception as e:
print(e)
os.remove(tempGenome)
def repair(fasta, gff3):
recs = {}
# seqids = {}
for record in GFF.parse(gff3):
# seqids[record.id] = ''
recs[record.id] = record
seqs = []
for seq in SeqIO.parse(fasta, "fasta"):
if seq.id not in recs:
continue
current = recs[seq.id]
for num, feat in enumerate(current.features):
if num == 0: # ignore first feature bc that's the full one
continue
cds, sd = get_CDS_and_SD(feat)
cds_start = seq.seq[cds.location.start:cds.location.start + 3]
broken_start = break_start(cds_start)
if (
cds_start != broken_start
): # try to break start sequence while keeping amino acid the same
seq.seq = (seq.seq[0:cds.location.start] + broken_start +
seq.seq[cds.location.start + 3:])
else: # if couldn't change start, must break SD
mod_sd_start = 0
mod_sd_end = 0
if (sd.location.start % 3) + 1 != 1:
mod_sd_start = next_first_frame(sd.location.start, 1)
if (sd.location.end % 3) + 1 != 1:
mod_sd_end = next_first_frame(sd.location.end, -1)
sd_seq = seq.seq[sd.location.start -
mod_sd_start:sd.location.end - mod_sd_end]
broken_sd = break_sd(sd_seq)
if sd_seq != broken_sd:
seq.seq = (seq.seq[0:(sd.location.start - mod_sd_start)] +
broken_sd +
seq.seq[(sd.location.end - mod_sd_end):])
seqs.append(seq)
SeqIO.write(seqs, sys.stdout, "fasta")
def main(seqFilepath, gffFilepath, outFilepath):
# load fasta
seqRec_lst = []
seqName_lst = []
for seqRec in SeqIO.parse(seqFilepath, "fasta"):
seqRec_lst.append(seqRec)
seqName_lst.append(seqRec.id)
print("LOADED {} seqs from {}".format(len(seqRec_lst), seqFilepath))
# load gff and distribute CDS
cds_lstlst = [[] for _ in range(len(seqName_lst))]
with open(gffFilepath) as f:
for rec in GFF.parse(f, target_lines=1):
assert len(rec.features) == 1
if rec.features[0].type == "CDS":
try:
idx = seqName_lst.index(rec.id)
cds_lstlst[idx].append(rec)
except ValueError:
pass # corresponding sequence does not exists in seqRec_lst
for idx, seqName in enumerate(seqName_lst):
print("\tLOADED {0} CDSs in {1}".format(len(cds_lstlst[idx]), seqName))
thres_lst = list(range(50, 1000 + 1, 50))
columns = ["+1", "+2", "+3", "-1", "-2", "-3"]
out_mat = np.zeros((len(thres_lst), len(columns))).astype(int)
for seqRec, cds_lst, seqName in zip(seqRec_lst, cds_lstlst, seqName_lst):
orf_df = get_orf_df(seqRec)
for i, thres in enumerate(thres_lst):
filtered_df = orf_df[orf_df["length"] >= thres]
pos_lst = get_pos_lst(cds_lst, filtered_df)
overlap_dctdct = get_overlap_dctdct(pos_lst)
for _, dct in overlap_dctdct.items():
out_mat[i, columns.index(dct["relLane"])] += (dct["oend"] -
dct["ostart"])
print("\tDONE with {}".format(seqName))
out_df = pd.DataFrame(out_mat, columns=columns)
out_df["thres"] = thres_lst
out_df = out_df[["thres"] + columns]
out_df.to_csv(outFilepath, index=False)
print("OUTPUT to {}".format(outFilepath))
def __init__(self, fasta, gtf):
self.fasta = fasta
self.gtf = gtf
sys.stderr.write("Reading FASTA file...\n")
with flexi_open(fasta, 'rU') as handle:
chromosomes = SeqIO.to_dict(SeqIO.parse(handle, 'fasta'))
self.intervaldict = dict()
sys.stderr.write("Reading GTF file (this will take some time)...\n")
limit_info = dict(gff_type=('CDS', ))
with flexi_open(gtf, 'r') as handle:
for rec in GFF.parse(handle,
limit_info=limit_info,
base_dict=chromosomes):
# Fix strand info.
for feature in rec.features: # Each top-level CDS
if hasattr(feature, 'strand'):
if feature.strand is None:
# An unfortunate effect of the GFF parser.
# Check subfeatures, take the first defined strand.
# Note: may be bad assumption in weird species (ciliates??).
for subfeat in get_subfeatures(feature):
if subfeat.strand is not None:
feature.strand = subfeat.strand
break
# Now, chromosomes['X'] is a SeqRecord with features each of
# which has a feature.extract method which can be used to
# access the underlying DNA sequence. CDS features are nested
# automatically. See also feature.qualifiers for a list of IDs
# (gene name, ID etc) associated with it.
chromosomes[rec.id] = rec
# We need to create some interval trees to identify
# affected CDS features etc.
self._index_record_in_intervaldict(rec)
self.chromosomes = chromosomes
self._precompute_chrlens()
sys.stderr.write("Object initialisation complete.\n")
def record_with_extracted_annotations_generator(
gff_file, white_list_of_annotation_types):
for record in GFF.parse(open(gff_file)):
#print("Extracting annotations from %s" % record.id)
new_record = deepcopy(record)
new_record.features = []
#print record.features
for feature in record.features:
#print ("%s\t%s" % (record.id, feature.id))
if (feature.id
in annotation_ids) and (feature.type
in white_list_of_annotation_types):
new_record.features.append(feature)
if len(new_record.features) > 0:
yield new_record
def do_import(self):
in_file = self.__gff_fasta_fn
in_handle = open(in_file)
# In DEBUG=True mode, Django keeps list of queries and blows up memory
# usage when doing a big import. The following line disables this
# logging.
connection.use_debug_cursor = False
for rec in GFF.parse(in_handle):
f = GFFFragmentImporter(rec).do_import()
self.__genome.genome_fragment_set.create(fragment=f,
inherited=False)
# Be nice and turn debug cursor back on
connection.use_debug_cursor = True
in_handle.close()
def __get_features(child, interpro=False):
child_features = {}
for rec in GFF.parse(child):
for feature in rec.features:
parent_feature_id = rec.id
if interpro:
if feature.type == 'polypeptide':
continue
if '_' in parent_feature_id:
parent_feature_id = parent_feature_id[parent_feature_id.
index('_') + 1:]
try:
child_features[parent_feature_id].append(feature)
except KeyError:
child_features[parent_feature_id] = [feature]
return child_features
def FindInitSites():
GFFgen = GFF.parse(
'/users/buskirk/documents/profiling/GFF/MG1655/coli3.gff')
chrom = GFFgen.next()
f_genome = chrom.seq
r_genome = chrom.seq.reverse_complement()
ORFlist = []
ORFlist.append([
'start', 'stop', 'retapa', 'onc112', 'codon1st', 'codonlast',
'peptide', 'strand'
])
# for the plus strand
f1 = 'retapa' #retapamulin data from Shura Mankin
pathi = "/users/buskirk/documents/profiling/projects/sORFs/wigfiles/"
density_filestring1 = pathi + f1
counts_f1 = readwig(density_filestring1 + "_plus")
f2 = 'onc' #Onc112 data
pathi = "/users/buskirk/documents/profiling/projects/sORFs/wigfiles/"
density_filestring2 = pathi + f2
counts_f2 = readwig(density_filestring2 + "_plus")
gp_plus = geneplot(chrom, 1) # 1 is plus
plusORFs = INIT_scan(f_genome, gp_plus, counts_f1, counts_f2, 'plus')
ORFlist.extend(plusORFs)
# for the minus strand
counts_f1 = readwig(density_filestring1 + "_minus")
counts_f1.reverse()
counts_f2 = readwig(density_filestring2 + "_minus")
counts_f2.reverse()
gp_minus = geneplot(chrom, -1) # -1 is minus
gp_minus.reverse()
minusORFs = INIT_scan(r_genome, gp_minus, counts_f1, counts_f2, 'minus')
ORFlist.extend(minusORFs)
writelisttoexcel(
ORFlist,
'/users/buskirk/documents/profiling/projects/sORFs/init_sites')
def get_features_from_file(
seq_record,
handle,
limit_to_seq_id: Union[bool, Dict[str, List[str]]] = False
) -> List[SeqFeature]:
""" Generates new SeqFeatures from a Record and a GFF file.
Arguments:
seq_record: the record that features belong to
limit_to_seq_id: False or a dictionary of GFF.parse options
Returns:
a list of SeqFeatures parsed from the GFF file
"""
features = []
for record in GFF.parse(handle, limit_info=limit_to_seq_id):
for feature in record.features:
if feature.type == 'CDS':
new_features = [feature]
else:
new_features = check_sub(feature, seq_record)
if not new_features:
continue
name = feature.id
locus_tag = feature.qualifiers.get("locus_tag")
for qtype in ["gene", "name", "Name"]:
if qtype in feature.qualifiers:
name_tmp = feature.qualifiers[qtype][0]
# Assume name/Name to be sane if they don't contain a space
if " " in name_tmp:
continue
name = name_tmp
break
for i, new_feature in enumerate(new_features):
variant = name
if len(new_features) > 1:
variant = "{0}_{1}".format(name, i)
new_feature.qualifiers['gene'] = [variant]
if locus_tag is not None:
new_feature.qualifiers["locus_tag"] = locus_tag
features.append(new_feature)
return features
def __get_features(child, interpro=False):
child_features = {}
for rec in GFF.parse(child):
log.info("Parsing %s", rec.id)
for feature in rec.features:
parent_feature_id = rec.id
if interpro:
if feature.type == "polypeptide":
continue
if "_" in parent_feature_id:
parent_feature_id = parent_feature_id[parent_feature_id.
index("_") + 1:]
try:
child_features[parent_feature_id].append(feature)
except KeyError:
child_features[parent_feature_id] = [feature]
return child_features
def examine(gff_file, fasta_file):
gff_handle = open(gff_file)
fasta_handle = open(fasta_file)
fasta_dict = SeqIO.to_dict(SeqIO.parse(fasta_handle, "fasta"))
for rec in GFF.parse(gff_handle):
#print rec.id
for one_feature in rec.features:
#print one_feature.qualifiers.keys()
locus = one_feature.qualifiers["Name"][0]
product = one_feature.qualifiers["product"][0]
seq = one_feature.extract(fasta_dict[rec.id].seq)
out = ">%s|%s|%s\n%s" % (rec.id, locus, product, seq)
print out
gff_handle.close()
fasta_handle.close()
def features(self):
"""list: Get the features from the feature file, metadata file, or in memory"""
if self.feature_file:
log.debug('{}: reading features from feature file {}'.format(self.id, self.feature_path))
with open(self.feature_path) as handle:
feats = list(GFF.parse(handle))
if len(feats) > 1:
log.warning('Too many sequences in GFF')
else:
return feats[0].features
elif self.metadata_file:
log.debug('{}: reading features from metadata file {}'.format(self.id, self.metadata_path))
tmp_sr = SeqIO.read(self.metadata_path, 'uniprot-xml')
return tmp_sr.features
else:
return self._features
def generator(gff_file, fasta_dict, exon_fd):
with open(gff_file, "r") as gff_fd:
for record in GFF.parse(gff_fd, target_lines=100000):
for feature in record.features:
#print (feature.type)
if feature.type == "transcript":
print(feature)
print(feature.sub_features)
exon_fd.write(feature.id + "\n")
exon_fd.write(str(feature.location))
exon_fd.write("\n")
exon_fd.write(str(feature.sub_features) + "\n")
if feature.type == "gene":
feature_record = feature.extract(fasta_dict[record.id])
feature_record.id = feature.qualifiers["gene_id"][0]
feature_record.description = ""
yield feature_record
def tableify(gff3, fasta):
names = {}
for fasta_rec in SeqIO.parse(fasta, 'fasta'):
names[fasta_rec.id] = []
for gff_rec in GFF.parse(gff3):
names[gff_rec.id].append(str(len(gff_rec.features) -
1)) # number of internal starts
starts = []
for feat in gff_rec.features:
feat_start = (feat.location.start - 9) / 3 + 1
if feat_start is not 1: # start codon position of each internal start
starts.append(str(feat_start))
names[gff_rec.id].append(starts)
for n in sorted(names):
if len(names[n]):
print '\t'.join([n, names[n][0], ', '.join(names[n][1])])
def t_basic_attributes(self):
"""Parse out basic attributes of GFF2 from Ensembl GTF.
"""
limit_info = dict(
gff_source_type = [('snoRNA', 'exon')]
)
rec_dict = SeqIO.to_dict(GFF.parse(self._ensembl_file,
limit_info=limit_info))
work_rec = rec_dict['I']
assert len(work_rec.features) == 1
test_feature = work_rec.features[0]
qual_keys = test_feature.qualifiers.keys()
qual_keys.sort()
assert qual_keys == ['Parent', 'exon_number', 'gene_id', 'gene_name',
'source', 'transcript_id', 'transcript_name']
assert test_feature.qualifiers['source'] == ['snoRNA']
assert test_feature.qualifiers['transcript_name'] == ['NR_001477.2']
assert test_feature.qualifiers['exon_number'] == ['1']
def ParseRecord(self, cn):
org = self._wa.organisms.findOrganismByCn(cn)
self._wa.annotations.setSequence(org['commonName'], org['id'])
data = io.StringIO(
self._wa.io.write(
exportType='GFF3',
seqType='genomic',
exportAllSequences=False,
exportGff3Fasta=True,
output="text",
exportFormat="text",
sequences=cn,
))
data.seek(0)
for record in GFF.parse(data):
yield WebApolloSeqRecord(record, self._wa)
def basic_parsing(gffFile):
""" GFF3 parse, extract information """
with open(gffFile) as in_handle:
for rec in GFF.parse(in_handle):
# iterate features
for feature in rec.features:
# iterate sub features
sub_features_temp = []
for sub_feature in feature.sub_features:
sub_features_temp.append(sub_feature)
yield SeqFeature(
type=feature.type,
location=feature.location,
strand=feature.strand,
qualifiers=feature.qualifiers,
sub_features=sub_features_temp,
)
def require_shinefind(gff3, fasta):
sd_finder = NaiveSDCaller()
# Load up sequence(s) for GFF3 data
seq_dict = SeqIO.to_dict(SeqIO.parse(fasta, "fasta"))
# Parse GFF3 records
for record in GFF.parse(gff3, base_dict=seq_dict):
# Reopen
genes = list(
feature_lambda(record.features,
feature_test_type, {"type": "gene"},
subfeatures=True))
good_genes = []
for gene in genes:
cdss = sorted(
list(
feature_lambda(
gene.sub_features,
feature_test_type,
{"type": "CDS"},
subfeatures=False,
)),
key=lambda x: x.location.start,
)
if len(cdss) == 0:
continue
cds = cdss[0]
sds, start, end, seq = sd_finder.testFeatureUpstream(cds,
record,
sd_min=5,
sd_max=15)
if len(sds) >= 1:
sd_features = sd_finder.to_features(sds,
gene.location.strand,
start,
end,
feature_id=gene.id)
gene.sub_features.append(sd_features[0])
good_genes.append(gene)
record.features = good_genes
yield record
def FindrRNA(path, tempFile):
in_handle = open(path)
for record in GFF.parse(in_handle, limit_info=limit_info):
if len(record.features) != 0:
for i in range(len(record.features)):
if '23S ribosomal RNA' in record.features[i].qualifiers["product"][0]:
#print(record.features[i])
#print(record.id)
#print(record.features[i].location.strand)
#print(record.features[i].location.start)
#print(record.features[i].location.end)
#print(type(record.features[i].location))
#feature_seq = record.seq[record.features[i].location.start:record.features[i].location.end].reverse_complement()
#print(feature_seq)
'''
seq_23S = SeqRecord(record.features[i].location.extract(record.seq),\
id=record.features[i].id,\
description=str(record.id +'-'+ record.features[i].qualifiers["product"][0]))
'''
seq_23S = SeqRecord(record.features[i].location.extract(record.seq),\
id=tempFile[:-4],\
description=str(record.id +'-'+record.features[i].id+'-'+ record.features[i].qualifiers["product"][0]+'-'+str(len(record.seq))))
#print(seq_23S)
#SeqIO.write(seq_23S, path[:-4]+"_23S_rRNA.fasta", "fasta")
SeqIO.write(seq_23S, "/home/junyuchen/Lab/16S-Prediction/"+tempFile[:-4]+"_23S_rRNA.fasta", "fasta")
elif '16S ribosomal RNA' in record.features[i].qualifiers["product"][0]:
#print(record.features[i])
#print(record.id)
seq_16S = SeqRecord(record.features[i].location.extract(record.seq),\
id=tempFile[:-4],\
description=str(record.id +'-'+record.features[i].id+'-'+ record.features[i].qualifiers["product"][0]+'-'+str(len(record.seq))))
#SeqIO.write(seq_16S, path[:-4]+"_16S_rRNA.fasta", "fasta")
SeqIO.write(seq_16S, "/home/junyuchen/Lab/16S-Prediction/"+tempFile[:-4]+"_16S_rRNA.fasta", "fasta")
#print(seq_16S)
#print
#print(record.features)
in_handle.close()
def main(fasta, gff3):
seq_dict = SeqIO.to_dict(SeqIO.parse(fasta, "fasta"))
codon_usage = {}
for rec in GFF.parse(gff3, base_dict=seq_dict):
for feat in feature_lambda(rec.features,
feature_test_type, {"type": "CDS"},
subfeatures=True):
seq = str(feat.extract(rec).seq)[0:3]
try:
codon_usage[seq] += 1
except KeyError:
codon_usage[seq] = 1
# TODO: print all actg combinations? Or just ones that are there
print "# Codon\tCount"
for key in sorted(codon_usage):
print "\t".join((key, str(codon_usage[key])))
def get_SNP_regions(my_genes, in_file="gencode.v24.annotation.gff3"):
in_handle = open(in_file)
regions = []
for index, rec in enumerate(GFF.parse(in_handle, target_lines=130)):
for x in rec.features:
if x.type == 'gene':
for y in my_genes:
if y == x.qualifiers['gene_name'][0]:
regions.append({
'chr': rec.id,
'start': x.location.start,
'end': x.location.end
})
in_handle.close()
return regions
def load_annotations(target_gff):
"""Loads genome annotations from specified GFF(s)."""
# Get chromosomes/contigs from GFF file
chromosomes = {}
# Load existing gene annotations
annotations_fp = open(target_gff)
for entry in GFF.parse(annotations_fp):
# For TriTrypDB 29 and above, there are no longer chromosome entries
# in the GFF files
# if len(entry.features) > 0 and entry.features[0].type in ['chromosome', 'contig']:
if len(entry.features) > 0:
chromosomes[entry.id] = entry
# clean up
annotations_fp.close()
return chromosomes
def record_with_extracted_transcripts_generator(gff_file,
transcript_ids):
for record in GFF.parse(open(gff_file)):
new_record = deepcopy(record)
new_record.features = []
for feature in record.features:
if (feature.type == "mRNA" or feature.type == "transcript") and (feature.id in transcript_ids):
new_record.features.append(feature)
elif feature.type == "gene":
new_feature = deepcopy(feature)
new_feature.sub_features = []
for subfeature in feature.sub_features:
if (subfeature.type == "mRNA" or subfeature.type == "transcript") and (subfeature.id in transcript_ids):
new_feature.sub_features.append(subfeature)
if len(new_feature.sub_features) > 0:
new_record.features.append(new_feature)
if len(new_record.features) > 0:
yield new_record
def run(sequence, options):
handle = open(options.gff3)
# If there's only one sequence in both, read all, otherwise, read only appropriate part of GFF3.
if options.single_entries:
limit_info = False
else:
limit_info = dict(gff_id=[sequence.id])
for record in GFF.parse(handle, limit_info=limit_info):
for feature in record.features:
if feature.type == 'CDS':
new_features = [feature]
else:
new_features = check_sub(feature, sequence)
if not new_features:
continue
name = feature.id
if len(name) > 40:
raise ValueError(
"Feature ID too long, < 40 characters required: %s" % name)
locus_tag = None
if "locus_tag" in feature.qualifiers:
locus_tag = feature.qualifiers["locus_tag"]
for qtype in ["gene", "name", "Name"]:
if qtype in feature.qualifiers:
name_tmp = feature.qualifiers[qtype][0]
# Assume name/Name to be sane if they don't contain a space
if " " in name_tmp:
continue
name = name_tmp
break
for i, n in enumerate(new_features):
variant = name
if len(new_features) > 1:
variant = "{0}_{1}".format(name, i)
n.qualifiers['gene'] = [variant]
if locus_tag is not None:
n.qualifiers["locus_tag"] = locus_tag
sequence.features.append(n)
def require_shinefind(gff3, fasta):
sd_finder = NaiveSDCaller()
# Load up sequence(s) for GFF3 data
seq_dict = SeqIO.to_dict(SeqIO.parse(fasta, "fasta"))
# Parse GFF3 records
for record in GFF.parse(gff3, base_dict=seq_dict):
# Reopen
genes = list(
feature_lambda(record.features,
feature_test_type, {'type': 'gene'},
subfeatures=True))
good_genes = []
for gene in genes:
cdss = list(
feature_lambda(gene.sub_features,
feature_test_type, {'type': 'CDS'},
subfeatures=False))
if len(cdss) == 0:
continue
# Someday this will bite me in the arse.
cds = cdss[0]
sds, start, end, seq = sd_finder.testFeatureUpstream(cds,
record,
sd_min=5,
sd_max=15)
if len(sds) >= 1:
# TODO
# Double plus yuck
sd_features = sd_finder.to_features(sds,
gene.location.strand,
start,
end,
feature_id=gene.id)
gene.sub_features.append(sd_features[0])
good_genes.append(gene)
# Yuck!
record.features = good_genes
yield record
