def parse_gff3(gff3_file):
assemblies = dict()
genes = dict()
for line in open(gff3_file):
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
## initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly(id=mol_id)
current_assembly = assemblies[mol_id]
rfmin = int(cols[3]) - 1
rfmax = int(cols[4])
rstrand = None
feat_id = column_9_value(cols[8], 'ID')
#print("Processing feature: ({0})".format(feat_id))
if cols[6] == '-':
strand = -1
elif cols[6] == '+':
strand = 1
else:
strand = 0
if cols[2] == 'gene':
gene = biothings.Gene(id=feat_id)
gene.locate_on(assembly=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
genes[feat_id] = gene
elif cols[2] == 'mRNA':
mRNA = biothings.mRNA(id=feat_id)
mRNA.locate_on(assembly=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_id = column_9_value(cols[8], 'Parent')
genes[parent_id].add_mRNA(mRNA)
return (assemblies, genes)
def main():
parser = argparse.ArgumentParser( description='Convert PASA GFF file to canonical gene models')
## output file to be written
parser.add_argument('-i', '--input', type=str, required=True, help='Path to a GFF file created by PASA' )
parser.add_argument('-o', '--output', type=str, required=True, help='Path to an output file to be created' )
parser.add_argument('-s', '--source', type=str, required=False, default='PASA', help='Value to use for the 2nd (source) column' )
args = parser.parse_args()
assemblies = dict()
current_assembly = None
gene = None
mRNA = None
gene_fmin = None
gene_fmax = None
gene_strand = None
## Used for tracking the exon count for each gene (for ID purposes)
exon_count_by_mRNA = dict()
fout = open(args.output, mode='wt', encoding='utf-8')
fout.write("##gff-version 3\n")
for line in open(args.input):
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
feat_type = cols[2]
feat_id = biocodegff.column_9_value(cols[8], 'ID')
# we expect all columns to be cDNA_match
if feat_type != 'cDNA_match':
raise Exception("ERROR: expected all columns to be of type 'cDNA_match' but found a {0}".format(feat_type))
## initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly( id=mol_id )
if gene is None or feat_id != gene.id:
if gene is not None:
# finish the previous one first
mRNA.locate_on( target=current_assembly, fmin=gene_fmin, fmax=gene_fmax, strand=gene_strand )
gene.locate_on( target=current_assembly, fmin=gene_fmin, fmax=gene_fmax, strand=gene_strand )
gene.add_mRNA(mRNA)
current_assembly.add_gene( gene )
gene.print_as(fh=fout, source=args.source, format='gff3')
# now start a new one
gene = biothings.Gene( id=feat_id )
mRNA = biothings.mRNA( id="{0}.mRNA".format(feat_id), parent=gene )
exon_count_by_mRNA[mRNA.id] = 0
gene_fmin = int(cols[3]) - 1
gene_fmax = int(cols[4])
gene_strand = cols[6]
current_assembly = assemblies[mol_id]
# each row is a new CDS/exon for the current mRNA
CDS = biothings.CDS( id="{0}.CDS".format(feat_id), parent=mRNA.id )
# FIX THIS PHASE
CDS.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6], phase='.' )
mRNA.add_CDS(CDS)
exon_count_by_mRNA[mRNA.id] += 1
exon_id = "{0}.exon{1}".format(mRNA.id, exon_count_by_mRNA[mRNA.id])
exon = biothings.Exon( id=exon_id, parent=mRNA.id )
exon.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6] )
mRNA.add_exon(exon)
if int(cols[3]) - 1 < gene_fmin:
gene_fmin = int(cols[3]) - 1
if int(cols[4]) > gene_fmax:
gene_fmax = int(cols[4])
# don't orphan the last one
if gene is not None:
# finish the previous one first
mRNA.locate_on( target=current_assembly, fmin=gene_fmin, fmax=gene_fmax, strand=gene_strand )
gene.locate_on( target=current_assembly, fmin=gene_fmin, fmax=gene_fmax, strand=gene_strand )
gene.add_mRNA(mRNA)
current_assembly.add_gene( gene )
gene.print_as(fh=fout, source=args.source, format='gff3')
def main():
parser = argparse.ArgumentParser(
description='Converts CEGMA GFF output to spec-legal GFF3')
# output file to be written
parser.add_argument('-i',
'--input_file',
type=str,
required=True,
help='Path to an input file to parse')
parser.add_argument('-o',
'--output_file',
type=str,
required=True,
help='Path to an output file to be created')
args = parser.parse_args()
fout = open(args.output_file, 'w')
fout.write("##gff-version 3\n")
assemblies = dict()
current_assembly = None
current_gene = None
current_mRNA = None
current_gene_fmin = None
current_gene_fmax = None
current_gene_strand = None
next_id_nums = {'gene': 1, 'mRNA': 1, 'CDS': 1, 'exon': 1}
exon_column_types = ['First', 'Internal', 'Terminal', 'Single']
for line in open(args.input_file, 'r'):
if line.startswith('#'):
continue
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
feat_type = cols[2]
feat_fmin = int(cols[3]) - 1
feat_fmax = int(cols[4])
if feat_type == 'Single' or feat_type == 'First':
# If there's an existing gene already, print it out
if current_gene is not None:
current_gene.locate_on(target=current_assembly,
fmin=current_gene_fmin,
fmax=current_gene_fmax,
strand=current_gene_strand)
current_mRNA.locate_on(target=current_assembly,
fmin=current_gene_fmin,
fmax=current_gene_fmax,
strand=current_gene_strand)
#current_gene.print_as(format='text')
current_gene.print_as(fh=fout, source='cegma', format='gff3')
# initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly(id=mol_id)
current_assembly = assemblies[mol_id]
feat_id = "cegma.gene.{0}".format(next_id_nums['gene'])
next_id_nums['gene'] += 1
gene = biothings.Gene(id=feat_id)
current_gene = gene
current_gene_strand = cols[6]
current_gene_fmin = feat_fmin
current_gene_fmax = feat_fmax
mRNA_id = "cegma.mRNA.{0}".format(next_id_nums['mRNA'])
next_id_nums['mRNA'] += 1
mRNA = biothings.mRNA(id=mRNA_id, parent=gene)
gene.add_mRNA(mRNA)
current_mRNA = mRNA
# CEGMA versions < 2.5 had two rows for each exon. We don't need to process both of them, so
# we skip the Exon one because its phase information is incorrect.
if feat_type in exon_column_types:
CDS_id = "cegma.CDS.{0}".format(next_id_nums['CDS'])
next_id_nums['CDS'] += 1
CDS = biothings.CDS(id=CDS_id, parent=current_mRNA)
CDS.locate_on(target=current_assembly,
fmin=feat_fmin,
fmax=feat_fmax,
strand=cols[6],
phase=cols[7])
current_mRNA.add_CDS(CDS)
exon_id = "cegma.exon.{0}".format(next_id_nums['exon'])
next_id_nums['exon'] += 1
exon = biothings.Exon(id=exon_id, parent=current_mRNA)
exon.locate_on(target=current_assembly,
fmin=feat_fmin,
fmax=feat_fmax,
strand=cols[6])
mRNA.add_exon(exon)
if feat_fmin < current_gene_fmin:
current_gene_fmin = feat_fmin
if feat_fmax > current_gene_fmax:
current_gene_fmax = feat_fmax
# don't forget the last gene
if current_gene is not None:
current_gene.locate_on(target=current_assembly,
fmin=current_gene_fmin,
fmax=current_gene_fmax,
strand=current_gene_strand)
current_mRNA.locate_on(target=current_assembly,
fmin=current_gene_fmin,
fmax=current_gene_fmax,
strand=current_gene_strand)
current_gene.print_as(fh=fout, source='cegma', format='gff3')
def get_gff3_features(gff3_file, assemblies=None):
'''
Parses the passed GFF3 file and returns two dicts, loaded with biocode.biothings objects:
1. The first dict are the Assembly objects, keyed on assembly ID. Each Assembly has all of the
children populated, so you can fully recover gene, RNA, exon and CDS features iterating on
the assembly.
2. The second dist is a flat structure of all the descendent feature objects of the Assemblies
keyed by the feature IDs.
See the documentation for each feature type in biocode.biothings for more info
'''
if assemblies is None:
assemblies = dict()
features = dict()
# these are related to parsing any embedded FASTA
in_fasta_section = False
is_assembly_fasta = False
current_fasta_id = None
for line in open(gff3_file):
#print("INFO: processing line: {0}".format(line))
if in_fasta_section == True:
m = re.search('>(\S+)\s*(.*)', line)
if m:
current_fasta_id = m.group(1)
if current_fasta_id in assemblies:
is_assembly_fasta = True
else:
is_assembly_fasta = False
else:
if is_assembly_fasta == True:
# must be a sequence line for an assembly
# python 2.6+ makes string concatenation amortized O(n)
# http://stackoverflow.com/a/4435752/1368079
assemblies[current_fasta_id].residues += str(line.rstrip())
assemblies[current_fasta_id].length = len(
assemblies[current_fasta_id].residues)
continue
elif line.startswith("##FASTA"):
# all data to the end of the file must be FASTA
in_fasta_section = True
continue
# ignore all other comments
if line.startswith('#'):
continue
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
# initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly(id=mol_id, residues='')
current_assembly = assemblies[mol_id]
rfmin = int(cols[3]) - 1
rfmax = int(cols[4])
rstrand = None
atts = column_9_dict(cols[8])
feat_id = atts.get('ID')
parent_id = atts.get('Parent')
parent_feat = None
# sanity check
if rfmin > rfmax:
raise Exception(
"ERROR: Coordinates in GFF for feature id {0} appear to be reversed and violate GFF3 specification: {1} > {2}"
.format(feat_id, cols[3], cols[4]))
if 'locus_tag' in atts:
locus_tag = atts['locus_tag']
else:
locus_tag = None
# shared features is not yet supported
if isinstance(parent_id, list):
raise Exception(
"This line contains a shared feature with multiple parents. This isn't yet supported:\n{0}"
.format(line))
if parent_id is not None:
if parent_id in features:
parent_feat = features[parent_id]
else:
raise Exception(
"Error in GFF3: Parent {0} referenced by a child feature before it was defined"
.format(parent_id))
if cols[6] == '-':
rstrand = -1
elif cols[6] == '+':
rstrand = 1
else:
rstrand = 0
if cols[2] == 'gene':
gene = biothings.Gene(id=feat_id, locus_tag=locus_tag)
gene.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
features[feat_id] = gene
current_assembly.add_gene(gene)
elif cols[2] == 'mRNA':
mRNA = biothings.mRNA(id=feat_id,
parent=parent_feat,
locus_tag=locus_tag)
mRNA.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_mRNA(mRNA)
features[feat_id] = mRNA
elif cols[2] == 'rRNA':
rRNA = biothings.rRNA(id=feat_id,
parent=parent_feat,
locus_tag=locus_tag)
rRNA.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_rRNA(rRNA)
features[feat_id] = rRNA
elif cols[2] == 'tRNA':
tRNA = biothings.tRNA(id=feat_id,
parent=parent_feat,
locus_tag=locus_tag)
tRNA.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_tRNA(tRNA)
features[feat_id] = tRNA
elif cols[2] == 'exon':
exon = biothings.Exon(id=feat_id, parent=parent_feat)
exon.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_exon(exon)
features[feat_id] = exon
elif cols[2] == 'CDS':
CDS = biothings.CDS(id=feat_id, parent=parent_feat)
CDS.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_CDS(CDS)
features[feat_id] = CDS
elif cols[2] == 'polypeptide':
polypeptide = biothings.Polypeptide(id=feat_id, parent=parent_feat)
polypeptide.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_polypeptide(polypeptide)
polypeptide.annotation = parse_annotation_from_column_9(cols[8])
features[feat_id] = polypeptide
elif cols[2] == 'five_prime_UTR':
utr = biothings.FivePrimeUTR(id=feat_id, parent=parent_feat)
utr.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_five_prime_UTR(utr)
features[feat_id] = utr
elif cols[2] == 'three_prime_UTR':
utr = biothings.ThreePrimeUTR(id=feat_id, parent=parent_feat)
utr.locate_on(target=current_assembly,
fmin=rfmin,
fmax=rfmax,
strand=rstrand)
parent_feat.add_three_prime_UTR(utr)
features[feat_id] = utr
else:
sys.stderr.write("Skipping feature {0} with type {1}\n".format(
feat_id, cols[2]))
continue
features[feat_id].length = rfmax - rfmin
return (assemblies, features)
def main():
parser = argparse.ArgumentParser(
description='Metagenemark GFF -> GFF3 conversion script')
## output file to be written
parser.add_argument('-i',
'--input',
type=str,
required=True,
help='Path to a GFF file created by Metagenemark')
parser.add_argument('-o',
'--output',
type=str,
required=True,
help='Path to an output file to be created')
parser.add_argument('-p',
'--prefix',
type=str,
required=True,
help='Prefix to use in ID generation')
parser.add_argument('-pf',
'--protein_fasta',
type=str,
required=False,
help='Optional protein FASTA to be written')
args = parser.parse_args()
assemblies = dict()
current_assembly = None
# key like 2 = SRS014890.polypeptide.2
polypeptide_lookup = dict()
writing_protein = False
gene = None
mRNAs = dict()
current_sequence = None
current_gene_comment_lines = list()
fout = open(args.output, mode='wt', encoding='utf-8')
fout.write("##gff-version 3\n")
if args.protein_fasta is not None:
protein_out = open(args.protein_fasta, mode='wt', encoding='utf-8')
for line in open(args.input):
if line.startswith("#"):
if line.startswith("##FASTA"):
current_gene_comment_lines.append("#{0}".format(line))
elif line.startswith("##end-Protein"):
writing_protein = False
current_gene_comment_lines.append(line)
# since we're already doing our own header, don't duplicate the old one
elif line.startswith("##gff-version"):
continue
else:
if line.startswith("##Protein "):
m = re.match("##Protein (\d+)", line)
if m:
writing_protein = True
protein_out.write(">{0}\n".format(
polypeptide_lookup[m.group(1)]))
else:
raise Exception(
"ERROR: Expected line to match: ##Protein N")
elif writing_protein == True:
protein_out.write(line[2:])
current_gene_comment_lines.append(line)
else:
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
mol_id_m = re.match('^(\S+) ', mol_id)
if mol_id_m:
print("MATCH!")
mol_id = mol_id_m.group(1)
feat_type = cols[2]
## we expect only gene types here
if feat_type not in ['gene', 'CDS']:
raise Exception(
"ERROR: expected only 'gene' or 'CDS' feature types as input (depending on metagenemark version)."
)
m_gene = re.match('gene_id[ =](\d+)', cols[8])
if m_gene:
gene_num = m_gene.group(1)
else:
raise Exception(
"ERROR: expected 9th column to have gene ids like: gene_id 5"
)
## initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly(id=mol_id)
current_assembly = assemblies[mol_id]
gene = biothings.Gene(
id="{0}.gene.{1}".format(args.prefix, gene_num))
gene.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
mRNA = biothings.mRNA(id="{0}.mRNA.{1}".format(
args.prefix, gene_num),
parent=gene.id)
mRNA.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
gene.add_mRNA(mRNA)
CDS = biothings.CDS(id="{0}.CDS.{1}".format(args.prefix, gene_num),
parent=mRNA.id)
CDS.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6],
phase=int(cols[7]))
mRNA.add_CDS(CDS)
exon = biothings.Exon(id="{0}.exon.{1}".format(
args.prefix, gene_num),
parent=mRNA.id)
exon.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
mRNA.add_exon(exon)
polypeptide_id = "{0}.polypeptide.{1}".format(
args.prefix, gene_num)
polypeptide = biothings.Polypeptide(id=polypeptide_id,
parent=mRNA.id)
polypeptide.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
mRNA.add_polypeptide(polypeptide)
polypeptide_lookup[gene_num] = polypeptide_id
gene.print_as(fh=fout, source='GeneMark.hmm', format='gff3')
fout.write("".join(current_gene_comment_lines))
current_gene_comment_lines = list()
def main():
parser = argparse.ArgumentParser( description='A GTF -> GFF3 conversion script for Cufflinks output')
## output file to be written
parser.add_argument('-i', '--input_file', type=str, required=True, help='Path to an input GTF file' )
parser.add_argument('-o', '--output_file', type=str, required=False, help='Path to an output GFF file to be created' )
parser.add_argument('-e', '--export_mode', type=str, required=False, default='model', help='Export mode for results (model or cDNA_match)' )
args = parser.parse_args()
if args.export_mode not in ['model', 'cDNA_match']:
raise Exception("ERROR: the only valid values for --export_mode are 'model' or 'cDNA_match'")
## output will either be a file or STDOUT
ofh = sys.stdout
if args.output_file is not None:
ofh = open(args.output_file, 'wt')
ofh.write("##gff-version 3\n")
assemblies = dict()
current_assembly = None
current_gene = None
current_RNA = None
current_match = None
rna_count_by_gene = defaultdict(int)
exon_count_by_RNA = defaultdict(int)
# each gb_record is a SeqRecord object
for line in open(args.input_file, "r"):
cols = line.split("\t")
if len(cols) != 9:
print("SKIPPING: {0}".format(line))
continue
mol_id = cols[0]
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly( id=mol_id )
current_assembly = assemblies[mol_id]
ftype = cols[2]
fmin = int(cols[3]) - 1
fmax = int(cols[4])
strand = cols[6]
col9 = cols[8]
# this makes it look like GFF column 9 so I can use biocodeutils.column_9_value(str, key)
col9 = col9.replace(' "', '="')
gene_id = biocodegff.column_9_value(col9, 'gene_id').replace('"', '')
transcript_id = biocodegff.column_9_value(col9, 'transcript_id').replace('"', '')
if ftype == 'transcript':
if args.export_mode == 'model':
if current_gene is not None and current_gene.id != gene_id:
gene.print_as(fh=ofh, source='Cufflinks', format='gff3')
if current_gene is None or current_gene.id != gene_id:
gene = biothings.Gene( id=gene_id )
gene.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
current_gene = gene
mRNA = biothings.mRNA( id=transcript_id, parent=current_gene )
mRNA.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
gene.add_mRNA(mRNA)
current_RNA = mRNA
exon_count_by_RNA[transcript_id] = 0
current_CDS_phase = 0
elif args.export_mode == 'cDNA_match':
if current_match is not None and current_match.id != transcript_id:
match.print_as( fh=ofh, source='Cufflinks', format='gff3' )
match = biothings.Match( id=transcript_id, subclass='cDNA_match', length=fmax - fmin )
match.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
current_match = match
elif ftype == 'exon':
exon_number = biocodegff.column_9_value(col9, 'exon_number').replace('"', '')
if args.export_mode == 'model':
exon_count_by_RNA[transcript_id] += 1
cds_id = "{0}.CDS.{1}".format( current_RNA.id, exon_count_by_RNA[current_RNA.id] )
CDS = biothings.CDS( id=cds_id, parent=current_RNA )
CDS.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand, phase=current_CDS_phase )
current_RNA.add_CDS(CDS)
# calculate the starting phase for the next CDS feature (in case there is one)
current_CDS_phase = 3 - (((fmax - fmin) - current_CDS_phase) % 3)
if current_CDS_phase == 3:
current_CDS_phase = 0
exon_id = "{0}.exon.{1}".format( current_RNA.id, exon_count_by_RNA[current_RNA.id] )
exon = biothings.Exon( id=exon_id, parent=current_RNA )
exon.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
current_RNA.add_exon(exon)
elif args.export_mode == 'cDNA_match':
mp_id = "{0}.match_part.{1}".format(transcript_id, exon_number)
mp = biothings.MatchPart( id=mp_id, parent=current_match, length=fmax - fmin )
mp.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
current_match.add_part(mp)
# don't forget to do the last gene, if there were any
if args.export_mode == 'model':
if current_gene is not None:
gene.print_as(fh=ofh, source='GenBank', format='gff3')
elif args.export_mode == 'cDNA_match':
if current_match is not None:
match.print_as( fh=ofh, source='Cufflinks', format='gff3' )
def main():
parser = argparse.ArgumentParser( description='Convert GenBank flat files to GFF3 format')
## output file to be written
parser.add_argument('-i', '--input_file', type=str, required=True, help='Path to an input GBK file' )
parser.add_argument('-o', '--output_file', type=str, required=False, help='Path to an output GFF file to be created' )
parser.add_argument('--with_fasta', dest='fasta', action='store_true', help='Include the FASTA section with genomic sequence at end of file. (default)' )
parser.add_argument('--no_fasta', dest='fasta', action='store_false' )
parser.set_defaults(fasta=True)
args = parser.parse_args()
## output will either be a file or STDOUT
ofh = sys.stdout
if args.output_file is not None:
ofh = open(args.output_file, 'wt')
ofh.write("##gff-version 3\n")
assemblies = dict()
current_assembly = None
current_gene = None
current_RNA = None
rna_count_by_gene = defaultdict(int)
exon_count_by_RNA = defaultdict(int)
seqs_pending_writes = False
features_skipped_count = 0
# each gb_record is a SeqRecord object
for gb_record in SeqIO.parse(open(args.input_file, "r"), "genbank"):
mol_id = gb_record.name
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly( id=mol_id )
if len(str(gb_record.seq)) > 0:
seqs_pending_writes = True
assemblies[mol_id].residues = str(gb_record.seq)
assemblies[mol_id].length = len(str(gb_record.seq))
current_assembly = assemblies[mol_id]
# each feat is a SeqFeature object
for feat in gb_record.features:
#print(feat)
fmin = int(feat.location.start)
fmax = int(feat.location.end)
if feat.location.strand == 1:
strand = '+'
elif feat.location.strand == -1:
strand = '-'
else:
raise Exception("ERROR: unstranded feature encountered: {0}".format(feat))
#print("{0} located at {1}-{2} strand:{3}".format( locus_tag, fmin, fmax, strand ) )
if feat.type == 'source':
continue
if feat.type == 'gene':
# print the previous gene (if there is one)
if current_gene is not None:
gene.print_as(fh=ofh, source='GenBank', format='gff3')
locus_tag = feat.qualifiers['locus_tag'][0]
gene = biothings.Gene( id=locus_tag )
gene.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
current_gene = gene
current_RNA = None
elif feat.type == 'mRNA':
locus_tag = feat.qualifiers['locus_tag'][0]
rna_count_by_gene[locus_tag] += 1
feat_id = "{0}.mRNA.{1}".format( locus_tag, rna_count_by_gene[locus_tag] )
mRNA = biothings.mRNA( id=feat_id, parent=current_gene )
mRNA.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
gene.add_mRNA(mRNA)
current_RNA = mRNA
if feat_id in exon_count_by_RNA:
raise Exception( "ERROR: two different RNAs found with same ID: {0}".format(feat_id) )
else:
exon_count_by_RNA[feat_id] = 0
elif feat.type == 'tRNA':
locus_tag = feat.qualifiers['locus_tag'][0]
rna_count_by_gene[locus_tag] += 1
feat_id = "{0}.tRNA.{1}".format( locus_tag, rna_count_by_gene[locus_tag] )
tRNA = biothings.tRNA( id=feat_id, parent=current_gene )
tRNA.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
gene.add_tRNA(tRNA)
current_RNA = tRNA
if feat_id in exon_count_by_RNA:
raise Exception( "ERROR: two different RNAs found with same ID: {0}".format(feat_id) )
else:
exon_count_by_RNA[feat_id] = 0
elif feat.type == 'rRNA':
locus_tag = feat.qualifiers['locus_tag'][0]
rna_count_by_gene[locus_tag] += 1
feat_id = "{0}.rRNA.{1}".format( locus_tag, rna_count_by_gene[locus_tag] )
rRNA = biothings.rRNA( id=feat_id, parent=current_gene )
rRNA.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
gene.add_rRNA(rRNA)
current_RNA = rRNA
if feat_id in exon_count_by_RNA:
raise Exception( "ERROR: two different RNAs found with same ID: {0}".format(feat_id) )
else:
exon_count_by_RNA[feat_id] = 0
elif feat.type == 'CDS':
locus_tag = feat.qualifiers['locus_tag'][0]
# If processing a prokaryotic GBK, we'll encounter CDS before mRNA, so we have to
# manually make one
if current_RNA is None:
feat_id = "{0}.mRNA.{1}".format( locus_tag, rna_count_by_gene[locus_tag] )
mRNA = biothings.mRNA( id=feat_id, parent=current_gene )
mRNA.locate_on( target=current_assembly, fmin=fmin, fmax=fmax, strand=strand )
gene.add_mRNA(mRNA)
current_RNA = mRNA
exon_count_by_RNA[current_RNA.id] += 1
cds_id = "{0}.CDS.{1}".format( current_RNA.id, exon_count_by_RNA[current_RNA.id] )
current_CDS_phase = 0
for loc in feat.location.parts:
subfmin = int(loc.start)
subfmax = int(loc.end)
CDS = biothings.CDS( id=cds_id, parent=current_RNA )
CDS.locate_on( target=current_assembly, fmin=subfmin, fmax=subfmax, strand=strand, phase=current_CDS_phase )
current_RNA.add_CDS(CDS)
# calculate the starting phase for the next CDS feature (in case there is one)
# 0 + 6 = 0 TTGCAT
# 0 + 7 = 2 TTGCATG
# 1 + 6 = 1 TTGCAT
# 2 + 7 = 1 TTGCATG
# general: 3 - ((length - previous phase) % 3)
current_CDS_phase = 3 - (((subfmax - subfmin) - current_CDS_phase) % 3)
if current_CDS_phase == 3:
current_CDS_phase = 0
exon_id = "{0}.exon.{1}".format( current_RNA.id, exon_count_by_RNA[current_RNA.id] )
exon = biothings.Exon( id=exon_id, parent=current_RNA )
exon.locate_on( target=current_assembly, fmin=subfmin, fmax=subfmax, strand=strand )
current_RNA.add_exon(exon)
exon_count_by_RNA[current_RNA.id] += 1
else:
print("WARNING: The following feature was skipped:\n{0}".format(feat))
features_skipped_count += 1
# don't forget to do the last gene, if there were any
if current_gene is not None:
gene.print_as(fh=ofh, source='GenBank', format='gff3')
if args.fasta is True:
if seqs_pending_writes is True:
ofh.write("##FASTA\n")
for assembly_id in assemblies:
ofh.write(">{0}\n".format(assembly_id))
ofh.write("{0}\n".format(biocodeutils.wrapped_fasta(assemblies[assembly_id].residues)))
if features_skipped_count > 0:
print("Warning: {0} unsupported feature types were skipped".format(features_skipped_count))
def main():
parser = argparse.ArgumentParser( description='Metagenemark GFF -> GFF3 conversion script')
## output file to be written
parser.add_argument('-i', '--input', type=str, required=True, help='Path to a GFF file created by Metagenemark' )
parser.add_argument('-o', '--output', type=str, required=True, help='Path to an output file to be created' )
parser.add_argument('-p', '--prefix', type=str, required=True, help='Prefix to use in ID generation')
args = parser.parse_args()
assemblies = dict()
current_assembly = None
gene = None
mRNAs = dict()
current_sequence = None
current_gene_comment_lines = list()
fout = open(args.output, mode='wt', encoding='utf-8')
fout.write("##gff-version 3\n")
for line in open(args.input):
if line.startswith("#"):
current_gene_comment_lines.append(line)
else:
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
mol_id_m = re.match('^(\S+) ', mol_id)
if mol_id_m:
print("MATCH!")
mol_id = mol_id_m.group(1)
feat_type = cols[2]
## we expect only gene types here
if feat_type not in ['gene', 'CDS']:
raise Exception("ERROR: expected only 'gene' or 'CDS' feature types as input (depending on metagenemark version).")
m_gene = re.match('gene_id[ =](\d+)', cols[8])
if m_gene:
gene_num = m_gene.group(1)
else:
raise Exception("ERROR: expected 9th column to have gene ids like: gene_id 5")
## initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly( id=mol_id )
current_assembly = assemblies[mol_id]
gene = biothings.Gene( id="{0}.gene.{1}".format(args.prefix, gene_num) )
gene.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6] )
mRNA = biothings.mRNA( id="{0}.mRNA.{1}".format(args.prefix, gene_num), parent=gene.id )
mRNA.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6] )
gene.add_mRNA(mRNA)
CDS = biothings.CDS( id="{0}.CDS.{1}".format(args.prefix, gene_num), parent=mRNA.id )
CDS.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6], phase=int(cols[7]) )
mRNA.add_CDS(CDS)
exon = biothings.Exon( id="{0}.exon.{1}".format(args.prefix, gene_num), parent=mRNA.id )
exon.locate_on( target=current_assembly, fmin=int(cols[3]) - 1, fmax=int(cols[4]), strand=cols[6] )
mRNA.add_exon(exon)
gene.print_as(fh=fout, source='GeneMark.hmm', format='gff3')
fout.write( "".join(current_gene_comment_lines) )
current_gene_comment_lines = list()
def main():
parser = argparse.ArgumentParser(
description=
'Convert native (GTF) or GFF output from Augustus into GFF3 format')
## output file to be written
parser.add_argument('-i',
'--input',
type=str,
required=True,
help='Path to a GFF file created by Augustus')
parser.add_argument('-o',
'--output',
type=str,
required=True,
help='Path to an output file to be created')
args = parser.parse_args()
assemblies = dict()
current_assembly = None
gene = None
mRNAs = dict()
in_sequence = False
current_sequence = None
current_gene_comment_lines = list()
## Used for tracking the exon count for each gene (for ID purposes)
exon_count_by_mRNA = dict()
fout = open(args.output, mode='wt', encoding='utf-8')
fout.write("##gff-version 3\n")
for line in open(args.input):
if line.startswith("#"):
current_gene_comment_lines.append(line)
if line.startswith("# end gene "):
## purge the comments, then write the gene
fout.write("".join(current_gene_comment_lines))
gene.print_as(fh=fout, source='AUGUSTUS', format='gff3')
gene = None
mRNAs = dict()
in_sequence = False
current_sequence = None
current_gene_comment_lines = list()
elif line.startswith("# protein sequence = ["):
pass
elif in_sequence is True:
# build 'current_sequence'
pass
else:
cols = line.split("\t")
if len(cols) != 9:
continue
mol_id = cols[0]
feat_type = cols[2]
if feat_type not in ['gene', 'transcript', 'CDS']:
continue
## The output format is GTF by default and (mostly) GFF if the --gff option is used.
# If GTF is detected, let's start by transforming the 9th column into GFF so the
# libraries can use it
# g1 -> ID=g1
# g1.t1 -> ID=g1.t1;Parent=g1
# transcript_id "g1.t1"; gene_id "g1"; -> ID=g1.t1.cds;Parent=g1.t1
m_gene = re.match('(g\d+)', cols[8])
m_transcript = re.match('((g\d+).t\d+)', cols[8])
m_CDS = re.match('transcript_id "(g\d+.t\d+)"; gene_id "g\d+";',
cols[8])
# the input can be in GTF or GFF. We need to reformat the 9th column for the GTF entries
if not cols[8].startswith('ID') and not cols[8].startswith(
'Parent'):
if feat_type == 'gene':
if m_gene:
cols[8] = "ID={0}".format(m_gene.group(1))
else:
raise Exception(
"ERROR: GTF detected but gene row has bad 9th column format: {0}"
.format(cols[8]))
elif feat_type == 'transcript':
if m_transcript:
cols[8] = "ID={0};Parent={1}".format(
m_transcript.group(1), m_transcript.group(2))
else:
raise Exception(
"ERROR: GTF detected but transcript row has bad 9th column format: {0}"
.format(cols[8]))
elif feat_type == 'CDS':
if m_CDS:
cols[8] = "ID={0}.cds;Parent={0}".format(
m_CDS.group(1))
else:
raise Exception(
"ERROR: GTF detected but CDS row has bad 9th column format: {0}"
.format(cols[8]))
feat_id = biocodegff.column_9_value(cols[8], 'ID')
## initialize this assembly if we haven't seen it yet
if mol_id not in assemblies:
assemblies[mol_id] = biothings.Assembly(id=mol_id)
current_assembly = assemblies[mol_id]
if feat_type == "gene":
gene = biothings.Gene(id=feat_id)
gene.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
elif feat_type == "transcript":
mRNA = biothings.mRNA(id=feat_id, parent=gene)
mRNA.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
gene.add_mRNA(mRNA)
mRNAs[mRNA.id] = mRNA
if feat_id in exon_count_by_mRNA:
raise Exception(
"ERROR: two different mRNAs found with same ID: {0}".
format(feat_id))
else:
exon_count_by_mRNA[feat_id] = 0
elif feat_type == "CDS":
parent_id = biocodegff.column_9_value(cols[8], 'Parent')
## sanity check that we've seen this parent
if parent_id not in mRNAs:
raise Exception(
"ERROR: Found CDS column with parent ({0}) mRNA not yet in the file"
.format(parent_id))
CDS = biothings.CDS(id=feat_id, parent=mRNAs[parent_id])
CDS.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6],
phase=int(cols[7]))
mRNA.add_CDS(CDS)
## exons weren't explicitly defined in the input file, so we need to derive new IDs for them
exon_count_by_mRNA[parent_id] += 1
exon_id = "{0}.exon{1}".format(parent_id,
exon_count_by_mRNA[parent_id])
exon = biothings.Exon(id=exon_id, parent=mRNAs[parent_id])
exon.locate_on(target=current_assembly,
fmin=int(cols[3]) - 1,
fmax=int(cols[4]),
strand=cols[6])
mRNA.add_exon(exon)