def build_parser(parser):
parser.add_argument('seqs', help="""Named sequences""")
parser.add_argument('seq_info', help="""Sequence info file""")
parser.add_argument('-t', '--taxonomy',
help="""Taxonomy as taxtable; optional
if a grouping term is available in seq_info""")
parser.add_argument('--seq-info-out',
help='subset of original seq_info')
parser.add_argument('--derep-map-out',
help=('mapping of input sequences to dereplicated '
'representatives. `group` column corresponds to '
'the field identified by --group-on.'))
parser.add_argument('--seqs-out',
default=sys.stdout,
type=util.file_opener('w'))
parser.add_argument('-g', '--group-on', default='species',
help='Field in seq_info on which to group sequences')
parser.add_argument('--id', default=1.0,
type=float, help="""Clustering identity between 0 and 1
[default: %(default).3f]""")
parser.add_argument('-i', '--include', type=util.file_opener('r'),
help=('Optional file containing list '
'of group labels to include'))
parser.add_argument('--threads',
help=('Number of threads to use for clustering each '
'group [default is one thread per '
'available CPU core]'))
def build_parser(p):
p.add_argument('fasta_file',
help="""sequence file""",
type=file_opener('r'))
p.add_argument('seqinfo_file',
help="""Sequence metadata""",
type=file_opener('r'))
p.add_argument('--named-seqs',
default='named.seqs.fasta',
help='[default %(default)s]')
p.add_argument('--named-info',
default='named.seq_info.csv',
help='[default %(default)s]')
p.add_argument('--unnamed-seqs',
default='unnamed.seqs.fasta',
help='[default %(default)s]')
p.add_argument('--unnamed-info',
default='unnamed.seq_info.csv',
help='[default %(default)s]')
flt = p.add_argument_group('Filtering options')
flt.add_argument(
'-a',
'--prop-ambig-cutoff',
default=0.01,
type=float,
help="""Maximum proportion of characters in sequence which may be
ambiguous [default: %(default).2f]""")
flt.add_argument('-l',
'--min-length',
type=int,
help="""Minimum sequence
length [default: %(default)d]""",
default=1200)
def build_parser(p):
p.add_argument('infile', type=util.file_opener('r'),
help="""Input file, gzipped""")
p.add_argument('database', help="""Path to taxonomy database""")
p.add_argument('fasta_out', type=util.file_opener('w'),
help="""Path to write sequences in FASTA format.
Specify '.gz' or '.bz2' extension to compress.""")
p.add_argument('output', metavar='tax_out', type=argparse.FileType('w'),
help="""Output path to write taxonomic
information in CSV format""")
p.add_argument('--no-header', action='store_false', dest='header',
default=True, help="""Don't write a header""")
def build_parser(p):
p.add_argument('fasta_file', help="""sequence file""", type=file_opener('r'))
p.add_argument('seqinfo_file', help="""Sequence metadata""", type=file_opener('r'))
p.add_argument('named_base')
p.add_argument('unnamed_base')
flt = p.add_argument_group('Filtering options')
flt.add_argument('-a', '--prop-ambig-cutoff', default=0.01, type=float,
help="""Maximum proportion of characters in sequence which may be
ambiguous [default: %(default).2f]""")
flt.add_argument('-l', '--min-length', type=int, help="""Minimum sequence
length [default: %(default)d]""", default=1200)
def build_parser(p):
p.add_argument('infile',
type=util.file_opener('r'),
help="""Input file, gzipped""")
p.add_argument('database', help="""Path to taxonomy database""")
p.add_argument('fasta_out',
type=util.file_opener('w'),
help="""Path to write sequences in FASTA format.
Specify '.gz' or '.bz2' extension to compress.""")
p.add_argument('output',
metavar='tax_out',
type=argparse.FileType('w'),
help="""Output path to write taxonomic
information in CSV format""")
p.add_argument('--no-header',
action='store_false',
dest='header',
default=True,
help="""Don't write a header""")
def action(args):
dtype = {'gi': str, 'tax_id': str, 'species': str}
seq_info = pandas.read_csv(args.seq_info, dtype=dtype, index_col='seqname')
log.info('reading sequences')
with util.file_opener()(args.sequences) as sequences_in:
seqhashes = dict()
for record in util.Counter(SeqIO.parse(sequences_in, 'fasta')):
seq = str(record.seq).replace('\n', '').upper()
seqhashes[record.name] = hashlib.sha1(seq).hexdigest()
seqhash = pandas.Series(data=seqhashes, name='seqhash')
seqhash.index.name = 'seqname'
seq_info = seq_info.join(seqhash)
group_by = ['seqhash']
if args.group_by:
group_by.extend(args.group_by.split(','))
def choose_rep(df):
if args.prefer_columns:
df = df.sort_values(by=args.prefer_columns.split(','))
rep = df.tail(1)
rep['weight'] = len(df)
return rep
log.info('choosing seq_info representatives')
seq_info = seq_info.groupby(
by=group_by, group_keys=False).apply(choose_rep)
seq_info = seq_info.drop('seqhash', axis=1)
log.info('writing seqinfo')
seq_info.to_csv(args.out_info, quoting=csv.QUOTE_NONNUMERIC)
log.info('writing dedup file')
with util.file_opener()(args.sequences) as sequences_in, \
util.file_opener('w')(args.out) as sequences_out:
for record in util.Counter(SeqIO.parse(sequences_in, 'fasta')):
if record.name in seq_info.index:
fasta_out = '>{}\n{}\n'.format(record.name, str(record.seq))
sequences_out.write(fasta_out)
def action(args):
seq_info = pandas.read_csv(args.seq_info, dtype=str, index_col='seqname')
log.info('reading sequences')
with util.file_opener()(args.sequences) as sequences_in:
seqhashes = dict()
for record in util.Counter(SeqIO.parse(sequences_in, 'fasta')):
seq = str(record.seq).replace('\n', '').upper()
seqhashes[record.name] = hashlib.sha1(seq).hexdigest()
seqhash = pandas.Series(data=seqhashes, name='seqhash')
seqhash.index.name = 'seqname'
seq_info = seq_info.join(seqhash)
group_by = ['seqhash']
if args.group_by:
group_by.extend(args.group_by.split(','))
def choose_rep(df):
if args.prefer_columns:
df = df.sort_values(by=args.prefer_columns.split(','))
rep = df.tail(1)
rep['weight'] = len(df)
return rep
log.info('choosing seq_info representatives')
seq_info = seq_info.groupby(
by=group_by, group_keys=False).apply(choose_rep)
seq_info = seq_info.drop('seqhash', axis=1)
log.info('writing seqinfo')
seq_info.to_csv(args.out_info, quoting=csv.QUOTE_NONNUMERIC)
log.info('writing dedup file')
with util.file_opener()(args.sequences) as sequences_in, \
util.file_opener('w')(args.out) as sequences_out:
for record in util.Counter(SeqIO.parse(sequences_in, 'fasta')):
if record.name in seq_info.index:
fasta_out = '>{}\n{}\n'.format(record.name, str(record.seq))
sequences_out.write(fasta_out)
def build_parser(p):
# inputs
p.add_argument(
'fasta',
metavar='FASTA',
help="""sequence file""",
type=util.file_opener('r'))
p.add_argument(
'seqinfo',
metavar='CSV',
help="""Sequence metadata""")
p.add_argument(
'--references',
metavar='CSV',
help='csv file with columns: version,pubmed_id')
# outputs
info_outs = p.add_argument_group('outputs for seq_info\'s')
info_outs.add_argument(
'--named-info',
metavar='CSV',
type=util.file_opener('w'),
help='taxid_classified column is True')
info_outs.add_argument(
'--unnamed-info',
metavar='CSV',
type=util.file_opener('w'),
help='taxid_classified column is False')
info_outs.add_argument(
'--type-info',
type=util.file_opener('w'),
metavar='CSV',
help='rows where is_type column is True')
info_outs.add_argument(
'--published-info',
type=util.file_opener('w'),
metavar='CSV',
help="""requires references.csv. Return seq_info with pubmed_ids""")
info_outs.add_argument(
'--references-info',
type=util.file_opener('w'),
metavar='CSV',
help=('requires references.csv. '
'Return columns [version, accession, pubmed_id]'))
seq_outs = p.add_argument_group('outputs for sequences')
seq_outs.add_argument(
'--named-seqs',
type=util.file_opener('w'),
metavar='FASTA',
help='where taxid_classified column is True')
seq_outs.add_argument(
'--unnamed-seqs',
type=util.file_opener('w'),
metavar='FASTA',
help='where taxid_classified column is False')
seq_outs.add_argument(
'--type-seqs',
type=util.file_opener('w'),
metavar='FASTA',
help='where is_type column is True')
seq_outs.add_argument(
'--published-seqs',
type=util.file_opener('w'),
metavar='FASTA',
help="""requires references.csv. Return sequences with pubmed_ids""")
# filtering switches
flt = p.add_argument_group('filtering options')
flt.add_argument('-a', '--prop-ambig-cutoff',
type=float,
help=('Maximum proportion of characters in '
'sequence which may be ambiguous'))
flt.add_argument('-l', '--min-length',
type=int,
help='Minimum sequence length')
