def format_read_as_fna(read, qual=False):
"""Format a single read from a binary SFF file as a FASTA record.
If qual is True, output the qual values instead of the bases.
"""
# TODO: Move to PyCogent
out = StringIO()
out.write('>%s' % read['Name'])
# Roche uses 1-based indexing, where the right index is inclusive.
# To transform to 0-based indices, where the right index is not
# inclusive, we subtract 1 from the left index, but leave the
# right index intact.
start_idx = read['clip_qual_left'] - 1
end_idx = read['clip_qual_right']
# A surprising result is produced if the number of cycles are
# adjusted such that no bases remain past clip_qual_left. In the
# clipping routine, the Roche software sets clip_qual_left to be
# equal to the number of bases. Using our indexing scheme, the
# resulting sequence is of length 1 after clipping (one would
# expect a length of 0). We would fix this issue, if the effect
# were not present in the output from Roche's sffinfo program. We
# retain this arguably incorrect behavior to be consistent with
# the reference implementation.
out.write(' length=%d' % (end_idx - start_idx))
timestamp, _, region, location = decode_accession(read['Name'])
out.write(' xy=%04d_%04d' % location)
out.write(' region=%d' % region)
out.write(' run=R_%d_%02d_%02d_%02d_%02d_%02d_' % timestamp)
out.write('\n')
if qual:
scores = read['quality_scores'][start_idx:end_idx]
out.write(' '.join(['%d' % s for s in scores]))
else:
bases = read['Bases'][start_idx:end_idx]
out.write(bases)
out.write('\n')
return out.getvalue()
def format_read_as_fna(read, qual=False):
"""Format a single read from a binary SFF file as a FASTA record.
If qual is True, output the qual values instead of the bases.
"""
# TODO: Move to PyCogent
out = StringIO()
out.write('>%s' % read['Name'])
# Roche uses 1-based indexing, where the right index is inclusive.
# To transform to 0-based indices, where the right index is not
# inclusive, we subtract 1 from the left index, but leave the
# right index intact.
start_idx = read['clip_qual_left'] - 1
end_idx = read['clip_qual_right']
# A surprising result is produced if the number of cycles are
# adjusted such that no bases remain past clip_qual_left. In the
# clipping routine, the Roche software sets clip_qual_left to be
# equal to the number of bases. Using our indexing scheme, the
# resulting sequence is of length 1 after clipping (one would
# expect a length of 0). We would fix this issue, if the effect
# were not present in the output from Roche's sffinfo program. We
# retain this arguably incorrect behavior to be consistent with
# the reference implementation.
out.write(' length=%d' % (end_idx - start_idx))
timestamp, _, region, location = decode_accession(read['Name'])
out.write(' xy=%04d_%04d' % location)
out.write(' region=%d' % region)
out.write(' run=R_%d_%02d_%02d_%02d_%02d_%02d_' % timestamp)
out.write('\n')
if qual:
scores = read['quality_scores'][start_idx:end_idx]
out.write(' '.join(['%d' % s for s in scores]))
else:
bases = read['Bases'][start_idx:end_idx]
out.write(bases)
out.write('\n')
return out.getvalue()
def test_decode_accession(self):
self.assertEqual(
decode_accession('GA202I001ER3QL'),
((2010, 1, 22, 13, 28, 56), '0', 1, (1843, 859)))
def test_decode_accession(self):
self.assertEqual(decode_accession('GA202I001ER3QL'),
((2010, 1, 22, 13, 28, 56), '0', 1, (1843, 859)))
