def find_protein_diff(read, ref, verbose=False, start_offset=3, end_trail=3):
# quality control
if read is None:
return None, None
ends = findEnds(read, ref, start_offset)
if not endMatch(read, ref, ends):
return None, None
newread = read
newref = ref
# scan reference triplet by triplet
# move letters when encountering an indel
prot_errors = []
prot_short = []
i = ends.get('aligned')
ref_index = int((ends.get('aligned') - start_offset) /
3) + 1 # reference amino acid index
max_i = len(ref) - end_trail
while i <= ends.get('end'):
if i > max_i:
break
if newread is None:
break
ref_codon = newref[i:i + 3]
read_codon = newread[i:i + 3]
if '-' in read_codon: # found a deletion
# Check if this is the last acid, and it's incomplete, ignore it.
if re.search('[ATGC]', str(newread[i + 3:])) is None:
break
if '-' in ref_codon: # something very broken
prot_errors.append((ref_index, 'f'))
prot_short.append('f')
break
elif read_codon == '---': # single codon deletion
if ref_index > 0:
prot_errors += [(ref_index, 'd')]
prot_short.append(str(ref_index) + 'Δ')
i += 3
ref_index += 1
else: # check it's not a frame shift
l = indel_len(newread, i)
if l % 3 != 0:
prot_errors.append((ref_index, 'f'))
prot_short.append('f')
break
# realign gap and repeat loop at same position to compare the codons
gap = findGap(newread[i - 1:])
gap = (gap[0] + i - 1, gap[1] + i - 1)
newread = gapAlign(newread, gap, start_offset)
continue
elif '-' in ref_codon: # found an insertion
l = indel_len(newref, i)
if l % 3 != 0:
prot_errors.append((ref_index, 'f'))
prot_short.append('f')
break
gap = findGap(newref[i - 1:])
if gap[0] == 1: # insertion after codon
insertion = newread[gap[0] + i - 1:gap[1] + i - 1]
if '-' in insertion:
prot_errors.append((ref_index, 'f'))
prot_short.append('f')
break
if ref_index > 0:
prot_errors.append(
(ref_index - 1, 'i', str(translate(insertion))
)) # position before + insertion
stop, inslist = format_insertion(ref_index - 1, insertion)
prot_short += inslist
if stop:
break
i += l
ref_index += 1
else: # realign gap and repeat loop at same position to compare the codons
gap = (gap[0] + i - 1, gap[1] + i - 1)
newref = gapAlign(newref, gap, start_offset)
continue
elif translate(read_codon) != translate(
ref_codon): # must be a substitution
if ref_index > 0:
prot_errors.append(
(ref_index, 's', str(translate(read_codon))))
prot_short.append(
str(
translate(ref_codon) + str(ref_index) +
str(translate(read_codon))))
if str(translate(read_codon)) == '*':
break
i += 3
ref_index += 1
else:
i += 3
ref_index += 1
if verbose:
print(prot_errors)
if prot_short == []:
short = 'wt'
else:
short = '/'.join(prot_short)
return tuple(prot_errors), short
def find_dna_hgvs(read,
ref,
refname,
verbose=False,
start_offset=3,
end_trail=3):
"""@ read, ref: MutableSeq objects
:return errors - tuple (position, expected triplet, actual triplet, ) / none if broken read
The assumption is that the reference includes an offset of 3 nt either side of the gene of interest, such that the
starting triplet is reported as 'amino acid 0'. If the offset is different, it needs to be set explicitly.
end_trail specifies the number of nt after end of gene and is ignored
"""
if read is None:
if verbose:
print('no read provided')
return
# No gap realignment at this point
prefix = str(refname) + ':c.'
# quality control that there are no mutations at ends of reads
ends = findEnds(read, ref, start_offset)
if not endMatch(read, ref, ends):
if verbose:
print('ends do not match')
return
# scan read & reference letter by letter, counting position in reference
# reads have been trimmed so that reference starts @ 3 (0,1,2 is the extra triplet)
# in the general case, reference starts @ offset in 0-count
# This is equal to the number of nt before ATG
# ref_index denotes HGVS DNA position labeling, i is used for accessing sequence
dna_errors = []
ref_index = ends.get(
'start'
) - start_offset + 1 # if the read starts at 3, this becomes nt 1 (1-based as is HGVS)
i = ends.get('start')
max_i = len(ref) - end_trail
while i < ends.get('end'):
if i > max_i: # the trailing nt are ignored when reading mutations
break
# check for differences
if read[i] == ref[i]:
ref_index += 1
i += 1
elif read[i] == '-':
# start of a deletion, format depends on length
l = indel_len(read, i)
if ref_index > 0:
if l == 1: # format is POSdel
dna_errors.append(str(ref_index) + 'del')
else:
# format is FIRST_LASTdel
dna_errors.append(
str(ref_index) + '_' + str(ref_index + l - 1) + 'del')
i += l
ref_index += l
elif ref[i] == '-':
# start of an insertion, format is FLANK_FLANKinsSEQ
l = indel_len(ref, i)
if ref_index > 0:
dna_errors.append(
str(ref_index - 1) + '_' + str(ref_index) + 'ins' +
str(read[i:i + l]))
i += l
else:
# substitution: need to include ref. sequence in format 8A>G
if ref_index > 0:
dna_errors.append(
str(ref_index) + str(ref[i]) + '>' + str(read[i]))
i += 1
ref_index += 1
# format the result including name of sequence
if len(dna_errors) == 1:
dna_hgvs = prefix + dna_errors[0]
else:
dna_hgvs = prefix + '[' + ';'.join(dna_errors) + ']'
return dna_hgvs
def find_dna_diff(read, ref, verbose=False, start_offset=3, end_trail=3):
"""
@ read, ref: MutableSeq objects
:return errors - tuple (position, expected triplet, actual triplet, ) / none if broken read
The default assumption is that the reference includes 3 nt either side of the gene of interest. The starting triplet is
reported as 'amino acid 0'. The number of such ignored triplets is set by start-offset.
As for HGVS, the starting offset and number of trailing nt are variable
Letter by letter report mutations in NGS read, all counts 1- based in result (code in 0-count).
- substitution: 78C = nt 78 in reference is changed to C
- deletions: 78d6 = 6 nt deleted starting with 78: 1-77, d6, 84-end
- insertion: 78iATC = after nt 78 inserted seq. ATC
"""
if read is None:
if verbose:
print('no read provided')
return
# No gap realignment at this point
# quality control that there are no mutations at ends of reads
ends = findEnds(read, ref, start_offset)
if not endMatch(read, ref, ends):
if verbose:
print('ends do not match')
return
# scan read & reference letter by letter, counting position in reference
# reads have been trimmed so that reference starts @ offset=3 by default (0,1,2 is the extra triplet)
dna_errors = []
ref_index = ends.get('start') - start_offset + 1
i = ends.get('start')
max_i = len(ref) - end_trail
while i < ends.get('end'):
if i > max_i:
break
# check for differences
if read[i] == ref[i]:
ref_index += 1
i += 1
elif read[i] == '-':
# start of a deletion
l = indel_len(read, i)
# now we know the length of a deletion, check for frameshifts
if l % 3 == 0:
if ref_index > 0:
dna_errors += [
(str(ref_index), 'd', str(l))
] # deletion length l starting at ref_index in 0-count
i += l
ref_index += l
else:
dna_errors += [(str(ref_index), 'f')]
break
elif ref[i] == '-':
# start of an insertion
l = indel_len(ref, i)
# check for frameshifts
if l % 3 == 0:
if ref_index > 0:
dna_errors += [(str(ref_index), 'i', str(read[i:i + l]))]
i += l
else:
dna_errors += [(str(ref_index), 'f')]
break
else:
# substitution
if ref_index > 0:
dna_errors += [(str(ref_index + 1), 's', str(read[i]))]
i += 1
ref_index += 1
return tuple(dna_errors)