def createMapping(set_a, set_b):
sol = {}
for sa in set_a:
a = sa.split()
candidate = None
candidate_val = 1e4
for sb in set_b:
b = sb.split()
assert len(a) == 2
assert len(b) == 2
v1 = edlib.align(a[0], b[0])['editDistance'] + edlib.align(
a[1], b[1])['editDistance']
v2 = edlib.align(a[0], b[1])['editDistance'] + edlib.align(
a[1], b[0])['editDistance']
d = min(v1, v2)
if d < candidate_val:
candidate = sb
candidate_val = d
sol[sa] = candidate
if candidate_val > 4:
sol[sa] = None
print(sa, candidate, candidate_val)
return sol
def revprimerStrip(file):
base = os.path.basename(file).split('.')[0]
goodseq = os.path.join(tmpdir, base + '.good')
badseq = os.path.join(tmpdir, base + '.bad')
with open(goodseq, 'w') as good:
with open(badseq, 'w') as bad:
for title, seq, qual in FastqGeneralIterator(open(file)):
foralign = edlib.align(args.rev_primer,
seq,
mode="HW",
k=args.primer_mismatch,
additionalEqualities=amptklib.degenNuc)
if foralign["editDistance"] >= 0:
ForCutPos = foralign["locations"][0][1] + 1
Seq = seq[ForCutPos:]
Qual = qual[ForCutPos:]
#align reverse
revalign = edlib.align(
RevForPrimer,
Seq,
mode="HW",
task="locations",
k=args.primer_mismatch,
additionalEqualities=amptklib.degenNuc)
if revalign["editDistance"] >= 0:
RevCutPos = revalign["locations"][0][0]
Seq = Seq[:RevCutPos]
Qual = Qual[:RevCutPos]
good.write("@%s\n%s\n+\n%s\n" % (title, Seq, Qual))
else:
bad.write("@%s\n%s\n+\n%s\n" % (title, seq, qual))
def remove_title_page(txt_file_without_tp, txt_file_with_tp, num_lines,
out_dir):
result = None
lines_without = txt_file_without_tp.readlines()
lines_with = txt_file_with_tp.readlines()
without_txt = ''.join(lines_without[:num_lines]).lower()
with_txt = ''.join(lines_with[:num_lines]).lower()
res = edlib.align(without_txt, with_txt)
prev_ld = res['editDistance']
for i in range(num_lines):
without_txt = ''.join(lines_without[:num_lines]).lower()
with_txt = ''.join(lines_with[i:num_lines]).lower()
res = edlib.align(without_txt, with_txt)
ld = res['editDistance']
if ld > prev_ld:
result = ''.join(lines_with[i - 1:])
break
elif ld < prev_ld:
prev_ld = ld
if result is None:
warnings.warn('No title page found')
out_file = out_file_name(out_dir, txt_file_with_tp.name)
shutil.copy2(txt_file_with_tp.name, out_file)
else:
out_file = out_file_name(out_dir, txt_file_with_tp.name)
with codecs.open(out_file, 'w', encoding='utf8') as f:
f.write(result)
def check_library_barcode(b1: str, b2: str) -> bool:
"""only some library with libary barcode."""
# if len(b1) == 0 or len(b2) == 0:
# return False
d1 = edlib.align(b1, _LIB5_SEQ, mode="NW", task="distance")["editDistance"]
d2 = edlib.align(b2, _LIB3_SEQ, mode="NW", task="distance")["editDistance"]
return d1 <= 3 and d2 <= 3
def detect_read_strand(read_5p_seq, upstream_context_fwd,
upstream_context_rev):
result_fwd = edlib.align(upstream_context_fwd, read_5p_seq, mode="HW", k=5)
result_rev = edlib.align(upstream_context_rev, read_5p_seq, mode="HW", k=5)
strand = "-"
if result_fwd["editDistance"] < result_rev["editDistance"]:
strand = "+"
return strand
def sequences_overlap(query, targets, min_match=30, check_revcom=True):
"""Check if a query overlaps with any sequence in targets."""
for target in targets:
cigar = align(query, target, 'HW', 'path')['cigar']
if cigar_to_max_operation(cigar) > min_match:
return True
elif check_revcom:
cigar = align(revcom(query), target, 'HW', 'path')['cigar']
if cigar_to_max_operation(cigar) > min_match:
return True
return False
def get_direction(query, ref):
"""
Determine the strand direction for query sequence by comparison with
ref sequence
"""
fwd = query
rev = query.reverse_complement()
d_fwd = edlib.align(str(fwd.seq), str(ref.seq), task="distance",
mode="NW")["editDistance"]
d_rev = edlib.align(str(rev.seq), str(ref.seq), task="distance",
mode="NW")["editDistance"]
return fwd if d_fwd < d_rev else rev
def compare_polished_sequences(self, final_sequences):
report_fn = os.path.join(self.params.outdir, 'report.txt')
with open(report_fn, 'w') as f:
for i in range(1, self.params.num_iters):
seq_i, seq_i1 = final_sequences[i], final_sequences[i+1]
alignment = edlib.align(seq_i, seq_i1)
print(f'Alignment polishing seq {i} vs {i+1}:', file=f)
print(alignment, file=f)
hpc_seq_i = compress_homopolymer(final_sequences[i])
hpc_seq_i1 = compress_homopolymer(final_sequences[i+1])
alignment = edlib.align(hpc_seq_i, hpc_seq_i1)
print(f'Alignment homopolymer compressed polishing seq {i} vs {i+1}:', file=f)
print(alignment, file=f)
def create_synteny_matrix_mul(self, gene_seq, g1, g2, n):
for gene in g1:
if gene == "NULL_GENE":
continue
try:
_ = gene_seq[gene]
except BaseException:
return np.zeros((n, n, 2)), np.zeros((n, n, 2))
for gene in g2:
if gene == "NULL_GENE":
continue
try:
_ = gene_seq[gene]
except BaseException:
return np.zeros((n, n, 2)), np.zeros((n, n, 2))
sm = np.zeros((n, n, 2))
sml = np.zeros((n, n, 2))
for i in range(n):
if g1[i] == "NULL_GENE":
continue
if gene_seq[g1[i]] == "":
return np.zeros((n, n, 2)), np.zeros((n, n, 2))
for j in range(n):
if g2[j] == "NULL_GENE":
continue
if gene_seq[g2[j]] == "":
return np.zeros((n, n, 2)), np.zeros((n, n, 2))
norm_len = max(len(gene_seq[g1[i]]), len(gene_seq[g2[j]]))
try:
result = ed.align(gene_seq[g1[i]],
gene_seq[g2[j]],
mode="NW",
task="distance")
sm[i][j][0] = result["editDistance"] / (norm_len)
result = ed.align(gene_seq[g1[i]],
gene_seq[g2[j]][::-1],
mode="NW",
task="distance")
sm[i][j][1] = result["editDistance"] / (norm_len)
_, result, _ = local_pairwise_align_ssw(
DNA(gene_seq[g1[i]]), DNA(gene_seq[g2[j]]))
sml[i][j][0] = result / (norm_len)
_, result, _ = local_pairwise_align_ssw(
DNA(gene_seq[g1[i]]), DNA(gene_seq[g2[j]][::-1]))
sml[i][j][1] = result / (norm_len)
except BaseException:
return np.zeros((n, n, 2)), np.zeros((n, n, 2))
return sm, sml
def replace_word(self, word, normalize=True):
"""uses levenshtein by default, no argument to avoid time spent on conditionals"""
"""edlib.align faster than levenshteinDistance"""
if normalize:
scores = []
for vocab_word in self.limited_vocab:
align_score = align(vocab_word, word)
edit_distance = align_score['editDistance']
if edit_distance != 0:
edit_distance /= align_score['alphabetLength']
scores.append(edit_distance)
scores = np.array([scores])
else:
scores = np.array([align(vocab_word, word)['editDistance'] for vocab_word in self.limited_vocab])
replacement_word = self.idx2word[np.argmin(scores)]
self.word2idx[word] = self.word2idx[replacement_word]
def edist(lst):
if len(str(lst[0])) == 0:
return len(str(lst[1]))
if len(str(lst[1])) == 0:
return len(str(lst[0]))
result = edlib.align(str(lst[0]),
str(lst[1]),
mode="NW",
additionalEqualities=[('U', 'T'), ('R', 'A'),
('R', 'G'), ('Y', 'C'),
('Y', 'T'), ('Y', 'U'),
('K', 'G'), ('K', 'T'),
('K', 'U'), ('M', 'A'),
('M', 'C'), ('S', 'C'),
('S', 'G'), ('W', 'A'),
('W', 'T'), ('W', 'U'),
('B', 'C'), ('B', 'G'),
('B', 'T'), ('B', 'U'),
('D', 'A'), ('D', 'G'),
('D', 'T'), ('D', 'U'),
('H', 'A'), ('H', 'C'),
('H', 'T'), ('H', 'U'),
('V', 'A'), ('V', 'C'),
('V', 'G'), ('N', 'C'),
('N', 'C'), ('N', 'G'),
('N', 'T'), ('N', 'U')])
return result["editDistance"]
def identity_test(self, target_identity, read_length, error_model, qscore_model):
target_errors = 1.0 - target_identity
read_delta = self.read_delta * target_errors
mean_delta = self.mean_delta * target_errors
if VERBOSE:
print(f'\nRead length: {read_length}, target identity: {target_identity}')
print(f' allowed error per read: {read_delta:.4f}')
print(f' allowed error in mean: {mean_delta:.4f}')
print(' identities: ', end='')
read_identities = []
for i in range(self.trials):
frag = badread.misc.get_random_sequence(read_length)
seq, qual, _, _ = badread.simulate.sequence_fragment(frag, target_identity,
error_model, qscore_model)
cigar = edlib.align(frag, seq, task='path')['cigar']
read_identity = badread.misc.identity_from_edlib_cigar(cigar)
read_identities.append(read_identity)
if VERBOSE:
print('{:.4f}'.format(read_identity), flush=True,
end='\n ' if (i+1) % 20 == 0 else ' ')
self.assertAlmostEqual(read_identity, target_identity, delta=read_delta)
mean_identity = statistics.mean(read_identities)
if VERBOSE:
print('\r' if self.trials % 20 == 0 else '\n', end='')
print(f' mean: {mean_identity:.4f}')
self.assertAlmostEqual(mean_identity, target_identity, delta=mean_delta)
if VERBOSE:
print(' PASS')
def get_aligement(self, fast5_path, ref_path, verbose, fasta_out_dir=None, ref=None):
t = perf_counter()
fasta_out = None
if fasta_out_dir is not None:
fasta_out = os.path.join(fasta_out_dir, os.path.splitext(fast5_path)[0].split('/')[-1] + ".fasta")
basecalled = self.basecall_sample(fast5_path, fasta_out=fasta_out, ref=ref)
with open(ref_path) as f:
target = f.readlines()[-1]
if verbose:
print("Basecalled:\n", basecalled)
print("Target\n", target)
self.logger.debug("fast5_path %s, ref_path %s", fast5_path, ref_path)
result = edlib.align(basecalled, target, task='path')
cigar_pairs = butils.cigar_str_to_pairs(result['cigar'])
self.logger.debug("\nBasecalled: %s\nTarget : %s",
butils.query_align_string(basecalled, cigar_pairs),
butils.reference_align_string(target, cigar_pairs)
)
self.logger.debug("extCigar %s", result['cigar'])
self.logger.debug("Whole time %.3f", perf_counter() - t)
cigar_stat = defaultdict(int)
total = 0
for num, op in breakCigar(result['cigar']):
total += num
cigar_stat[op] += num
acc = cigar_stat['='] / total
nedit = result['editDistance'] / len(target)
return nedit, acc, len(basecalled), cigar_stat
def custom_sequencing_adapter_check(self, r1_seq):
''' Check for custom sequencing adapter on r1
:param bytes r1_seq: R1 sequence
:rtype int
:returns end pos of adapter, -1 if not found
'''
# NOTE: Will check multiple adapters for all reads
# Should the adapter be identified first and only that sequence used going forward ?
# Maybe add this logic in the trim_custom_sequencing_adapter function.
best_adapter = {"seq": None, "align": None, "score": None}
for a in self.custom_seq_adapter:
alignment = edlib.align(a,
r1_seq[0:len(a) + 3],
mode="SHW",
task="locations")
score = float(alignment["editDistance"]) / len(
self.custom_seq_adapter)
if best_adapter["seq"] is None or score < best_adapter["score"]:
best_adapter["seq"] = a
best_adapter["align"] = alignment
best_adapter["score"] = score
if best_adapter["score"] < 0.18:
return best_adapter["align"]["locations"][-1][1]
else:
return -1
def _cutr(reads, adp, th, r, len_list=[]):
iden_max = -1
match_num = 0
cut_pos = []
skip_num = 0
repat = re.compile(r'(\d+)[DHIMNPSX=]{1}')
if len(reads[0]) > 2:
has_qual = True
else:
has_qual = False
for read in reads:
read_length = len(read[1])
if not len_list:
pass
else:
len_list.append(read_length)
if read_length < 2 * r:
#logger.debug("%s is too short: %d bp. Skip." % (read[0], read_length))
skip_num += 1
continue
result = edlib.align(adp, read[1][-r:], mode="HW", task='path')
identity = 1.0 - float(result['editDistance'] / np.sum(
[int(i) for i in repat.findall(result['cigar'])]))
if identity > th:
start = len(read[1]) - r + result['locations'][0][0]
cut_pos.append(r - result['locations'][0][0])
match_num += 1
if identity > iden_max:
iden_max = identity
read[1] = read[1][:start]
if has_qual:
read[2] = read[2][:start]
logger.info("%d reads were skipped due to their short lengths." % skip_num)
return (iden_max, match_num, cut_pos)
def get_pairwise_alignments(seqs):
section_header('Pairwise global alignments')
explanation(
'Trycycler uses the edlib aligner to get global alignments between all pairs of '
'sequences. This can help you to spot any problematic sequences that should be '
'excluded before continuing. If you see any sequences with notably worse '
'identities or max indels, you can remove them (delete the contig\'s FASTA) and '
'run this command again.')
seq_names = list(seqs.keys())
max_seq_name_len = max(len(x) for x in seq_names)
pairwise_cigars, percent_identities, max_indels = {}, {}, {}
for i, a in enumerate(seq_names):
seq_a = seqs[a]
for j in range(i + 1, len(seq_names)):
b = seq_names[j]
seq_b = seqs[b]
log(' ' * (max_seq_name_len - len(a)) + a, end='')
log(' vs ', end='')
log(b + '...' + ' ' * (max_seq_name_len - len(b)), end=' ')
result = edlib.align(seq_a, seq_b, mode="NW", task="path")
cigar = result['cigar']
percent_identity, max_indel = identity_and_max_indel_from_cigar(
cigar)
log(f'{percent_identity:.2f}% identity, max indel = {max_indel}')
pairwise_cigars[(a, b)] = cigar
percent_identities[(a, b)] = percent_identity
percent_identities[(b, a)] = percent_identity
max_indels[(a, b)] = max_indel
max_indels[(b, a)] = max_indel
log()
return pairwise_cigars, percent_identities, max_indels
def get_global_aln_results(ref_seq, query_seq, min_seq_len):
"""
Aligns two sequences globally, and returns (delta_len, idt, cov) values
compatible with the legacy deduplication code.
Currently unused - it was used in an intermediate version, and might be useful
at some point in the future.
"""
log('Aligning (Edlib): len(ref_seq) = %d, len(query_seq) = %d' % (len(ref_seq), len(query_seq)))
delta_len = len(query_seq) - len(ref_seq)
idt = 0.0
cov = 0.0
if len(ref_seq) < min_seq_len or len(query_seq) < min_seq_len:
return delta_len, idt, cov
result = edlib.align(query_seq, ref_seq, mode="NW", task="path")
cigar = result['cigar']
num_m, num_i, num_d, total_len = count_cigar_ops(result['cigar'])
num_eq = (num_m + num_i + num_d) - result['editDistance']
num_x = num_m - num_eq
idt_query = float(num_eq) / float(num_eq + num_x + num_i)
idt_ref = float(num_eq) / float(num_eq + num_x + num_d)
idt = min(idt_query, idt_ref)
cov = 1.0
log(' - Alignment stats: num_m = %d, num_i = %d, num_d = %d, total_len = %d, num_eq = %d, num_x = %d' % (num_m, num_i, num_d, total_len, num_eq, num_x))
return delta_len, idt, cov
def sequences_match(seq1, seq2, compare='5p_clip', min_length=4):
"""
Compare 2 sequences and determine whether they are likely the same.
>>> s1, s2 = 'AAAAAAAAAA', 'AAAAAAAAATGC'
>>> sequences_match(s1, s2, compare='3p_clip')
True
>>> sequences_match('TTGAAAAAAA', s2, compare='3p_clip')
False
>>> sequences_match(s1[::-1], s2[::-1], compare='5p_clip')
True
>>> sequences_match('A', 'A')
False
"""
if compare == '3p_clip':
seq1 = seq1[:10]
seq2 = seq2[:10]
else:
seq1 = seq1[-10:]
seq2 = seq2[-10:]
if len(seq1) > min_length and len(seq2) > min_length:
if len(seq1) >= 10 and len(seq2) >= 10:
if align(seq1, seq2)['editDistance'] < 2:
return True
elif compare == '3p_clip':
return seq1.startswith(seq2) or seq2.startswith(seq1)
else:
return seq1.endswith(seq2) or seq2.endswith(seq1)
return False
def print_alignment(s1, s2, k=None, mode='NW', width=60):
if k is None:
k = max(len(s1), len(s2))
alignment = edlib.align(s1, s2, task='path', k=k, mode=mode)
cigar = alignment['cigar']
locs = alignment['locations']
st, en = locs[0]
_, _, a1, a2 = parse_cigar(cigar, s1, s2[st:])
status = []
for c1, c2 in zip(a1, a2):
if c1 == c2:
status.append('|')
elif c1 == '-' or c2 == '-':
status.append('-')
elif c1 != c2:
status.append('X')
else:
assert False
a1 = ''.join(a1)
status = ''.join(status)
a2 = ''.join(a2)
a1 = textwrap.wrap(a1, width=width)
status = textwrap.wrap(status, width=width)
a2 = textwrap.wrap(a2, width=width)
for wa1, wst, wa2 in zip(a1, status, a2):
print(wa1)
print(wst)
print(wa2)
print("")
def min_ed(max_insertion, q_ins):
result = edlib.align(max_insertion, q_ins, task="path", mode="NW")
cigar = result["cigar"]
tuples = []
# do not allow deletions in max_insertion: because we do not want to alter this sequence
if "D" in cigar:
return ""
matches = re.split(r'[=DXSMI]+', cigar)
i = 0
for length in matches[:-1]:
i += len(length)
type_ = cigar[i]
i += 1
tuples.append((int(length), type_))
q_insertion_modified = ""
q_ins_pos = 0
for length, type_ in tuples:
# if we reach here we are guaranteed no deletions in alignment of max_insertion
# we therefore simply thread in the matching or mismatching characters (if type '=' or 'X')
# or we put a "-" (if type is 'I')
if type_ == "I":
q_insertion_modified += "-" * length
else:
q_insertion_modified += q_ins[q_ins_pos:q_ins_pos + length]
q_ins_pos += length
return q_insertion_modified
def getRatio(self, stringA=None, stringB=None):
"""
With default(None) atributes - returns editdistance ratio based on main files.
With different strings can be helpful for substring comparision (.getRatio(SubS1,Subs2))
>>> ED1 = EditDistance("testABC1", "testABB2")
>>> ED1.getRatio()
0.75
>>> ED1.getRatio("Ala ma kota", "Alan ma psa")
0.6363636363636364
>>> ED1.getRatio(1,"999")
-1
>>> ED1.getRatio(3987,4789)
-1
"""
if (stringA == None and stringB == None):
stringA, stringB = self.string1, self.string2
alignmentEditDistance = self.editDistance
else:
if not (type(stringA) == str and type(stringB) == str):
return -1
alignmentEditDistance = edlib.align(stringA, stringB,
task="path")["editDistance"]
stringLenMax = max(len(stringA), len(stringB))
ratioAB = 1 - (float(alignmentEditDistance) / float(stringLenMax))
return ratioAB
def align(query, pattern_info, max_ed, normalise=False):
pattern, wildcard, equalities, forward = pattern_info
result = edlib.align(
pattern,
query,
task="path",
mode="HW",
k=max_ed,
additionalEqualities=equalities,
)
if result["editDistance"] == -1:
return None, None
ed = result["editDistance"]
if not normalise:
locs = result["locations"][0]
umi = query[locs[0]:locs[1] + 1]
return ed, umi
# Extract and normalise UMI
umi = ""
align = edlib.getNiceAlignment(result, pattern, query)
for q, t in zip(align["query_aligned"], align["target_aligned"]):
if q != wildcard:
continue
if t == "-":
umi += "N"
else:
umi += t
if len(umi) != 16:
raise RuntimeError("UMI length incorrect: {}".format(umi))
return ed, umi
def edlib_traceback_allow_ends(x, y, mode="NW", task="path", k=1, end_threshold = 0):
result = edlib.align(x, y, mode=mode, task=task, k=k)
ed = result["editDistance"]
locations = result["locations"]
cigar = result["cigar"]
if cigar:
tuples = []
result = re.split(r'[=DXSMI]+', cigar)
i = 0
for length in result[:-1]:
i += len(length)
type_ = cigar[i]
i += 1
tuples.append((length, type_ ))
ed_ignore_ends = ed
if tuples[0][1] == "D" or tuples[0][1] == "I":
begin_snippet = int(tuples[0][0])
if begin_snippet <= end_threshold:
ed_ignore_ends -= int(begin_snippet)
if tuples[-1][1] == "D" or tuples[-1][1] == "I":
end_snippet = int(tuples[-1][0])
if end_snippet <= end_threshold:
ed_ignore_ends -= int(end_snippet)
# if ed > ed_ignore_ends:
# print("ed global:", ed, "ed after:", ed_ignore_ends)
ed = ed_ignore_ends
# if ed ==0:
# print("here")
return ed, locations, cigar
def edist(lst):
if len(str(lst[0])) == 0:
return -1, ""
if len(str(lst[1])) == 0:
return -1, ""
result = edlib.align(str(lst[0]), str(lst[1]), mode="NW", task="path")
return result["editDistance"], result["cigar"]
def get_align_index_path(query: Iterable,
target: Iterable) -> List[CharAlignToken]:
path_ = edlib.align(query, target, task="path")["cigar"]
if path_ is None:
return []
path_ = expand_cigar_format(path_)
index_out = 0
index_path = 0
out = []
for index_query in range(len(query)):
while path_[index_path] == "D":
index_out += 1
index_path += 1
action = path_[index_path]
out.append(CharAlignToken(index_out, action))
if action == "=":
assert query[index_query] == target[index_out]
if action in ["=", "X"]:
index_out += 1
index_path += 1
return out
def exec(peptide, time_node):
file = open("../src/public/jobs/service3/service3.fasta", "w")
file.write(peptide)
file.close()
fasta = SeqIO.parse("../src/public/jobs/service3/service3.fasta", "fasta")
if(any(fasta) == False): #False when `fasta` is empty
return "error"
count = 0
for record in SeqIO.parse("../src/public/jobs/service3/service3.fasta", "fasta"):
sequence_input = str(record.seq)
count = count+1
print(count)
print(sequence_input)
if (count > 1):
return "error"
dataset = pd.read_csv("data_values_activity_non_modified.csv")
dict_response = []
for i in range(len(dataset)):
align_result = edlib.align(sequence_input, dataset['sequence'][i], mode = "HW", task = "path")
view_alignment = edlib.getNiceAlignment(align_result, sequence_input, dataset['sequence'][i])
dict_aligment = {"input_sequence":view_alignment['query_aligned'], "space_format":view_alignment['matched_aligned'], "compare_sequence": view_alignment['target_aligned'], "id_sequence" : str(dataset['index_sequence'][i]), 'distance_sequences':str(align_result['editDistance'])}
dict_response.append(dict_aligment)
dict_data_results = {"summary_alignment":dict_response}
#export result alignment
return dict_data_results
def edist_nw(lst):
if len(str(lst[0])) == 0:
return -1, ""
if len(str(lst[1])) == 0:
return -1, ""
result = edlib.align(str(lst[0]), str(lst[1]), mode="NW", k=500)
return result["editDistance"]
def distance_matrix(sequences):
"""
Construct a distance matrix from pairwise alignments sequences
"""
dists = np.array([
np.array([0 for _ in range(len(sequences))])
for _ in range(len(sequences))
])
if dists.shape[0] == 0:
return dists
base_seq = sequences[0]
adjusted_sequences = []
for s in tqdm(sequences, desc="{:<10}".format("prescan")):
adjusted_sequences.append(get_direction(s, base_seq))
for i in tqdm(range(len(dists)), desc="{:<10}".format("align")):
query = str(adjusted_sequences[i].seq)
for j in range(i, len(dists[i])):
if i != j:
target = str(adjusted_sequences[j].seq)
d = edlib.align(query, target, task="distance",
mode="NW")["editDistance"]
dists[i][j] = d
dists[j][i] = d
return dists
def find_barcode_locations(center, barcodes, primer_max_ed):
"Find barcodes in a center using edlib"
# Creation of a IUPAC equivalence map for edlib to allow IUPAC code in primers
# The IUPAC map was created with:
# from Bio.Data import IUPACData
# IUPAC_map = [(i, k) for i, j in IUPACData.ambiguous_dna_values.items() for k in j]
IUPAC_map = [('A', 'A'), ('C', 'C'), ('G', 'G'), ('T', 'T'), ('M', 'A'), ('M', 'C'),
('R', 'A'), ('R', 'G'), ('W', 'A'), ('W', 'T'), ('S', 'C'), ('S', 'G'),
('Y', 'C'), ('Y', 'T'), ('K', 'G'), ('K', 'T'), ('V', 'A'), ('V', 'C'),
('V', 'G'), ('H', 'A'), ('H', 'C'), ('H', 'T'), ('D', 'A'), ('D', 'G'),
('D', 'T'), ('B', 'C'), ('B', 'G'), ('B', 'T'), ('X', 'G'), ('X', 'A'),
('X', 'T'), ('X', 'C'), ('N', 'G'), ('N', 'A'), ('N', 'T'), ('N', 'C')]
all_locations = []
for primer_acc, primer_seq in barcodes.items():
# print(primer_acc, primer_seq,center)
# Add additionalEqualities=IUPAC_map allow edlib to understand IUPAC code
result = edlib.align(primer_seq, center,
mode="HW", task="locations", k=primer_max_ed,
additionalEqualities=IUPAC_map)
ed = result["editDistance"]
locations = result["locations"]
print(locations, ed)
if locations:
all_locations.append((primer_acc, locations[0][0], locations[0][1], ed))
return all_locations
def edist(lst):
if len(str(lst[0])) == 0:
return 100500
if len(str(lst[1])) == 0:
return 100500
result = edlib.align(str(lst[0]), str(lst[1]), mode="NW")
return result["editDistance"]
def correct_basecalled(bucketed_basecall, reference, nedit_tol=0.2):
basecalled = "".join(bucketed_basecall)
origin = np.zeros(len(basecalled), dtype=np.int32)
idx = 0
for i, b in enumerate(bucketed_basecall):
origin[idx:idx + len(b)] = i
idx += len(b)
result_set = edlib.align(basecalled, reference, task="path")
nedit = result_set['editDistance'] / len(reference)
if nedit > nedit_tol:
raise TooLargeEditDistance(
"Normalized edit distance is large...%.3f" % nedit
)
result = ["" for _ in bucketed_basecall]
idx_ref = 0
idx_bcalled = 0
for num, op in breakCigar(result_set['cigar']):
for _ in range(num):
if op in CIGAR_MATCH_MISSMATCH:
result[origin[idx_bcalled]] += reference[idx_ref]
idx_bcalled = min(idx_bcalled + 1, len(basecalled) - 1)
idx_ref += 1
elif op in CIGAR_INSERTION:
idx_bcalled += 1
elif op in CIGAR_DELETION:
result[origin[idx_bcalled]] += reference[idx_ref]
idx_ref += 1
return result
def get_aln_data(t_seq, q_seq):
aln_data = []
K = 8
seq0 = t_seq
lk_ptr = kup.allocate_kmer_lookup(1 << (K * 2))
sa_ptr = kup.allocate_seq(len(seq0))
sda_ptr = kup.allocate_seq_addr(len(seq0))
kup.add_sequence(0, K, seq0, len(seq0), sda_ptr, sa_ptr, lk_ptr)
q_id = "dummy"
kmer_match_ptr = kup.find_kmer_pos_for_seq(
q_seq, len(q_seq), K, sda_ptr, lk_ptr)
kmer_match = kmer_match_ptr[0]
if kmer_match.count != 0:
aln_range_ptr = kup.find_best_aln_range(kmer_match_ptr, K, K * 5, 12)
aln_range = aln_range_ptr[0]
s1, e1, s2, e2 = aln_range.s1, aln_range.e1, aln_range.s2, aln_range.e2
log('Mapped (q, s1 = {}, e1 = {}, len1 = {}, (e1 - s1) = {}, t, s2 = {}, e2 = {}, (e2 - s2) = {}, len2 = {})'.format(
s1, e1, e1 - s1, len(q_seq), s2, e2, e2 - s2, len(t_seq)))
if e1 - s1 > 100:
log('Calling edlib.align(q, s1 = {}, e1 = {}, len1 = {}, (e1 - s1) = {}, t, s2 = {}, e2 = {}, (e2 - s2) = {}, len2 = {})'.format(
s1, e1, e1 - s1, len(q_seq), s2, e2, e2 - s2, len(t_seq)))
# Align using Edlib instead of DWA.
edlib_result = edlib.align(q_seq[s1:e1], seq0[s2:e2], mode="NW")
delta_l = len(q_seq) - len(t_seq)
cov = float(e1 - s1) / float(len(q_seq))
idt = float(e1 - s1 - edlib_result['editDistance']) / float(e1 - s1)
aln_data.append((q_id, 0, s1, e1, len(q_seq), s2, e2, len(seq0),
delta_l, idt, cov))
kup.free_aln_range(aln_range_ptr)
kup.free_kmer_match(kmer_match_ptr)
kup.free_kmer_lookup(lk_ptr)
kup.free_seq_array(sa_ptr)
kup.free_seq_addr_array(sda_ptr)
return aln_data #, x, y
except:
print("ERROR: PLEASE RUN pip install edlib")
sys.exit(1)
#read_length is 2nd argument; default 150 bp
try:
read_length = int(sys.argv[2])
except:
read_length = 150
# Max # of mismatches is 3rd argument; default 3
try:
mismatch = int(sys.argv[3])
except:
mismatch = 3
#FASTA file is 1st argument
for rec in SeqIO.parse(sys.argv[1], 'fasta'):
start = str(rec.seq[:read_length])
end = str(rec.seq[len(rec.seq)-read_length:])
if edlib.align(start,end)['editDistance'] <= mismatch:
if sys.argv[4] == 'contigs':
print(">" + str(rec.id))
print(str(rec.seq))
else:
print(rec.id)
with open('../../test_data/Enterobacteria_Phage_1/mutated_90_perc_oneline.fasta', 'r') as f:
queryFull = f.readline()
print('Read query: ', len(queryFull) ,' characters.')
with open('../../test_data/Enterobacteria_Phage_1/Enterobacteria_phage_1_oneline.fa', 'r') as f:
targetFull = f.readline()
print('Read target: ', len(targetFull) ,' characters.')
for seqLen in [30, 100, 1000, 10000, 50000]:
query = queryFull[:seqLen]
target = targetFull[:seqLen]
numRuns = max(1000000000 // (seqLen**2), 1)
print('Sequence length: ', seqLen)
edlibTime = timeit.timeit(stmt="edlib.align(query, target)",
number=numRuns, globals=globals()) / numRuns
print('Edlib: ', edlibTime)
print(edlib.align(query, target))
editdistanceTime = timeit.timeit(stmt="editdistance.eval(query, target)",
number=numRuns, globals=globals()) / numRuns
print('editdistance: ', editdistanceTime)
levenshteinTime = timeit.timeit(stmt="Levenshtein.distance(query, target)",
number=numRuns, globals=globals()) / numRuns
print('levenshtein: ', levenshteinTime)
print('edlib is %f times faster than editdistance.' % (editdistanceTime / edlibTime))
print('edlib is %f times faster than Levenshtein.' % (levenshteinTime / edlibTime))
import sys
import edlib
testFailed = False
result = edlib.align("telephone", "elephant")
if not (result and result["editDistance"] == 3):
testFailed = True
result = edlib.align(b"telephone", b"elephant")
if not (result and result["editDistance"] == 3):
testFailed = True
result = edlib.align("ACTG", "CACTRT", mode="HW", task="path", additionalEqualities=[("R", "A"), ("R", "G")])
if not (result and result["editDistance"] == 0):
testFailed = True
if testFailed:
print("Some of the tests failed!")
else:
print("All tests passed!")
sys.exit(testFailed)