def annotate_guides_with_score(candidates_count_dictionary, jellyfish_filename,
priors, posteriors, max_hd, target_string,
target_coverage):
iteration_count = 0
list_candidates = []
for candidate in list(candidates_count_dictionary.keys()):
strand_type = candidates_count_dictionary[candidate]
trie = generate_adjacent_mers(candidate, max_hd)
value1 = value2 = 0.0
print('processing candidate ' + candidate)
flag = True
for mer in trie.keys():
if strand_type == '+':
cp = get_score(candidate, mer)
else:
cp = get_score(reverse_complement(candidate),
reverse_complement(mer))
qf = jellyfish.QueryMerFile(jellyfish_filename)
merDNA = jellyfish.MerDNA(mer)
rev_comp_merDNA = jellyfish.MerDNA(reverse_complement(mer))
k = max(qf[merDNA], qf[rev_comp_merDNA])
if k <= 0:
continue
if k >= max_k:
flag = False
break
p = float(target_string.count(mer))
accum = 0.0
for count in range(1, max_limit_count):
probability = get_probability(count, k)
p_count = priors[count]
p_k = posteriors[k]
new_val = 1.0 * probability * count * p_count / p_k
accum = accum + new_val
value1 = value1 + cp * p
value2 = value2 + cp * accum
if value1 <= 0.0 or flag is False:
continue
score = 1.0 * value2 / (value1 * target_coverage)
qf = jellyfish.QueryMerFile(jellyfish_filename)
merDNA = jellyfish.MerDNA(candidate)
k = max(qf[merDNA],
qf[jellyfish.MerDNA(reverse_complement(candidate))])
list_candidates.append((candidate, score, k, trie, strand_type))
iteration_count = iteration_count + 1
print('processed ' + str(iteration_count) + 'th gRNA: ' + candidate +
' with score= ' + str(score))
print('DONE! Sorting...')
list_candidates.sort(key=sort_second)
print('Final list:')
f = open('scores', 'w')
for annotated_candidate in list_candidates:
print(annotated_candidate)
f.write(str(annotated_candidate[1]) + '\n')
f.close()
return list_candidates
def query(self, seq):
"""Fetch kmer count data from database."""
kmer = jellyfish.MerDNA(seq)
if (self.canonical):
kmer.canonicalize()
return self.jf[kmer]
def generate_k_spectrum_of_target_and_count(target_string,
jellyfish_count_file,
max_k_limit=200,
k=15):
"""
k-spectrum of target, then count the k-mers found within the target, then generate the histogram
:type max_k_limit: int
:param target_string: the target string
:param k: value of k
:param jellyfish_count_file: jellyfish binary file name
:param max_k_limit: max value upto which the histogram is to be generated
:return: the histogram data in a dictionary
"""
target = target_string
length = len(target)
a = set()
for i in range(length - k):
a.add(target[i:i + k])
lst = []
qf = jellyfish.QueryMerFile(jellyfish_count_file)
for substr in a:
mer = jellyfish.MerDNA(substr)
count = qf[mer]
lst.append(count)
dic = {}
for i in range(max_k_limit):
dic[i + 1] = lst.count(i + 1)
return dic
def get_kmer_presence(kmerF):
mer = jellyfish.MerDNA(kmerF)
mer.canonicalize()
kmer_pres = []
pres = int(qjellies[mer] > 0)
if pres:
return None
else:
kmer_pres.append(1)
return kmer_pres
def test_canonical_mers(self):
good = True
mers = jf.string_canonicals(self.str)
for count, m in enumerate(mers):
m2 = jf.MerDNA(self.str[count:count + self.k])
rm2 = m2.get_reverse_complement()
good = good and (m == m2 or m == rm2)
good = good and (not (m > m2)) and (not (m > rm2))
# count += 1
self.assertTrue(good)
self.assertEqual(len(self.str) - self.k + 0, count)
def test_all_mers(self):
count = 0
good1 = True
good2 = True
mers = jf.string_mers(self.str)
for m in mers:
m2 = jf.MerDNA(self.str[count:count + self.k])
good1 = good1 and m == m2
good2 = good2 and self.str[count:count + self.k].upper() == str(m2)
count += 1
self.assertTrue(good1)
self.assertTrue(good2)
self.assertEqual(len(self.str) - self.k + 1, count)
def test_add(self):
mer = jf.MerDNA()
good = True
for i in range(1000):
mer.randomize()
val = random.randrange(1000)
good = good and self.hash.add(mer, val)
if not good: break
if i % 3 > 0:
nval = random.randrange(1000)
val = val + nval
if i % 3 == 1:
good = good and (not self.hash.add(mer, nval))
else:
good = good and self.hash.update_add(mer, nval)
if not good: break
good = good and (val == self.hash.get(mer)) and (val == self.hash[mer])
if not good: break
self.assertTrue(good)
def annotate_guides_with_score_parallel(candidates_count_dictionary, jellyfish_filename, priors, posteriors,
max_hd,
target_string,
return_list):
index = 0
list_candidates = []
for candidate in list(candidates_count_dictionary.keys()):
strand_type = candidates_count_dictionary[candidate][0]
trie = generate_adjacent_mers(candidate, max_hd)
value1 = value2 = 0.0
flag = True
for mer in trie.keys():
if strand_type == '+':
cp = get_score(candidate, mer)
else:
cp = get_score(reverse_complement(candidate), reverse_complement(mer))
qf = jellyfish_filename
merDNA = jellyfish.MerDNA(mer)
merDNA.canonicalize()
k = qf[merDNA]
if k <= 0:
continue
if k >= max_k:
flag = False
break
p = float(target_string.count(mer))
accum = 0.0
for count in range(1, max_limit_count):
probability = get_probability(count, k)
p_count = priors[count]
p_k = posteriors[k]
new_val = 1.0 * probability * count * p_count / p_k
accum = accum + new_val
value1 = value1 + cp * p
value2 = value2 + cp * accum
if value1 <= 0.0 or flag is False:
continue
score = 1.0 * value2 / value1
return_list[index] = score
index = index + 1
def generate_k_spectrum_of_target_and_count(target_string, jellyfish_count_file, max_k_limit):
"""
k-spectrum of target, then count the k-mers found within the target, then generate the histogram
:type max_k_limit: int
:param target_string: the target string
:param jellyfish_count_file: jellyfish binary file (jellyfish.QueryMerFile)
:param max_k_limit: max value upto which the histogram is to be generated
:return: the histogram data in a dictionary as k_spectrum, and the counts of k-mers indexed as positions
"""
# a pair is returned
# pair.first = the k-spectrum histogram of k-mers taken only from the target region
# pair.second = a hash-map that has keys:positions in target, values:count of a k-mer in that position
k = candidate_length
target = target_string
length = len(target)
a = set()
counts_in_positions = {}
k_spectrum = {}
#qf = jellyfish.QueryMerFile(jellyfish_count_file)
qf = jellyfish_count_file
for i in range(length - k + 1):
subst = target[i:i + k]
mer = jellyfish.MerDNA(subst)
mer.canonicalize()
count = qf[mer]
counts_in_positions[i] = count
if count == 0:
logging.info("Count = 0 for substring " + subst)
continue
if subst not in a:
a.add(subst)
if count in k_spectrum.keys():
k_spectrum[count] += 1
else:
k_spectrum[count] = 1
return k_spectrum, counts_in_positions
def compute_hybrid(self, first_var, var_wgts):
import dna_jellyfish
r = self.r
chrom = self.variants[first_var].chrom
pos = self.variants[first_var].pos
#if self.variants[first_var].pos < pos or self.variants[first_var].pos >= pos+r:
# return
# Number of variants in window starting at this one
k = 1
while first_var + k < self.num_v and self.variants[
first_var +
k].chrom == chrom and self.variants[first_var +
k].pos < pos + r:
k += 1
#if k > 14:
# sys.stdout.write('Processing variant %d with %d neighbors' % (first_var, k))
if k > self.max_v_in_window:
alt_freqs = [(sum(self.variants[first_var + j].probs),
first_var + j) for j in range(1, k)]
ids = [first_var] + [
f[1] for f in sorted(alt_freqs, reverse=True)
[:self.max_v_in_window - 1]
]
it = PseudocontigIterator(self.genome[chrom],
[self.variants[v] for v in ids], self.r)
else:
ids = range(first_var, first_var + k)
it = PseudocontigIterator(self.genome[chrom],
self.variants[first_var:first_var + k],
r)
pseudocontig = it.next()
while pseudocontig:
vec = it.curr_vec
p = self.prob_read(self.variants, ids, vec)
for i in range(len(pseudocontig) - self.r + 1):
mer = dna_jellyfish.MerDNA(pseudocontig[i:i + r])
mer.canonicalize()
c_linear = self.h_ref[mer]
if not c_linear:
c_linear = 0
c_added = self.h_added[mer]
if not c_added:
c_added = 0
if c_added == 0:
print(
'Error! Read %s from added pseudocontigs could not be found (SNPs %d - %d)'
%
(pseudocontig[i:i + r], first_var, first_var + k))
for j in range(first_var, first_var + k):
print('%s: %d, %s --> %s' %
(self.variants[j].chrom,
self.variants[j].pos, self.variants[j].orig,
','.join(self.variants[j].alts)))
exit()
c_total = c_linear + c_added
if c_total == 0:
print('Variants %d -%d / %d' %
(first_var, first_var + k - 1, self.num_v))
print('Vector: ' + str(vec))
print('Pseudocontig: ' + str(pseudocontig))
print('Read: ' + str(pseudocontig[i:i + r]))
exit()
# Average relative probability of this read's other mappings
avg_wgt = c_linear * self.wgt_ref + (c_added -
1) * self.wgt_added
hybrid_wgt = (p - avg_wgt) / (c_total)
for j in range(len(ids)):
if vec[j]:
var_wgts[ids[j]] -= hybrid_wgt
pseudocontig = it.next()
def string_to_kmer(self, sequence):
binary_kmer = jf.MerDNA(sequence)
return binary_kmer