def profile_most_probable_kmer(self, profile, dna=None):
"""takes a profile matrix. validate it, removes 0 error.
:returns profile most probable kmer."""
if dna is None:
dna = self.DNA
if True in [0 in pro for pro in profile]:
print('profile contains 0. changing it...')
for i in range(4):
for j in range(len(profile[i])):
profile[i][j] += 0.2
kmers = ngram(dna.upper(), len(profile[0]))
pro = []
for i in range(len(kmers)):
b = list(kmers[i])
x = 1
for j in range(len(b)):
latter = b[j]
if latter == 'A':
x *= profile[0][j]
if latter == 'C':
x *= profile[1][j]
if latter == 'G':
x *= profile[2][j]
if latter == 'T':
x *= profile[3][j]
pro.append(x)
return kmers[pro.index(max(pro))]
def approximate_matched_pattern(self, pattern, max_mismatch, dna=None):
""":returns a list of starting indexes of approximate matched patterns"""
if dna is None:
dna = self.DNA
k_gram = ngram(dna, len(pattern))
return [
i for i, j in enumerate(k_gram)
if hamming_distance(j, pattern) <= max_mismatch
]
def most_frequent_k_mer(self, k):
""":returns the list of most frequent kmers in a DNA sequence."""
result_list = []
k_gram = ngram(self.DNA, k)
set_gram = list(set(k_gram))
set_count = []
for a in set_gram:
set_count.append(k_gram.count(a))
m = max(set_count)
for a in range(len(set_count)):
if set_count[a] == m:
result_list.append(set_gram[a])
return result_list
def most_frequent_k_mer_with_mismatch(self, k, max_mismatch):
""":returns the list of most frequent kmers of length k in
a DNA sequence allowing maximum mismatch of max_mismatch"""
k_mers = list(set(ngram(self.DNA, k)))
frequency_counts = [
len(self.approximate_matched_pattern(gram, max_mismatch))
for gram in k_mers
]
m = max(frequency_counts)
return [
k_mers[i] for i in range(len(frequency_counts))
if frequency_counts[i] == m
]
def most_frequent_k_mer_with_mismatch_and_complements(
self, k, max_mismatch):
""":returns the list of most frequent kmers of length k in a DNA sequence allowing
maximum mismatch of max_mismatch also including occurrences of there reverse compliments"""
k_mers = list(set(ngram(self.DNA, k)))
frequency_counts = [
(len(self.approximate_matched_pattern(gram, max_mismatch)) + len(
self.approximate_matched_pattern(
self.reverse_complement(dna=gram), max_mismatch)))
for gram in k_mers
]
m = max(frequency_counts)
return [
k_mers[i] for i in range(len(frequency_counts))
if frequency_counts[i] == m
]
def lt_clump(self, k, l, t, dna=None):
"""
:param k: length of kmers
:param l: length L
:param t:minimum number of kmers in L
:param dna: optional DNA sequence
:return: list of kmers for which L-t clump is present in the dna sequence.
"""
if dna is None:
dna = self.DNA
clumps = []
k_mers = list(set(ngram(dna, k)))
for each in k_mers:
for i in range(len(dna) - l + 1):
if dna[i] == each[0]:
if self.count_k_mer(each, dna=dna[i:i + l]) >= t:
clumps.append(each)
break # if present no need to check further for a kmer
return clumps