def generate_null_distribution(seq_x, seq_y, scoring_matrix, num_trials):
'''
Takes as input two sequences seq_x and seq_y, a scoring matrix scoring_matrix,
and a number of trials num_trials.
Return a dictionary scoring_distribution that represents an un-normalized
distribution generated by performing the following process num_trials times:
1.Generate a random permutation rand_y of the sequence seq_y using random.shuffle()
2.Compute the maximum value score for the local alignment of seq_x and rand_y
3.Increment the entry score in the dictionary scoring_distribution by one.
'''
scoring_distribution = {}
for _ in range(num_trials):
#shuffle the seq_y
list_y = list(seq_y)
random.shuffle(list_y)
rand_y = "".join(list_y)
alignment_matrix = compute_alignment_matrix(
seq_x=seq_x,
seq_y=rand_y,
scoring_matrix=scoring_matrix,
global_flag=False)
score = max([max(item) for item in alignment_matrix])
#score, align_human, align_fruitfly = compute_local_alignment(seq_x=seq_x, seq_y=rand_y, scoring_matrix=scoring_matrix, alignment_matrix=alignment_matrix)
#import pdb; pdb.set_trace()
try:
scoring_distribution[score] += 1
except (KeyError):
scoring_distribution[score] = 1
scoring_file = open('scoring_distribution.p', 'wb')
pickle.dump(scoring_distribution, scoring_file)
scoring_file.close()
return scoring_distribution
def consensus_alignment(scoring_matrix, align_sequence):
'''
compute the similarity of the two sequences in the local alignment computed to
the PAX Domain.
global alignment of local align_x vs PAX Domain, local align_y vs PAX Domain
return the alignment and the persentage of elements in these two sequences that agree
'''
pax_domain = read_protein('ConsensusPAXDomain.txt')
align_sequence = align_sequence.replace("-", "")
alignment_matrix = compute_alignment_matrix(seq_x=align_sequence,
seq_y=pax_domain,
scoring_matrix=scoring_matrix,
global_flag=True)
score, global_align_sequence, global_align_pax = compute_global_alignment(
seq_x=align_sequence,
seq_y=pax_domain,
scoring_matrix=scoring_matrix,
alignment_matrix=alignment_matrix)
seq = difflib.SequenceMatcher(None, global_align_sequence,
global_align_pax)
ratio = seq.ratio()
#import pdb; pdb.set_trace()
return global_align_sequence, ratio
def random_alignment(scoring_matrix):
'''
Take two random amino acids, compute the alignment and the consensus with pax
Return the ratio of similarity
To examine the solution of the homework
'''
alphabet = "ACBEDGFIHKMLNQPSRTWVYXZ"
length_human = 422
length_fruitfly = 857
random_human = ''
for _ in range(length_human):
random_human += random.choice(alphabet)
random_fruitfly = ''
for _ in range(length_fruitfly):
random_fruitfly += random.choice(alphabet)
alignment_matrix = compute_alignment_matrix(seq_x=random_human,
seq_y=random_fruitfly,
scoring_matrix=scoring_matrix,
global_flag=False)
score, align_human, align_fruitfly = compute_local_alignment(
seq_x=random_human,
seq_y=random_fruitfly,
scoring_matrix=scoring_matrix,
alignment_matrix=alignment_matrix)
global_align_human, ratio_human_random = consensus_alignment(
scoring_matrix, align_human)
global_align_fruitfly, ratio_fruitfly_random = consensus_alignment(
scoring_matrix, align_fruitfly)
return ratio_human_random, ratio_fruitfly_random
def generate_null_distribution(seq_x, seq_y, scoring_matrix, num_trials):
'''
Takes as input two sequences seq_x and seq_y, a scoring matrix scoring_matrix,
and a number of trials num_trials.
Return a dictionary scoring_distribution that represents an un-normalized
distribution generated by performing the following process num_trials times:
1.Generate a random permutation rand_y of the sequence seq_y using random.shuffle()
2.Compute the maximum value score for the local alignment of seq_x and rand_y
3.Increment the entry score in the dictionary scoring_distribution by one.
'''
scoring_distribution = {}
for _ in range(num_trials):
#shuffle the seq_y
list_y = list(seq_y)
random.shuffle(list_y)
rand_y = "".join(list_y)
alignment_matrix = compute_alignment_matrix(seq_x=seq_x, seq_y=rand_y, scoring_matrix=scoring_matrix, global_flag=False)
score = max([max(item) for item in alignment_matrix])
#score, align_human, align_fruitfly = compute_local_alignment(seq_x=seq_x, seq_y=rand_y, scoring_matrix=scoring_matrix, alignment_matrix=alignment_matrix)
#import pdb; pdb.set_trace()
try:
scoring_distribution[score] += 1
except(KeyError):
scoring_distribution[score] = 1
scoring_file = open('scoring_distribution.p', 'wb')
pickle.dump(scoring_distribution, scoring_file)
scoring_file.close()
return scoring_distribution
def random_alignment(scoring_matrix):
'''
Take two random amino acids, compute the alignment and the consensus with pax
Return the ratio of similarity
To examine the solution of the homework
'''
alphabet = "ACBEDGFIHKMLNQPSRTWVYXZ"
length_human = 422
length_fruitfly = 857
random_human = ''
for _ in range(length_human):
random_human += random.choice(alphabet)
random_fruitfly = ''
for _ in range(length_fruitfly):
random_fruitfly += random.choice(alphabet)
alignment_matrix = compute_alignment_matrix(seq_x=random_human, seq_y=random_fruitfly, scoring_matrix=scoring_matrix, global_flag=False)
score, align_human, align_fruitfly = compute_local_alignment(seq_x=random_human, seq_y=random_fruitfly, scoring_matrix=scoring_matrix, alignment_matrix=alignment_matrix)
global_align_human, ratio_human_random = consensus_alignment(scoring_matrix, align_human)
global_align_fruitfly, ratio_fruitfly_random = consensus_alignment(scoring_matrix, align_fruitfly)
return ratio_human_random, ratio_fruitfly_random
def get_edit_distance(seq_x, seq_y, scoring_matrix):
'''
compute the seq_x and seq_y global alignment with scoring matrix
return the edit distance can be expressed in term of:
|x| + |y| - score(x, y)
'''
alignment_matrix = compute_alignment_matrix(seq_x, seq_y, scoring_matrix, True)
score, align_x, align_y = compute_global_alignment(seq_x, seq_y, scoring_matrix, alignment_matrix)
return len(seq_x) + len(seq_y) - score
def protein_alignment(scoring_matrix):
'''
compute the human eyeless protein and fruitfly eyeless portein alignment
'''
human_protein = read_protein('HumanEyelessProtein.txt')
fruitfly_protein = read_protein('FruitflyEyelessprotein.txt')
alignment_matrix = compute_alignment_matrix(seq_x=human_protein, seq_y=fruitfly_protein, scoring_matrix=scoring_matrix, global_flag=False)
score, align_human, align_fruitfly = compute_local_alignment(seq_x=human_protein, seq_y=fruitfly_protein, scoring_matrix=scoring_matrix, alignment_matrix=alignment_matrix)
return score, align_human, align_fruitfly
def get_edit_distance(seq_x, seq_y, scoring_matrix):
'''
compute the seq_x and seq_y global alignment with scoring matrix
return the edit distance can be expressed in term of:
|x| + |y| - score(x, y)
'''
alignment_matrix = compute_alignment_matrix(seq_x, seq_y, scoring_matrix,
True)
score, align_x, align_y = compute_global_alignment(seq_x, seq_y,
scoring_matrix,
alignment_matrix)
return len(seq_x) + len(seq_y) - score
def protein_alignment(scoring_matrix):
'''
compute the human eyeless protein and fruitfly eyeless portein alignment
'''
human_protein = read_protein('HumanEyelessProtein.txt')
fruitfly_protein = read_protein('FruitflyEyelessprotein.txt')
alignment_matrix = compute_alignment_matrix(seq_x=human_protein,
seq_y=fruitfly_protein,
scoring_matrix=scoring_matrix,
global_flag=False)
score, align_human, align_fruitfly = compute_local_alignment(
seq_x=human_protein,
seq_y=fruitfly_protein,
scoring_matrix=scoring_matrix,
alignment_matrix=alignment_matrix)
return score, align_human, align_fruitfly
def consensus_alignment(scoring_matrix, align_sequence):
'''
compute the similarity of the two sequences in the local alignment computed to
the PAX Domain.
global alignment of local align_x vs PAX Domain, local align_y vs PAX Domain
return the alignment and the persentage of elements in these two sequences that agree
'''
pax_domain = read_protein('ConsensusPAXDomain.txt')
align_sequence = align_sequence.replace("-", "")
alignment_matrix = compute_alignment_matrix(seq_x=align_sequence, seq_y=pax_domain, scoring_matrix=scoring_matrix, global_flag=True)
score, global_align_sequence, global_align_pax = compute_global_alignment(seq_x=align_sequence, seq_y=pax_domain, scoring_matrix=scoring_matrix, alignment_matrix=alignment_matrix)
seq = difflib.SequenceMatcher(None, global_align_sequence, global_align_pax)
ratio = seq.ratio()
#import pdb; pdb.set_trace()
return global_align_sequence, ratio
