def get_clusters_from_seqlist(seqlist, dist_threshold=0.05):
"""Cluster a list of sequences by a distance identity threshold
Parameters
----------
seqlist : list
list of sequences as str
dist_threshold : float
Max distance value to retain, branches above this length in the
hierarchical clustering tree will be cut.
Returns
-------
list
list of lists - input sequences now grouped by cluster
list
list of int - cluster memberships of the originally input list
"""
if len(seqlist) == 1:
# Skip alignment if there is only one sequence
return([seqlist], [0])
else:
aligner = PairwiseAligner()
aligner.mode = "local"
# Convert sequence list to distance matrix
distmatrix = []
for seq1 in seqlist:
row = []
for seq2 in seqlist:
maxlen = max([len(seq1), len(seq2)])
# Take percentage identity of pairwise alignment score (match base
# +1, all other operations +0) over the longer sequence in pair
idval = aligner.align(seq1, seq2).score / maxlen
distval = 1 - idval # convert to distance fraction
row.append(distval)
distmatrix.append(row)
# Hierarchical clustering from the distance matrix
htree = treecluster(data=None, distancematrix=array(distmatrix))
# Find number of branches with length longer than threshold, and add 1
# to get number of cuts
cuts = 1 + len([htree[i].distance for i in range(len(htree))
if htree[i].distance > dist_threshold])
clust_ids = list(htree.cut(cuts))
clust_seqs_dict = defaultdict(list)
for i in range(len(seqlist)):
clust_seqs_dict[clust_ids[i]] += [seqlist[i]]
# Convert dict of lists to list of lists
clust_seqs = [clust_seqs_dict[i] for i in clust_seqs_dict]
return(clust_seqs, clust_ids)
def nw_bio_mat(seq1, seq2, cost_mat, key):
aligner = PairwiseAligner(alphabet=key)
matrix = {}
for i in range(len(key)):
for j in range(0, len(key)):
matrix[(key[i], key[j])] = cost_mat[i * len(key) + j]
aligner.substitution_matrix = substitution_matrices.Array(data=matrix)
aligner.gap_score = cost_mat[len(key)**2]
alignments = aligner.align(seq1, seq2)
formated_alignments = []
for i in range(len(alignments)):
als = str(alignments[i]).split("\n")
formated_alignments.append([als[0], als[2], int(alignments[i].score)])
return formated_alignments
def make_aligner() -> PairwiseAligner:
aligner = PairwiseAligner(match_score=MATCH_SCORE,
mismatch_score=MISMATCH_SCORE,
end_open_gap_score=END_GAP_PENALTY,
end_extend_gap_score=END_GAP_EXTEND_PENALTY,
internal_open_gap_score=GAP_PENALTY,
internal_extend_gap_score=GAP_EXTEND_PENALTY)
return aligner
def bioPython_default_local_aligner(a, b):
aligner = PairwiseAligner()
aligner.mode = 'local'
aligner.match_score = 2
aligner.mismatch_score = -3
aligner.open_gap_score = -7
aligner.extend_gap_score = -2
sequence1 = SeqIO.read('./resource/fasta' + str(a) + '.fasta', 'fasta')
sequence2 = SeqIO.read('./resource/fasta' + str(b) + '.fasta', 'fasta')
alignments = aligner.align(sequence1.seq, sequence2.seq)
def setUp(self):
aligner = PairwiseAligner()
aligner.internal_open_gap_score = -1
aligner.internal_extend_gap_score = -0.0
aligner.match_score = +1
aligner.mismatch_score = -1
aligner.mode = "local"
self.aligner = aligner
def create_aligner() -> PairwiseAligner:
"""
Creates an aligner that can be used to search for proteins.
"""
aligner = PairwiseAligner(mode="local")
# By default we want matches and penalize mismatches.
aligner.mismatch_score = -1
aligner.match_score = 1
# left or right gaps shouldn't count negatively due to the local search.
aligner.query_left_gap_score = 0
aligner.query_right_gap_score = 0
aligner.target_right_gap_score = 0
aligner.target_left_gap_score = 0
# Gaps in the middle should count negatively to narrow down the search space.
aligner.query_internal_gap_score = -1
aligner.target_internal_gap_score = -1
return aligner
def pairwise(self, potential_parent):
"""
Парное выравнивание последовательности листа на потенциальных родителей и сохранение скора
Args:
potential_parent (str): последовательность потенциального родителя
"""
aligner = PairwiseAligner()
# используем локальное выравнивание
aligner.mode = 'local'
# заменим дефолтные символы пропуска на другие
seq = self.seq.replace('-', '')
potential_parent_undersores = potential_parent.replace('-', '')
# выравнивание
score = aligner.score(seq, potential_parent_undersores)
# сохраняем скор
self.parent_scores[potential_parent] = score
def nw_bio(seq1, seq2, cost_table):
aligner = PairwiseAligner(alphabet=list(set(seq1 + seq2)))
aligner.match_score = cost_table[0]
aligner.mismatch_score = cost_table[1]
aligner.gap_score = cost_table[2]
alignments = aligner.align(seq1, seq2)
formated_alignments = []
for i in range(len(alignments)):
als = str(alignments[i]).split("\n")
formated_alignments.append([als[0], als[2], int(alignments[i].score)])
return formated_alignments
def _remove_missing_res(self, record: SeqRecord, pdb: Path):
structure = PDBParser().get_structure(record.id, pdb)
sequence = ''.join([
str(_.get_sequence())
for _ in CaPPBuilder().build_peptides(structure, aa_only=False)
])
path = PairwiseAligner().align(record.seq.ungap('-'), sequence)[0].path
gaps = []
for i, _ in enumerate(path[:-1]):
if path[i][1] == path[i + 1][1]:
gaps.append((path[i][0], path[i + 1][0]))
gaps = list(reversed(gaps))
mut = record.seq.tomutable()
for gap in gaps:
i = 0
for k, res in enumerate(mut):
if res == '-':
continue
if gap[0] <= i < gap[1]:
mut[k] = '-'
i += 1
record.seq = mut.toseq()
return record
def perform_randomized_tests(n=1000):
"""Perform randomized tests and compare to pslMap.
Run this function to perform 8 x n mappings for alignments of randomly
generated sequences, get the alignment in PSL format, and compare the
result to that of pslMap.
"""
aligner = PairwiseAligner()
aligner.internal_open_gap_score = -1
aligner.internal_extend_gap_score = -0.0
aligner.match_score = +1
aligner.mismatch_score = -1
aligner.mode = "local"
for i in range(n):
nBlocks1 = random.randint(1, 10)
nBlocks2 = random.randint(1, 10)
test_random(aligner, nBlocks1, nBlocks2, "+", "+")
test_random(aligner, nBlocks1, nBlocks2, "+", "-")
test_random(aligner, nBlocks1, nBlocks2, "-", "+")
test_random(aligner, nBlocks1, nBlocks2, "-", "-")
test_random_sequences("+", "+")
test_random_sequences("+", "-")
test_random_sequences("-", "+")
test_random_sequences("-", "-")
type=str,
required=True)
parser.add_argument('-r',
'--reference',
help='Reference to be aligned to',
type=str,
required=True)
parser.add_argument('-n',
'--seq_name',
help='Name of the aligned sequence',
type=str,
required=True)
args = parser.parse_args()
aligner = PairwiseAligner()
aligner.mode = 'global'
aligner.match_score = 1
aligner.mismatch_score = 0
aligner.open_gap_score = -2
aligner.extend_gap_score = -1
ref = SeqIO.read(args.reference, "fasta")
ref.seq = str(ref.seq.upper()).replace('-', 'N')
cons = SeqIO.read(args.infile, "fasta")
aln = aligner.align(ref.seq, cons.seq)
with open(args.outfile, 'w') as out:
print(">", args.seq_name, file=out)
print(str(aln[0]).strip().split('\n')[2], file=out)
try:
GAP_PENALTY = scores_dict['gap penalty']
GAP_EXTEND_PENALTY = scores_dict['gap extend penalty']
END_GAP_PENALTY = scores_dict['end gap penalty']
END_GAP_EXTEND_PENALTY = scores_dict['end gap extend penalty']
MATCH_SCORE = scores_dict['match score']
MISMATCH_SCORE = scores_dict['mismatch score']
except KeyError as ex:
raise ValueError(f"'{ex.args[0]}' is missing in data/scores.tab") from ex
score_matrix = np.empty((16, 16))
for i, j in itertools.product(range(0, 16), range(0, 16)):
score_matrix[i, j] = MATCH_SCORE if i & j else MISMATCH_SCORE
aligner = PairwiseAligner(substitution_matrix=score_matrix, end_open_gap_score=END_GAP_PENALTY,
end_extend_gap_score=END_GAP_EXTEND_PENALTY, internal_open_gap_score=GAP_PENALTY, internal_extend_gap_score=GAP_EXTEND_PENALTY)
seq_read_dict = {
'-':0,
'A':1,
'C':2,
'G':4,
'T':8,
'R':5,
'Y':10,
'S':6,
'W':9,
'K':12,
'M':3,
'B':14,
'D':13,
# Add parse arguments
parser = argparse.ArgumentParser(description='Computes a pairwise similarity matrix from a fasta file.')
parser.add_argument('-f',help='name of the fasta file',required=True)
parser.add_argument('-s',help='name of subsitution matrix from BioPython',required=True,choices=MatrixInfo.available_matrices)
parser.add_argument('-go',help='gap opening score',type=float,required=True)
parser.add_argument('-ge',help='gap extension score',type=float,required=True)
args = parser.parse_args()
# Parse fasta file #
seqs = list(SeqIO.parse(args.f,'fasta'))
# Get substitution matrix
substitution_matrix = getattr(MatrixInfo,args.s)
#Pairwise alignment
aligner = PairwiseAligner()
aligner.open_gap_score, aligner.extend_gap_score = args.go, args.ge
aligner.substitution_matrix = substitution_matrix
# Align sequences and build matrix
def similarity_matrix(seqs,n=len(seqs)):
similarity_matrix = np.zeros([n,n])
for i in range(len(seqs)):
for j in range(len(seqs)):
alignment = aligner.align(seqs[i].seq,seqs[j].seq)
similarity_matrix[i][j] = alignment.score
return similarity_matrix
m = similarity_matrix(seqs)
def print_matrix(m):
args = parser.parse_args()
for file in (args.query_seq, args.target_seq):
if not path.isfile(file):
parser.error("File %s doesn't exist" % file)
# alphabet
if args.seq_type == 'dna':
args.seq_abc = IUPACAmbiguousDNA()
elif args.seq_type == 'rna':
args.seq_abc = IUPACAmbiguousRNA()
else:
args.seq_abc = ExtendedIUPACProtein()
# Aligners setup
aligners = {'global': PairwiseAligner(), 'local': None}
aligners['global'].mode = 'global'
if args.seq_type in ('dna', 'rna'):
aligners['global'].match = args.match_score
aligners['global'].mismatch = args.mismatch_score
if not args.open_gap_score:
args.open_gap_score = -5
if not args.extend_gap_score:
args.extend_gap_score = -2
else:
sub_matrix = getattr(import_module('Bio.SubsMat.MatrixInfo'),
args.sub_matrix)
aligners['global'].substitution_matrix = sub_matrix
if not args.open_gap_score:
args.open_gap_score = -11
if not args.extend_gap_score:
# coding: utf-8
# # DEMO: Izračun lastne matrike in njena uporaba
#
# V spodnjem primeru je prikazan izračun lastne matrike (enako, kot pri prejšnji vaji - [VAJA: Izračun matrike zamenjav (Python)](matrika_zamenjav.ipynb)) ter uporaba tako izračunane matrike za poravnavo dveh zaporedij.
#
# ## Izračun matrike
# In[1]:
from Bio import SeqIO
sequence1 = SeqIO.read('vhod/matrika_zamenjav-myoglobin_horse.fasta', 'fasta')
sequence2 = SeqIO.read('vhod/matrika_zamenjav-myoglobin_rat.fasta', 'fasta')
# v mapi vhod sta tudi zaporedji mišjega in človeškega nebulina, ki sta bistveno daljši
from Bio.Align import PairwiseAligner
aligner = PairwiseAligner()
aligner.mode = 'local'
aligner.match_score = 2
aligner.mismatch_score = -3
aligner.open_gap_score = -7
aligner.extend_gap_score = -2
alignments = aligner.align(sequence1.seq, sequence2.seq)
alignment = alignments[0]
from Bio.Align.substitution_matrices import Array
frequency = Array('ACGT', dims=2)
for (start1, end1), (start2, end2) in zip(*alignment.aligned):
seq1 = sequence1[start1:end1]
seq2 = sequence2[start2:end2]
for c1, c2 in zip(seq1, seq2):
frequency[c1, c2] += 1
import numpy
