def test_multiprocessing(self):
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=1)
dvec_multi = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=2)
self.assertTrue(np.all(dvec == dvec_multi))
def test_multiprocessing_more(self):
dvec_multi = pwsd.apply_pairwise_sq(mixed_seqs,
pwsd.metrics.nw_metric,
matrix='blosum62',
ncpus=2)
dvec = pwsd.apply_pairwise_sq(mixed_seqs,
pwsd.metrics.nw_metric,
matrix='blosum62',
ncpus=1)
self.assertTrue(np.all(dvec == dvec_multi))
def test_pw_sq_nonuniq(self):
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=1)
dmat = squareform(dvec)
dvec2 = pwsd.apply_pairwise_sq(seqs[:10] + seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=1)
dmat2 = squareform(dvec2)
self.assertTrue(np.all(dmat2[:10, :][:, :10] == dmat))
def test_nb_pw_sq_hamming():
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=1)
dvec_nb = pwsd.numba_tools.nb_pairwise_sq(
seqs[:10], pwsd.numba_tools.nb_hamming_distance)
assert (np.all(dvec == dvec_nb))
def test_pw_sq_subst(self):
subst_dict = pwsd.matrices.dict_from_matrix(parasail.blosum62)
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.str_subst_metric,
subst_dict=subst_dict,
ncpus=1)
dmat = squareform(dvec)
self.assertTrue(dmat.shape[0] == 10 and dmat.shape[1] == 10)
def test_ex6():
import pwseqdist as pw
from scipy.spatial.distance import squareform
import Levenshtein
dvec = pw.apply_pairwise_sq(seqs=['homer', 'home', 'rome'],
metric=Levenshtein.distance,
ncpus=1)
dmat = squareform(dvec)
def test_nb_pw_sq():
subst_dict = pwsd.matrices.dict_from_matrix(parasail.blosum62)
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.str_subst_metric,
subst_dict=subst_dict,
ncpus=1)
subst_dict = pwsd.numba_tools.nb_dict_from_matrix(parasail.blosum62)
dvec_nb = pwsd.numba_tools.nb_pairwise_sq(seqs[:10],
pwsd.numba_tools.nb_subst_metric,
subst_dict)
assert (np.all(dvec == dvec_nb))
def generate_peptide_data(L=5, n=300, seed=110820):
"""Attempt to generate some random peptide data with a
phenotype enrichment associated with a motif"""
np.random.seed(seed)
alphabet = 'ARNDCQEGHILKMFPSTWYVBZ'
probs = np.random.rand(len(alphabet))
probs = probs / np.sum(probs)
seqs = [''.join(np.random.choice(list(alphabet), size=4, p=probs)) for i in range(n)]
def _assign_trait2(seq):
if seq[1] in 'KRQ' or seq[3] in 'KRQ':
pr = 0.99
elif seq[0] in 'QA':
pr = 0.01
else:
pr = 0.03
return np.random.choice([1, 0], p=[pr, 1-pr])
def _assign_trait1(seq):
d = np.sum([i for i in map(operator.__ne__, seq, seqs[0])])
return {0:'ZERO', 1:'ONE'}[int((d <= 3) * (np.random.rand() < 0.6))]
def _assign_trait3(seq):
return np.random.choice(['A', 'B', 'C'], p=[0.2, 0.4, 0.4])
pw = pwsd.apply_pairwise_sq(seqs, metric=pwsd.metrics.hamming_distance)
Z = sch.linkage(pw, method='complete')
labels = sch.fcluster(Z, 50, criterion='maxclust')
dat = pd.DataFrame({'seq':seqs,
'trait1':np.array([_assign_trait1(p) for p in seqs]),
'trait2':np.array([_assign_trait2(p) for p in seqs]),
'trait3':np.array([_assign_trait3(p) for p in seqs]),
'cluster':labels,
'count':np.random.randint(4, 10, size=n)})
return dat, pw
def test_README_example1():
import numpy as np
import pwseqdist as pw
import multiprocessing
from scipy.spatial.distance import squareform
peptides = [
'CACADLGAYPDKLIF', 'CACDALLAYTDKLIF', 'CACDAVGDTLDKLIF',
'CACDDVTEVEGDKLIF', 'CACDFISPSNWGIQSGRNTDKLIF', 'CACDPVLGDTRLTDKLIF'
]
dvec = pw.apply_pairwise_sq(seqs=peptides,
metric=pw.metrics.nw_hamming_metric,
ncpus=multiprocessing.cpu_count())
dmat = squareform(dvec).astype(int)
exp = np.array([[0, 4, 6, 7, 15, 8], [4, 0, 5, 7, 14, 7],
[6, 5, 0, 6, 14, 4], [7, 7, 6, 0, 14, 8],
[15, 14, 14, 14, 0, 11], [8, 7, 4, 8, 11, 0]])
assert np.all(dmat == exp)
def test_pw_sq(self):
dvec = pwsd.apply_pairwise_sq(seqs[:10],
pwsd.metrics.hamming_distance,
ncpus=1)
dmat = squareform(dvec)
self.assertTrue(dmat.shape[0] == 10 and dmat.shape[1] == 10)
def test_pw_sq_nonuniq_tcrdist(self):
tmp = [
'PNSSL', 'KEKRN', 'KEKRN', 'PNASF', 'PNASF', 'PNASF', 'EKKES',
'EKKER', 'IRTEH'
]
res = np.array([[
0,
5,
5,
2,
2,
2,
5,
5,
5,
], [
5,
0,
0,
5,
5,
5,
4,
4,
5,
], [
5,
0,
0,
5,
5,
5,
4,
4,
5,
], [
2,
5,
5,
0,
0,
0,
5,
5,
5,
], [
2,
5,
5,
0,
0,
0,
5,
5,
5,
], [
2,
5,
5,
0,
0,
0,
5,
5,
5,
], [
5,
4,
4,
5,
5,
5,
0,
1,
4,
], [
5,
4,
4,
5,
5,
5,
1,
0,
4,
], [
5,
5,
5,
5,
5,
5,
4,
4,
0,
]])
dvec = pwsd.apply_pairwise_sq(tmp,
pwsd.metrics.nw_hamming_metric,
ncpus=1)
dmat = squareform(dvec).astype(int)
#print(dmat)
#print(res)
#print(tmp[0], tmp[3], res[0, 3], dmat[0, 3])
self.assertTrue(np.all(dmat == res))