def test_empty_iter(self):
seq1 = []
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq2[0:3] == seq1[0:3]
try:
seq2.next()
self.fail('StopIteration expected')
except StopIteration:
pass
def test_empty_iter(self):
seq1 = []
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq2[0:3] == seq1[0:3]
try:
seq2.next()
self.fail('StopIteration expected')
except StopIteration:
pass
def test_short_iter(self):
seq1 = range(3)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
first = seq2.next()
assert first == 0
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 1
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 2
assert seq2[0:4] == seq1[0:4]
try:
item = seq2.next()
self.fail('StopIteration expected')
except StopIteration:
pass
seq1 = range(2)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
def filter_snvs(self, snvs):
snps = RandomAccessIterator(snvs, rnd_access_win=self.snps_in_window)
rates = _calculate_segregation_rates(snps,
self.pop_type,
self.snps_in_window,
samples=self.samples)
max_zero_dist = self.max_zero_dist_recomb
for snp, chrom, pos, rates in rates:
self.tot_snps += 1
dists, recombs = zip(*[(rate.pos - pos, rate.recomb_rate)
for rate in rates])
if len(dists) < self.min_num_snps:
continue
if self.debug_plot_dir is None:
plot_fhand = None
else:
chrom_dir = pjoin(self.debug_plot_dir, str(chrom))
if not exists(chrom_dir):
mkdir(chrom_dir)
fname = str(chrom) + '_' + str(pos) + '.png'
plot_fhand = open(pjoin(chrom_dir, fname), 'w')
res = _calc_ajusted_recomb(dists,
recombs,
max_recomb=self.max_recomb_curve_fit,
max_zero_dist_recomb=max_zero_dist,
alpha_recomb_0=self.alpha_recomb_0,
plot_fhand=plot_fhand)
self._store_log_info(*res)
if res[1]:
self.passed_snps += 1
yield snp
def test_short_iter(self):
seq1 = range(3)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
first = seq2.next()
assert first == 0
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 1
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 2
assert seq2[0:4] == seq1[0:4]
try:
item = seq2.next()
self.fail('StopIteration expected')
except StopIteration:
pass
seq1 = range(2)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
def _smooth_genotypes(self, snp_ab_genotypes, samples):
big_win = self.window * 2 - 1
snp_ab_genotypes = RandomAccessIterator(snp_ab_genotypes,
rnd_access_win=big_win)
for idx, (snp, ab_genotype) in enumerate(snp_ab_genotypes):
wins = self._create_windows(idx)
for win in wins:
#transpose here
for sample in samples:
#keep memory of the samples already smothed
snps_in_win = snp_ab_genotypes[win[0]:win[1]]
print list(snps_in_win)
#if all samples are smoothed break the win loop
pass
def __call__(self, snvs):
# TODO: Randon acess iterator based on physical distance
# RandomAcessRegionIterator(items, location_getter)
# once we do this we can remove max_num_snps
snvs = RandomAccessIterator(snvs, self._max_num_snps)
half_win = (snvs._rnd_access_win - 1) // 2
half_win_in_bp = (self.window_in_bp - 1) // 2
for idx, snv in enumerate(snvs):
self._clean_filter(snv)
chrom = snv.chrom
pos = snv.pos
start = idx - half_win
if start < 0:
start = 0
end = idx + half_win
snvs_in_win = snvs[start:end]
def snv_is_close(snv2):
if snv2.chrom != chrom:
return False
if abs(snv2.pos - pos) < half_win_in_bp:
return True
else:
return False
close_snvs = filter(snv_is_close, snvs_in_win)
num_snvs = len(close_snvs)
win_len = self.window_in_bp
# The studied window could be smaller than expected if it is
# located at the beginning of the chromosome
dist_from_0 = pos
if dist_from_0 < half_win_in_bp:
win_len -= (half_win_in_bp - dist_from_0)
# The studied window could be smaller than expected if it is
# located at the end of the chromosome
ref_len = self._lengths[chrom]
end = ref_len - 1
len_not_studied_at_end = pos + half_win_in_bp - end
if len_not_studied_at_end > 0:
win_len -= len_not_studied_at_end
freq = num_snvs / win_len
if freq >= self.max_variability:
snv.add_filter(self.name)
yield snv
def _smooth_genotypes_old(self, snp_ab_genotypes, samples):
win = self.window
snp_ab_genotypes = RandomAccessIterator(snp_ab_genotypes,
rnd_access_win=win)
half_win = (win - 1) // 2
for idx, (snp, ab_genotype) in enumerate(snp_ab_genotypes):
chrom = snp.CHROM
start = idx - half_win
if start < 0:
start = 0
end = idx + half_win + 1
snp_gts_in_win = snp_ab_genotypes[start:end]
smoothed_genos = self._smooth(idx - start, snp_gts_in_win, samples)
smoothed_genos = OrderedDict(zip(samples, smoothed_genos))
yield snp, smoothed_genos
def _recode_parent_genotypes(self, samples=None):
get_coding = GetCoding(self.parents_a, self.parents_b)
def mapper(snp):
return snp, get_coding(snp)
win = self.window
snp_and_coding = RandomAccessIterator(imap(mapper, self._reader),
rnd_access_win=win)
offspring = self.offspring
half_win = (win - 1) // 2
for idx, (snp1, coding1) in enumerate(snp_and_coding):
snp1_calls = [snp1.genotype(sample) for sample in offspring]
start = idx - half_win
if start < 0:
start = 0
end = idx + half_win + 1
snp2_idxs = []
for snp2_idx in range(start, end):
try:
snp2_chrom = snp_and_coding[snp2_idx][0].CHROM
except IndexError:
continue
if snp2_chrom == snp1.CHROM:
snp2_idxs.append(snp2_idx)
coding1 = self._deduce_coding(snp_and_coding, snp1_calls,
snp2_idxs)
if coding1 is None:
# We haven't manage to deduce the AB coding for this snp
continue
coding1['.'] = '.'
if samples is None:
calls = snp1.samples
else:
calls = [snp1.genotype(sample) for sample in samples]
recoded = OrderedDict((call.sample, self._map_to_ab(call, coding1))
for call in calls)
yield snp1, recoded
def smooth_genotypes(self, snp_ab_genotypes, samples):
win = self.window
snp_ab_genotypes = RandomAccessIterator(snp_ab_genotypes,
rnd_access_win=win)
half_win = (win - 1) // 2
for idx, (snp, ab_genotype) in enumerate(snp_ab_genotypes):
chrom = snp.CHROM
start = idx - half_win
if start < 0:
start = 0
end = idx + half_win + 1
# remove snps in other chromosomes
snp_gts_in_win = [
snp_gt for snp_gt in snp_ab_genotypes[start:end]
if snp_gt[0].CHROM == chrom
]
smoothed_genos = self._smooth(idx - start, snp_gts_in_win, samples)
smoothed_genos = OrderedDict(zip(samples, smoothed_genos))
yield snp, smoothed_genos
def test_next_items(self):
seq1 = range(10)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
assert seq2[2] == seq1[2]
first = seq2.next()
assert first == 0
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 1
assert seq2[0:5] == seq1[0:5]
item = seq2.next()
assert item == 2
assert seq2[0:6] == seq1[0:6]
item = seq2.next()
assert item == 3
assert seq2[0:7] == seq1[0:7]
try:
seq2[0:8]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 4
assert seq2[1:8] == seq1[1:8]
assert seq2[1] == seq1[1]
try:
seq2[0]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 5
assert seq2[2:9] == seq1[2:9]
try:
seq2[1:8]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 6
assert seq2[3:10] == seq1[3:10]
item = seq2.next()
assert item == 7
assert seq2[3:10] == seq1[3:10]
item = seq2.next()
assert item == 8
item = seq2.next()
assert item == 9
assert seq2[3:10] == seq1[3:10]
def filter_snvs_by_ld(snvs,
samples=None,
r_sqr=DEF_R_SQR_THRESHOLD,
p_val=DEF_P_VAL,
bonferroni=True,
snv_win=DEF_SNV_WIN,
min_phys_dist=MIN_PHYS_DIST,
log_fhand=None):
if not snv_win % 2:
msg = 'The window should have an odd number of snvs'
raise ValueError(msg)
half_win = (snv_win - 1) // 2
if bonferroni:
p_val /= (snv_win - 1)
snvs = RandomAccessIterator(snvs, rnd_access_win=snv_win)
linked_snvs = set()
total_snvs = 0
passed_snvs = 0
prev_chrom = None
stats_cache = _LDStatsCache()
for snv_i, snv in enumerate(snvs):
total_snvs += 1
if snv_i in linked_snvs:
yield snv
passed_snvs += 1
linked_snvs.remove(snv_i)
continue
linked = None
win_start = snv_i - half_win
this_chrom = snv.chrom
if prev_chrom is None:
prev_chrom = this_chrom
if prev_chrom != this_chrom:
stats_cache = _LDStatsCache()
if win_start < 0:
win_start = 0
for snv_j in range(snv_i + half_win, win_start - 1, -1):
try:
snv_2 = snvs[snv_j]
except IndexError:
continue
if snv_i == snv_j:
continue
try:
linked = stats_cache.get_stat(snv_i, snv_j)
in_cache = True
except KeyError:
in_cache = False
if in_cache:
pass
elif snv.chrom != snv_2.chrom:
# different chroms, they're not linked
linked = False
elif abs(snv.pos - snv_2.pos) < min_phys_dist:
# Too close, they could be errors due to the same reads
# so no independent errors
linked = None
else:
stats = calculate_ld_stats(snv, snv_2, samples=samples)
if stats.r_sqr >= r_sqr and stats.fisher < p_val:
linked = True
if snv_j > snv_i:
linked_snvs.add(snv_j)
break
else:
linked = False
if not linked:
stats_cache.set_stat(snv_i, snv_j, linked)
if linked:
yield snv
passed_snvs += 1
stats_cache.del_lower_than(win_start)
if log_fhand is not None:
_write_log(log_fhand, total_snvs, passed_snvs)
def test_next_items(self):
seq1 = range(10)
seq2 = RandomAccessIterator(iter(seq1), 7)
assert seq1[0] == seq1[0]
assert seq2[0:3] == seq1[0:3]
assert seq2[:3] == seq1[:3]
assert seq2[2] == seq1[2]
first = seq2.next()
assert first == 0
assert seq2[0:4] == seq1[0:4]
item = seq2.next()
assert item == 1
assert seq2[0:5] == seq1[0:5]
item = seq2.next()
assert item == 2
assert seq2[0:6] == seq1[0:6]
item = seq2.next()
assert item == 3
assert seq2[0:7] == seq1[0:7]
try:
seq2[0:8]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 4
assert seq2[1:8] == seq1[1:8]
assert seq2[1] == seq1[1]
try:
seq2[0]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 5
assert seq2[2:9] == seq1[2:9]
try:
seq2[1:8]
self.fail('IndexError expexted')
except IndexError:
pass
item = seq2.next()
assert item == 6
assert seq2[3:10] == seq1[3:10]
item = seq2.next()
assert item == 7
assert seq2[3:10] == seq1[3:10]
item = seq2.next()
assert item == 8
item = seq2.next()
assert item == 9
assert seq2[3:10] == seq1[3:10]
def test_iter_access(self):
seq1 = range(100)
seq2 = RandomAccessIterator(iter(seq1), 11)
assert list(seq1) == list(seq2)