def _get_adj_list_directional(self, umis, counts):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
adj_list = {umi: [] for umi in umis}
if self.fuzzy_match:
for umi1 in umis:
# we need a second regex for some insertions,
# e.g UMI1 = "ATCG", UMI2 = "ATTC"
comp_regex_err = regex.compile("(%s){e<=1}" % str(umi1))
comp_regex_del = regex.compile("(%s){i<=1}" % str(umi1)[::-1])
for umi2 in umis:
if umi1 == umi2:
continue
if counts[umi1] >= (counts[umi2] * self.dir_threshold):
if (max(len(umi1), len(umi2)) -
min(len(umi1), len(umi2))) > 1:
continue
if (comp_regex_err.match(str(umi2))
or comp_regex_del.match(str(umi2))):
adj_list[umi1].append(umi2)
else:
for umi1, umi2 in itertools.combinations(umis, 2):
if edit_distance(umi1, umi2) <= 1:
if counts[umi1] >= (counts[umi2] * 2) - 1:
adj_list[umi1].append(umi2)
if counts[umi2] >= (counts[umi1] * 2) - 1:
adj_list[umi2].append(umi1)
return adj_list
def get_average_umi_distance(umis):
if len(umis) == 1:
return -1
dists = [edit_distance(x, y) for x, y in itertools.combinations(umis, 2)]
return float(sum(dists)) / (len(dists))
def _get_adj_list_adjacency(self, umis, counts, threshold):
''' identify all umis within hamming distance threshold'''
return {umi: [umi2 for umi2 in umis if
edit_distance(umi.encode('utf-8'),
umi2.encode('utf-8')) <= threshold]
for umi in umis}
def getErrorCorrectMapping(cell_barcodes, whitelist, threshold=1):
''' Find the mappings between true and false cell barcodes based
on an edit distance threshold.
Any cell barcode within the threshold to more than one whitelist
barcode will be excluded'''
true_to_false = collections.defaultdict(set)
whitelist = set([str(x).encode("utf-8") for x in whitelist])
for cell_barcode in cell_barcodes:
match = None
barcode_in_bytes = str(cell_barcode).encode("utf-8")
for white_cell in whitelist:
if barcode_in_bytes in whitelist: # don't check if whitelisted
continue
if edit_distance(barcode_in_bytes, white_cell) <= threshold:
if match is not None: # already matched one barcode
match = None # set match back to None
break # break and don't add to maps
else:
match = white_cell.decode("utf-8")
if match is not None:
true_to_false[match].add(cell_barcode)
return true_to_false
def _get_adj_list_adjacency(self, umis, counts, threshold):
''' identify all umis within hamming distance threshold'''
return {umi: [umi2 for umi2 in umis if
edit_distance(umi.encode('utf-8'),
umi2.encode('utf-8')) <= threshold]
for umi in umis}
def _get_adj_list_directional(self, umis, counts):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
adj_list = {umi: [] for umi in umis}
if self.fuzzy_match:
for umi1 in umis:
# we need a second regex for some insertions,
# e.g UMI1 = "ATCG", UMI2 = "ATTC"
comp_regex_err = regex.compile("(%s){e<=1}" % str(umi1))
comp_regex_del = regex.compile("(%s){i<=1}" % str(umi1)[::-1])
for umi2 in umis:
if umi1 == umi2:
continue
if counts[umi1] >= (counts[umi2]*self.dir_threshold):
if (max(len(umi1), len(umi2)) -
min(len(umi1), len(umi2))) > 1:
continue
if (comp_regex_err.match(str(umi2)) or
comp_regex_del.match(str(umi2))):
adj_list[umi1].append(umi2)
else:
for umi1, umi2 in itertools.combinations(umis, 2):
if edit_distance(umi1, umi2) <= 1:
if counts[umi1] >= (counts[umi2]*2)-1:
adj_list[umi1].append(umi2)
if counts[umi2] >= (counts[umi1]*2)-1:
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_directional_adjacency(self, umis, counts, threshold):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
return {umi: [umi2 for umi2 in umis if
edit_distance(umi.encode('utf-8'),
umi2.encode('utf-8')) == 1 and
counts[umi] >= (counts[umi2]*2)-1] for umi in umis}
def get_average_umi_distance(umis):
if len(umis) == 1:
return -1
dists = [edit_distance(x.encode('utf-8'), y.encode('utf-8')) for
x, y in itertools.combinations(umis, 2)]
return float(sum(dists))/(len(dists))
def _get_adj_list_directional(self, umis, counts, threshold=1):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
return {umi: [umi2 for umi2 in umis if
edit_distance(umi.encode('utf-8'),
umi2.encode('utf-8')) == threshold and
counts[umi] >= (counts[umi2]*2)-1] for umi in umis}
def _get_adj_list_adjacency(self, umis, counts, threshold):
''' identify all umis within hamming distance threshold'''
adj_list = {umi: [] for umi in umis}
for umi1, umi2 in itertools.combinations(umis, 2):
if edit_distance(umi1, umi2) <= threshold:
adj_list[umi1].append(umi2)
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_directional(self, umis, counts, threshold=1):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
adj_list = {umi: [] for umi in umis}
for umi1, umi2 in itertools.combinations(umis, 2):
if edit_distance(umi1, umi2) <= threshold:
if counts[umi1] >= (counts[umi2] * 2) - 1:
adj_list[umi1].append(umi2)
if counts[umi2] >= (counts[umi1] * 2) - 1:
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_adjacency(self, umis, counts, threshold):
''' identify all umis within hamming distance threshold'''
adj_list = {umi: [] for umi in umis}
if len(umis) > 25:
umi_length = len(umis[0])
substr_idx = build_substr_idx(umis, umi_length, threshold)
iter_umi_pairs = iter_nearest_neighbours(umis, substr_idx)
else:
iter_umi_pairs = itertools.combinations(umis, 2)
for umi1, umi2 in iter_umi_pairs:
if edit_distance(umi1, umi2) <= threshold:
adj_list[umi1].append(umi2)
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_adjacency(self, umis, counts, threshold):
''' identify all umis within hamming distance threshold'''
adj_list = {umi: [] for umi in umis}
if len(umis) > 25:
umi_length = len(umis[0])
substr_idx = build_substr_idx(umis, umi_length, threshold)
iter_umi_pairs = iter_nearest_neighbours(umis, substr_idx)
else:
iter_umi_pairs = itertools.combinations(umis, 2)
for umi1, umi2 in iter_umi_pairs:
if edit_distance(umi1, umi2) <= threshold:
adj_list[umi1].append(umi2)
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_directional(self, umis, counts, threshold=1):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
adj_list = {umi: [] for umi in umis}
if len(umis) > 25:
umi_length = len(umis[0])
substr_idx = build_substr_idx(umis, umi_length, threshold)
iter_umi_pairs = iter_nearest_neighbours(umis, substr_idx)
else:
iter_umi_pairs = itertools.combinations(umis, 2)
for umi1, umi2 in iter_umi_pairs:
if edit_distance(umi1, umi2) <= threshold:
if counts[umi1] >= (counts[umi2]*2)-1:
adj_list[umi1].append(umi2)
if counts[umi2] >= (counts[umi1]*2)-1:
adj_list[umi2].append(umi1)
return adj_list
def _get_adj_list_directional(self, umis, counts, threshold=1):
''' identify all umis within the hamming distance threshold
and where the counts of the first umi is > (2 * second umi counts)-1'''
adj_list = {umi: [] for umi in umis}
if len(umis) > 25:
umi_length = len(umis[0])
substr_idx = build_substr_idx(umis, umi_length, threshold)
iter_umi_pairs = iter_nearest_neighbours(umis, substr_idx)
else:
iter_umi_pairs = itertools.combinations(umis, 2)
for umi1, umi2 in iter_umi_pairs:
if edit_distance(umi1, umi2) <= threshold:
if counts[umi1] >= (counts[umi2]*2)-1:
adj_list[umi1].append(umi2)
if counts[umi2] >= (counts[umi1]*2)-1:
adj_list[umi2].append(umi1)
return adj_list
