def filter(self, input_filepath, output_filepath):
self.load_loci(input_filepath)
self.filtered_loci = {}
for locus_name, locus in iteritems(self.loci):
locus = self.filter_locus(locus)
if locus is not False:
self.filtered_loci[locus_name] = locus
self.write_output(output_filepath)
def filter(self, input_filepath, output_filepath):
self.load_loci(input_filepath)
self.filtered_loci = {}
for locus_name, locus in iteritems(self.loci):
locus = self.filter_locus(locus)
if locus is not False:
self.filtered_loci[locus_name] = locus
self.write_output(output_filepath)
def generate_value_list(self, data):
values = []
# Iterate through the data and only record values
# for k-repeats that have some support
for k, count in iteritems(data):
if count > 0:
values += [k] * count
return sorted(values)
def __normalized_subset(self, subset, data):
total = self.get_support(subset)
normalized = {}
for k, count in iteritems(data):
if total == 0:
normalized[k] = 0.0
else:
normalized[k] = (1.0 * count) / total
return normalized
def generate_value_list(self,data):
values = []
# Iterate through the data and only record values
# for k-repeats that have some support
for k, count in iteritems(data):
if count > 0:
values += [k] * count
return sorted(values)
def __normalized_subset(self, subset, data):
total = self.get_support(subset)
normalized = {}
for k, count in iteritems(data):
if total == 0:
normalized[k] = 0.0
else:
normalized[k] = (1.0 * count) / total
return normalized
def write_output(self, output_filepath):
fileout = open(output_filepath, 'w')
header = ['Locus', 'Repeats', 'Normal', 'Tumor']
fileout.write('\t'.join(header) + '\n')
for l, locus in sorted(iteritems(self.filtered_loci)):
output = locus.generate_output()
for line in output:
fileout.write('\t'.join(line) + '\n')
fileout.close()
return True
def subset_outlier_filter(self, data, sds):
values = self.generate_value_list(data)
output = {}
if len(values):
mean = numpy.mean(values)
std = numpy.std(values)
# Round the min and max for the window to allow for
# some leniency in the filter.
min_k = int(math.floor(mean - (sds * std)))
max_k = int(math.ceil(mean + (sds * std)))
for k, count in iteritems(data):
if min_k <= k <= max_k:
# Acceptable
output[k] = count
return output
def subset_outlier_filter(self, data, sds):
values = self.generate_value_list(data)
output = {}
if len(values):
mean = numpy.mean(values)
std = numpy.std(values)
# Round the min and max for the window to allow for
# some leniency in the filter.
min_k = int(math.floor(mean - (sds * std)))
max_k = int(math.ceil(mean + (sds * std)))
for k, count in iteritems(data):
if min_k <= k <= max_k:
# Acceptable
output[k] = count
return output
def k_values(self, subset=False):
k_values = set()
if subset is False:
for k in self.__k:
if (k in self.__normal and self.__normal[k] > 0.0) or \
(k in self.__tumor and self.__tumor[k] > 0.0):
k_values.add(k)
else:
if subset.upper()[0] == 'N':
# Normal data set
subset = self.__normal
elif subset.upper()[0] == 'T':
# Tumor data set
subset = self.__tumor
for k, v in iteritems(subset):
if v > 0.0:
k_values.add(k)
k_values = sorted(k_values)
return k_values
def k_values(self, subset = False):
k_values = set()
if subset is False:
for k in self.__k:
if (k in self.__normal and self.__normal[k] > 0.0) or \
(k in self.__tumor and self.__tumor[k] > 0.0):
k_values.add(k)
else:
if subset.upper()[0] == 'N':
# Normal data set
subset = self.__normal
elif subset.upper()[0] == 'T':
# Tumor data set
subset = self.__tumor
for k, v in iteritems(subset):
if v > 0.0:
k_values.add(k)
k_values = sorted(k_values)
return k_values
def expand_kmer_counts(d):
new_list = []
for k, v in iteritems(d):
new_list.extend([k] * v)
return new_list
output_filepath = os.path.abspath(args.output)
status_filepath = output_filepath + '.status'
loci = load_loci(input_filepath)
fileout = open(output_filepath, 'w')
line = '\t'.join([
'Locus', 'Normal_Reads', 'Tumor_Reads', 'Difference', 'Distance',
'Dissimilarity'
])
fileout.write(line + '\n')
# Iterate through all the results to generate the output. As part of the
# loop, count the weighted values for each metric.
values = {'difference': [], 'distance': [], 'dissimilarity': []}
for l, locus in sorted(iteritems(loci)):
# Calculate post-normalization metrics
locus.normalize()
difference = Difference.get(locus)
distance = EuclideanDistance.get(locus)
dissimilarity = CosineDissimilarity.get(locus)
# Generate output line.
line = '\t'.join([
str(x) for x in [
locus.locus(),
locus.get_support('N'),
locus.get_support('T'),
round(difference, 4),
round(distance, 4),
round(dissimilarity, 4)
def expand_kmer_counts(d):
new_list = []
for k, v in iteritems(d):
new_list.extend([k] * v)
return new_list
