def plot_matrix(self, matrix):
fig, ax = plt.subplots(1, 1)
ax.plot(range(len(matrix)), np.sort(matrix))
ax.set(xlabel="Genes", ylabel="log2(TPM+1)")
fig.savefig(join_paths([self.images_dir, 'gene_count_distribution.png']))
fig, ax = plt.subplots(1,1)
ax.hist(matrix)
fig.savefig(join_paths([self.images_dir, 'gene_count_histogram.png']))
def __init__(self,
dir_path_10x=BasePaths.Pilot19,
cache_directory=BasePaths.Cache):
super(UploadMatExpressionMatrix,
self).__init__(cache_dir=cache_directory)
self.dir_path_10x = dir_path_10x
self.cache_directory = cache_directory
self.expression_matrix_path = join_paths(
[self.dir_path_10x, 'matrix.mtx'])
self.cell_matrix_path = join_paths([self.dir_path_10x, 'barcodes.tsv'])
self.genes_matrix_path = join_paths([self.dir_path_10x, 'genes.tsv'])
self.composing_items.append(self.expression_matrix_path)
self.composing_items.append(self.cell_matrix_path)
def plot_distance_distribution(self, distances, name=''):
all_distances = distances.values
upper_distances = np.triu(all_distances, k=0)
upper_distances = upper_distances.reshape(-1)
upper_distances = upper_distances[upper_distances > 0]
plt.subplot()
sns.distplot(upper_distances)
plt.savefig(join_paths([self.outfile_path, name + 'distance_distribution.png']))
def __init__(self, filter_by_cell_line=True):
self.expression_matrix = load_if_cached(
join_paths([BasePaths.ExpressionProcessed]),
pre_process_expression_matrix)
if filter_by_cell_line:
self.expression_matrix, self.cell_lines_matrix, self.cell_lines_set = filter_expression_by_cell_line(
self.expression_matrix)
self.separated_expressions = {}
for cell_line in self.cell_lines_set:
self.separated_expressions[
cell_line] = self.CellLineExpression(
self.expression_matrix, self.cell_lines_matrix,
cell_line)
def transform(self, expression_object, *args, **kwargs):
try:
if self.DEBUG and not self.ignore_debug:
raise IOError
with open(self.out_file_path(expression_object), 'rb') as in_file:
ret = pickle.load(in_file)
if ret.composing_items[-1] == self and ret.composing_items[:-1] == expression_object.composing_items:
return ret
else:
logging.log("Wrong code version of transformer, running transformer again")
raise Exception
except:
logging.info("Running transformer {}".format(type(self)))
ret = self.transform_aux(expression_object, args, kwargs)
ret.composing_items.append(self)
with open(join_paths([self.cache_dir, ret.name]), 'wb') as out_file:
pickle.dump(ret, out_file)
logging.info(
"Saved result from transformer {} with parameters {} in {}".format(type(self), self.composing_items,
self.out_file_path(
expression_object)))
return ret
def out_file_path(self, expression_object=None, *args, **kwargs):
if expression_object is None:
return join_paths([self.cache_dir, self.out_file_name()])
else:
return join_paths([self.cache_dir, self.out_file_name(expression_object.name)])
def out_file_path(self, expression_object=None, *args, **kwargs):
return join_paths([self.cache_directory, self.file_suffix])