def _process_input_files(self):
interactions_file = open(self.args[0], "rb")
annotations_file = open(self.args[1], "rb")
# Create interaction graph
logger.info("Parsing interactions from {0}.".format(interactions_file.name))
self.interactions_graph = parsers.parse_interactions_file_to_graph(interactions_file)
logger.info(
"{0} genes (products) with {1} interactions "
"parsed.".format(len(self.interactions_graph), self.interactions_graph.number_of_edges())
)
# Create dictionary of annotations to genes, but only for genes in
# the interaction graph
logger.info("Parsing annotations from {0}.".format(annotations_file.name))
if annotations_file.name.endswith(".gmt"):
self.annotations_dict = parsers.parse_gmt_to_dict(annotations_file)
else:
self.annotations_dict = parsers.parse_annotations_to_dict(annotations_file)
self.annotations_stats = structures.get_annotations_stats(self.annotations_dict)
logger.info(
(
"{num_total_annotations} annotations processed, "
"for {num_genes} genes (or gene products), by "
"{num_annotation_terms} different terms.".format(**self.annotations_stats)
)
)
# Remove from the graph the set of nodes that have no annotation.
logger.info("Pruning unannotated genes (products) from " "interaction graph.")
self.interactions_graph.prune_unannotated_genes(self.annotations_dict)
logger.info(
"{0} genes (products) with {1} interactions "
"remaining in graph.".format(len(self.interactions_graph), self.interactions_graph.number_of_edges())
)
# Remove from the annotations any genes which are not in the graph.
logger.info("Removing genes with no interactions from the " "sets of annotated genes.")
self.interactions_graph.prune_non_network_genes_from_annotations(self.annotations_dict)
self.annotations_stats = structures.get_annotations_stats(self.annotations_dict)
logger.info(
"{num_total_annotations} annotations, "
"for {num_genes} genes (or gene products), by "
"{num_annotation_terms} different terms "
"remain.".format(**self.annotations_stats)
)
# Sanity test: the number of genes (products) in the
# interactions_graph should equal the union of all the sets in
# annotations_dict
assert len(self.interactions_graph) == self.annotations_stats["num_genes"], (
"interactions_graph and annotations_dict have unequal " "numbers of genes!"
)
for term, genes in self.annotations_dict.iteritems():
assert len(genes) > 0, "%s has no genes!" % term
interactions_file.close()
annotations_file.close()
def _process_input_files(self):
super(ContextualCli, self)._process_input_files()
expression_file = open(self.args[2], "rb")
# Get the expression values.
logger.info("Parsing expression values from %s." % expression_file.name)
expression_values = parsers.parse_expression_file(expression_file)
# Apply the expression values to the interaction graph, removing
# any nodes lacking expression values from the graph.
logger.info("Removing genes without expression values from " "interaction graph and annotation sets.")
self.interactions_graph.apply_expression_values_to_interactions_graph(expression_values)
# Re-synchronize the interaction graph and annotations dictionary.
self.interactions_graph.prune_non_network_genes_from_annotations(self.annotations_dict)
expression_file.close()
self.annotations_stats = structures.get_annotations_stats(self.annotations_dict)
gene_stats = {"num_interactions": self.interactions_graph.number_of_edges()}
gene_stats.update(self.annotations_stats)
logger.info(
"%(num_genes)d genes (products) with "
"%(num_interactions)d interactions remaining in "
"graph, with %(num_total_annotations)d annotations by "
"%(num_annotation_terms)d terms." % gene_stats
)
