def analyze(matrixtype, depth, processors,):
"""
Performs an analysis of the type given by matrixtype with the given background set and
:param matrixtype:
:param depth:
:param processors:
:return:
"""
source_bulbs_ids = get_source_bulbs_ids()
background_bulbs_ids = get_background_bulbs_ids()
# TODO: CRICIAL: inject background usage when background switch is available.
# Refer to the analysis pipeline example for an example
interactome_interface_instance = InteractomeInterface(main_connex_only=True, full_impact=False)
interactome_interface_instance.fast_load()
ref_param_set = [['biological_process'], background_bulbs_ids, (1, 1), True, 3]
if matrixtype == 'interactome':
interactome_analysis(source_bulbs_ids, depth, processors, background_bulbs_ids)
elif matrixtype == 'annotome':
knowledge_analysis(source=source_bulbs_ids,
desired_depth=depth,
processors=processors,
param_set=ref_param_set)
print "analsysis is finished, current results are stored " \
"in the outputs directory"
def rebuildlaplacians():
"""
Extracts the Laplacian matrices from the master graph database.
\f
:return:
"""
from bioflow.utils.top_level import rebuild_the_laplacians
from bioflow.utils.io_routines import get_background_bulbs_ids
background_bulbs_ids = get_background_bulbs_ids()
rebuild_the_laplacians(all_detectable_genes=background_bulbs_ids)
def map_and_save_gene_ids(hit_genes_location, all_detectable_genes_location=''):
cast_analysis_set_to_bulbs_ids(hit_genes_location)
hit_genes_ids = get_source_bulbs_ids()
if all_detectable_genes_location:
cast_background_set_to_bulbs_id(
background_set_csv_location=all_detectable_genes_location,
analysis_set_csv_location=hit_genes_location)
all_detectable_genes_ids = get_background_bulbs_ids()
else:
all_detectable_genes_ids = []
writer(open(Dumps.background_set_bulbs_ids, 'w'), delimiter='\n').writerow(all_detectable_genes_ids)
return hit_genes_ids, all_detectable_genes_ids
def extractmatrix(matrixtype):
"""
Extracts the matrix interface object for the computation routine.
:param matrixtype:
:return:
"""
if matrixtype == 'interactome':
local_matrix = InteractomeInterface(main_connex_only=True, full_impact=False)
local_matrix.full_rebuild()
if matrixtype == 'annotome':
local_matrix = InteractomeInterface(main_connex_only=True, full_impact=False)
local_matrix.fast_load()
ref_param_set = [['biological_process'], get_background_bulbs_ids(), (1, 1), True, 3]
annot_matrix = AnnotomeInterface(*ref_param_set)
annot_matrix.full_rebuild()
def interactomeanalysis(depth, processors, skipsampling, skiphitflow):
"""
Performs interactome analysis given background set given earlier.
\f
:param depth:
:param processors:
:param skipsampling:
:param skiphitflow:
:return:
"""
from bioflow.utils.io_routines import get_background_bulbs_ids, get_source_bulbs_ids
from bioflow.molecular_network.interactome_analysis import auto_analyze as interactome_analysis
source_bulbs_ids = get_source_bulbs_ids()
background_bulbs_ids = get_background_bulbs_ids()
interactome_analysis([source_bulbs_ids],
desired_depth=depth,
processors=processors,
background_list=background_bulbs_ids,
skip_sampling=skipsampling,
from_memoization=skiphitflow)
those that share a significant amount of reached_uniprots_bulbs_id_list in common
"""
self.Indep_Lapl = lil_matrix((len(self.All_GOs), len(self.All_GOs)))
for GO_list in self.UP2GO_Reachable_nodes.itervalues():
for GO1, GO2 in combinations(GO_list, 2):
idx1, idx2 = (self.GO2Num[GO1], self.GO2Num[GO2])
self.Indep_Lapl[idx1, idx2] += -1
self.Indep_Lapl[idx2, idx1] += -1
self.Indep_Lapl[idx2, idx2] += 1
self.Indep_Lapl[idx1, idx1] += 1
if __name__ == '__main__':
# Creates an instance of MatrixGetter and loads pre-computed values
go_interface_instance = GeneOntologyInterface(uniprot_node_ids=get_background_bulbs_ids())
go_interface_instance.full_rebuild()
# loading takes 1-6 seconds.
# fill for reach only is done in 2 seconds,
# tepping takes another 15,
# inverting + info computation - 1 more second
# Laplacian building =>
##
# full computation - 3 minutes 18 seconds; save 7 seconds, retrieval - 3
# seconds
# go_interface_instance.load()
# print go_interface_instance.pretty_time()
# go_interface_instance.get_indep_linear_groups()
those that share a significant amount of reached_uniprots_neo4j_id_list in common
"""
self.Indep_Lapl = lil_matrix((len(self.All_GOs), len(self.All_GOs)))
for GO_list in self.UP2GO_Reachable_nodes.itervalues():
for GO1, GO2 in combinations(GO_list, 2):
idx1, idx2 = (self.GO2Num[GO1], self.GO2Num[GO2])
self.Indep_Lapl[idx1, idx2] += -1
self.Indep_Lapl[idx2, idx1] += -1
self.Indep_Lapl[idx2, idx2] += 1
self.Indep_Lapl[idx1, idx1] += 1
if __name__ == '__main__':
# Creates an instance of MatrixGetter and loads pre-computed values
go_interface_instance = GeneOntologyInterface(uniprot_node_ids=get_background_bulbs_ids())
go_interface_instance.full_rebuild()
# loading takes 1-6 seconds.
# fill for reach only is done in 2 seconds,
# tepping takes another 15,
# inverting + info computation - 1 more second
# Laplacian building =>
##
# full computation - 3 minutes 18 seconds; save 7 seconds, retrieval - 3
# seconds
# go_interface_instance.load()
# print go_interface_instance.pretty_time()
# go_interface_instance.get_indep_linear_groups()
for group in nr_groups:
log.info(group)
log.info('\t Node_ID \t Name \t current \t connectedness \t p_value')
for node in nr_nodes:
log.info('\t %s \t %s \t %s \t %s \t %s', *node)
if __name__ == "__main__":
# pprinter = PrettyPrinter(indent=4)
# interactome_interface_instance = MatrixGetter(True, False)
# interactome_interface_instance.fast_load()
# dumplist = undump_object(Dumps.RNA_seq_counts_compare)
# MG1.randomly_sample([150], [1], chromosome_specific=15, No_add=True)
# nr_nodes, nr_groups = compare_to_blanc(150, [0.5, 0.6], MG1, p_val=0.9)
# MG1.export_conduction_system()
# for group in nr_groups:
# print group
# for node in nr_nodes:
# print node
source = get_source_bulbs_ids()
background_list = get_background_bulbs_ids()
auto_analyze([source],
desired_depth=5,
processors=6,
background_list=background_list,
skip_sampling=True)
# # building the neo4j database
# build_db()
# # set the source file of the ids of perturbed proteins and background set:
# "/home/andrei/2nd_pass_2x.txt"
# "/home/andrei/Linhao_imaging.txt"
# "/home/andrei/HS_30_Linhao_outliers.txt"
# cast_analysis_set_to_bulbs_ids("/home/andrei/Linhao_imaging.txt")
#
# cast_background_set_to_bulbs_id(
# background_set_csv_location=None,
# analysis_set_csv_location="/home/andrei/HS_30_Linhao_outliers.txt")
# # get the bulbs ids oif the nodes we would like to analyze
source_bulbs_ids = get_source_bulbs_ids()
background_bulbs_ids = get_background_bulbs_ids()
print len(source_bulbs_ids)
print len(background_bulbs_ids)
# # building the interactome interface object
# local_matrix = InteractomeInterface(main_connex_only=True, full_impact=False)
# local_matrix.full_rebuild()
# # perform the interactome analysis
# interactome_analysis([source_bulbs_ids], desired_depth=24, processors=3,
# background_list=background_bulbs_ids, skip_sampling=False)
# # building the reference parameters set
_filter = ['biological_process']
ref_param_set = [_filter, background_bulbs_ids, (1, 1), True, 3]
"""
self.Indep_Lapl = lil_matrix((len(self.All_GOs), len(self.All_GOs)))
for GO_list in self.UP2GO_Reachable_nodes.itervalues():
for GO1, GO2 in combinations(GO_list, 2):
idx1, idx2 = (self.GO2Num[GO1], self.GO2Num[GO2])
self.Indep_Lapl[idx1, idx2] += -1
self.Indep_Lapl[idx2, idx1] += -1
self.Indep_Lapl[idx2, idx2] += 1
self.Indep_Lapl[idx1, idx1] += 1
if __name__ == '__main__':
# Creates an instance of MatrixGetter and loads pre-computed values
go_interface_instance = GeneOntologyInterface(
uniprot_node_ids=get_background_bulbs_ids())
go_interface_instance.full_rebuild()
# loading takes 1-6 seconds.
# fill for reach only is done in 2 seconds,
# tepping takes another 15,
# inverting + info computation - 1 more second
# Laplacian building =>
##
# full computation - 3 minutes 18 seconds; save 7 seconds, retrieval - 3
# seconds
# go_interface_instance.load()
# print go_interface_instance.pretty_time()
# go_interface_instance.get_indep_linear_groups()