def undump_eigen(self):
"""
undumps self.adj_eigenvals and self.laplacian_matrix
"""
self.adj_eigenvals, self.adj_eigenvects = undump_object(
Dumps.val_eigen)
self.cond_eigenvals, self.cond_eigenvects = undump_object(
Dumps.cond_eigen)
def undump_matrices(self):
"""
undumps self.adjacency_Matrix and self.laplacian_matrix
"""
self.adjacency_Matrix = undump_object(
Dumps.interactome_adjacency_matrix)
self.laplacian_matrix = undump_object(
Dumps.interactome_laplacian_matrix)
def setUpClass(cls):
parser_object = UniProtParser(['199310', '405955'])
cls.uniprot_dict = parser_object.parse_uniprot(cls.up_to_parse)
cls.acces_dict = parser_object.get_access_dicts()
# dump_object(cls.ref_parses, (cls.uniprot_dict, cls.acces_dict))
cls.ref_uniprot_dict, cls.ref_acces_dict = undump_object(
cls.ref_parses)
def setUpClass(cls):
cls.actual_parser = ReactomeParser(cls.reactome_to_parse)
cls.actual_parser.parse_all()
# print cls.actual_parser.parsed
# dump_object(cls.ref_parse, cls.actual_parser)
cls.ref_parser = undump_object(cls.ref_parse)
# to parse
cls.maxDiff = None
def setUpClass(cls):
cls.actual_parser = ReactomeParser(cls.reactome_to_parse)
cls.actual_parser.parse_all()
# print cls.actual_parser.parsed
# dump_object(cls.ref_parse, cls.actual_parser)
cls.ref_parser = undump_object(cls.ref_parse)
# to parse
cls.maxDiff = None
def undump_maps(self):
"""
undumps all the elements required for the mapping between the types and ids of
database entries and matrix columns
"""
log.debug("undumping from %s", Dumps.interactome_maps)
self.bulbs_id_2_matrix_index, self.matrix_index_2_bulbs_id, \
self.bulbs_id_2_display_name, self.bulbs_id_2_legacy_id, self.bulbs_id2_node_type, \
self.bulbs_id_2_localization, self.reached_uniprots_bulbs_id_list,\
self.all_uniprots_bulbs_id_list, self.Uniprot_attachments, self.UP2Chrom, \
self.chromosomes_2_uniprot, self.uniprot_matrix_index_list,\
self.entry_point_uniprots_bulbs_ids = \
undump_object(Dumps.interactome_maps)
log.debug("post-undump e_p_u_b_i length: %s", len(self.entry_point_uniprots_bulbs_ids))
def auto_analyze(source=None, go_interface_instance=None, processors=3, desired_depth=24,
skip_sampling=False, param_set=ref_param_set):
"""
Automatically analyzes the GO annotation of the RNA_seq results.
:param source:
:param go_interface_instance:
:param processors:
:param desired_depth:
:param skip_sampling: uses existing mongoDB content without spawning a sampler
:param param_set:
"""
if source is None:
dumplist = undump_object(Dumps.RNA_seq_counts_compare)
else:
dumplist = source
if desired_depth % processors != 0:
desired_depth = desired_depth / processors + 1
else:
desired_depth = desired_depth / processors
# noinspection PyTypeChecker
for my_list in dumplist:
if go_interface_instance is None:
go_interface_instance = get_go_interface_instance(param_set)
go_interface_instance.set_uniprot_source(my_list)
if not skip_sampling:
log.info("spawning a sampler for %s proteins @ %s compops/sec",
len(go_interface_instance.analytic_uniprots), estimated_comp_ops)
# TODO: restructure to spawn a sampler pool that does not share an object in the Threading
if len(go_interface_instance.analytic_uniprots) < 200:
if not skip_sampling:
log.info('length: %s \t sampling depth: %s \t, estimated round time: %s min',
len(go_interface_instance.analytic_uniprots),
'full',
len(go_interface_instance.analytic_uniprots)**2 / estimated_comp_ops / 60)
spawn_sampler_pool(processors,
[len(go_interface_instance.analytic_uniprots)],
[desired_depth],
go_interface_instance=None,
param_set=param_set)
go_interface_instance.build_extended_conduction_system()
nr_nodes, nr_groups = compare_to_blank(
len(go_interface_instance.analytic_uniprots),
[1100, 1300],
p_val=0.9,
go_interface_instance=go_interface_instance,
param_set=param_set)
else:
sampling_depth = max(200 ** 2 / len(go_interface_instance.analytic_uniprots), 5)
if not skip_sampling:
log.info('length: %s \t sampling depth: %s \t, estimated round time: %s min',
len(go_interface_instance.analytic_uniprots),
sampling_depth,
len(go_interface_instance.analytic_uniprots) * sampling_depth / 2 / 60)
spawn_sampler_pool(processors,
[len(go_interface_instance.analytic_uniprots)],
[desired_depth],
sparse_rounds=sampling_depth,
go_interface_instance=None,
param_set=param_set)
go_interface_instance.build_extended_conduction_system(sparse_samples=sampling_depth)
# go_interface_instance.export_conduction_system()
nr_nodes, nr_groups = compare_to_blank(
len(go_interface_instance.analytic_uniprots),
[1100, 1300],
p_val=0.9, sparse_rounds=sampling_depth,
go_interface_instance=go_interface_instance,
param_set=param_set)
go_interface_instance.export_conduction_system()
for group in nr_groups:
log.info(group)
log.info('\t NodeID \t Name \t current \t informativity \t confusion_potential \t p_val \t '
'UP_list')
for node in nr_nodes:
log.info('\t %s \t %s \t %s \t %s \t %s \t %s \t %s', *node)
def undump_independent_linear_sets(self):
self.Indep_Lapl = undump_object(Dumps.GO_Indep_Linset)
def undump_memoized():
"""
:return: undumped memoized analysis
"""
return undump_object(Dumps.GO_Analysis_memoized)
def setUpClass(cls):
parser_object = UniProtParser(['199310', '405955'])
cls.uniprot_dict = parser_object.parse_uniprot(cls.up_to_parse)
cls.acces_dict = parser_object.get_access_dicts()
# dump_object(cls.ref_parses, (cls.uniprot_dict, cls.acces_dict))
cls.ref_uniprot_dict, cls.ref_acces_dict = undump_object(cls.ref_parses)
def undump_independent_linear_sets(self):
self.Indep_Lapl = undump_object(Dumps.GO_Indep_Linset)
def undump_inflated_elements(self):
self.inflated_Laplacian, self.inflated_idx2lbl, \
self.inflated_lbl2idx, self.binding_intensity = \
undump_object(Dumps.GO_Inflated)
def undump_matrices(self):
self.adjacency_matrix, self.dir_adj_matrix, self.laplacian_matrix = undump_object(
Dumps.GO_Mats)
def undump_statics():
return undump_object(Dumps.GO_builder_stat)
def undump_memoized():
""" undumps memoized analysis """
return undump_object(Dumps.Interactome_Analysis_memoized)
def undump_statics():
return undump_object(Dumps.GO_builder_stat)
def undump_core(self):
self.UP2GO_Dict, self.GO2UP, self.SeedSet, self.Reachable_nodes_dict,\
self.GO_Names, self.GO_Legacy_IDs, self.rev_GO_IDs, self.All_GOs,\
self.GO2Num, self.Num2GO, self.UP_Names, self.UPs_without_GO =\
undump_object(Dumps.GO_dump)
def undump_core(self):
self.UP2GO_Dict, self.GO2UP, self.SeedSet, self.Reachable_nodes_dict,\
self.GO_Names, self.GO_Legacy_IDs, self.rev_GO_IDs, self.All_GOs,\
self.GO2Num, self.Num2GO, self.UP_Names, self.UPs_without_GO =\
undump_object(Dumps.GO_dump)
def undump_informativities(self):
self.UP2GO_Reachable_nodes, self.GO2UP_Reachable_nodes, self.UP2GO_step_Reachable_nodes, \
self.GO2UP_step_Reachable_nodes, self.GO2_Pure_Inf, self.GO2_Weighted_Ent = \
undump_object(Dumps.GO_Infos)
def undump_matrices(self):
self.adjacency_matrix, self.dir_adj_matrix, self.laplacian_matrix = undump_object(
Dumps.GO_Mats)
def undump_memoized():
"""
:return: undumped memoized analysis
"""
return undump_object(Dumps.GO_Analysis_memoized)
def undump_informativities(self):
self.UP2GO_Reachable_nodes, self.GO2UP_Reachable_nodes, self.UP2GO_step_Reachable_nodes, \
self.GO2UP_step_Reachable_nodes, self.GO2_Pure_Inf, self.GO2_Weighted_Ent = \
undump_object(Dumps.GO_Infos)
def auto_analyze(source=None,
go_interface_instance=None,
processors=3,
desired_depth=24,
skip_sampling=False,
param_set=ref_param_set):
"""
Automatically analyzes the GO annotation of the RNA_seq results.
:param source:
:param go_interface_instance:
:param processors:
:param desired_depth:
:param skip_sampling: uses existing mongoDB content without spawning a sampler
:param param_set:
"""
if source is None:
dumplist = undump_object(Dumps.RNA_seq_counts_compare)
else:
dumplist = source
if desired_depth % processors != 0:
desired_depth = desired_depth / processors + 1
else:
desired_depth = desired_depth / processors
# noinspection PyTypeChecker
for my_list in dumplist:
if go_interface_instance is None:
go_interface_instance = get_go_interface_instance(param_set)
go_interface_instance.set_uniprot_source(my_list)
if not skip_sampling:
log.info("spawning a sampler for %s proteins @ %s compops/sec",
len(go_interface_instance.analytic_uniprots),
estimated_comp_ops)
# TODO: restructure to spawn a sampler pool that does not share an object in the Threading
if len(go_interface_instance.analytic_uniprots) < 200:
if not skip_sampling:
log.info(
'length: %s \t sampling depth: %s \t, estimated round time: %s min',
len(go_interface_instance.analytic_uniprots), 'full',
len(go_interface_instance.analytic_uniprots)**2 /
estimated_comp_ops / 60)
spawn_sampler_pool(
processors, [len(go_interface_instance.analytic_uniprots)],
[desired_depth],
go_interface_instance=None)
go_interface_instance.build_extended_conduction_system()
nr_nodes, nr_groups = compare_to_blank(
len(go_interface_instance.analytic_uniprots), [1100, 1300],
p_val=0.9,
go_interface_instance=go_interface_instance,
param_set=param_set)
else:
sampling_depth = max(
200**2 / len(go_interface_instance.analytic_uniprots), 5)
if not skip_sampling:
log.info(
'length: %s \t sampling depth: %s \t, estimated round time: %s min',
len(go_interface_instance.analytic_uniprots),
sampling_depth,
len(go_interface_instance.analytic_uniprots) *
sampling_depth / 2 / 6 / 60)
spawn_sampler_pool(
processors, [len(go_interface_instance.analytic_uniprots)],
[desired_depth],
sparse_rounds=sampling_depth,
go_interface_instance=None)
go_interface_instance.build_extended_conduction_system(
sparse_samples=sampling_depth)
# go_interface_instance.export_conduction_system()
nr_nodes, nr_groups = compare_to_blank(
len(go_interface_instance.analytic_uniprots), [1100, 1300],
p_val=0.9,
sparse_rounds=sampling_depth,
go_interface_instance=go_interface_instance,
param_set=param_set)
go_interface_instance.export_conduction_system()
for group in nr_groups:
log.info(group)
log.info(
'\t NodeID \t Name \t current \t informativity \t confusion_potential \t p_val \t '
'UP_list')
for node in nr_nodes:
log.info('\t %s \t %s \t %s \t %s \t %s \t %s \t %s', *node)
def undump_inflated_elements(self):
self.inflated_Laplacian, self.inflated_idx2lbl, \
self.inflated_lbl2idx, self.binding_intensity = \
undump_object(Dumps.GO_Inflated)
