def run(args):
subnetworks_file = '%s/viz_files/%s' % (str(hotnet2.__file__).rsplit('/', 1)[0], VIZ_SUBNETWORKS)
# create output directory if doesn't exist; warn if it exists and is not empty
outdir = args.output_directory
if not os.path.exists(outdir):
os.makedirs(outdir)
if len(os.listdir(outdir)) > 0:
print("WARNING: Output directory is not empty. Any conflicting files will be overwritten. "
"(Ctrl-c to cancel).")
ks = set()
output = dict(deltas=[], subnetworks=dict(), mutation_matrices=dict(), stats=dict())
subnetworks = dict()
for results_file in args.results_files:
results = json.load(open(results_file))
ccs = results['components']
heat_file = json.load(open(results['parameters']['heat_file']))
gene2heat = heat_file['heat']
heat_parameters = heat_file['parameters']
d_score = hnio.load_display_score_tsv(args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(args.display_name_file) if args.display_name_file else dict()
edges = hnio.load_ppi_edges(args.edge_file, hnio.load_index(results['parameters']['infmat_index_file']))
delta = format(results['parameters']['delta'], 'g')
output['deltas'].append(delta)
subnetworks[delta] = ccs
output["subnetworks"][delta] = []
for cc in ccs:
output['subnetworks'][delta].append(viz.get_component_json(cc, gene2heat, edges,
args.network_name, d_score, d_name))
# make oncoprints if heat file was generated from mutation data
if 'heat_fn' in heat_parameters and heat_parameters['heat_fn'] == 'load_mutation_heat':
output['mutation_matrices'][delta] = list()
samples = hnio.load_samples(heat_parameters['sample_file']) if heat_parameters['sample_file'] else None
genes = hnio.load_genes(heat_parameters['gene_file']) if heat_parameters['gene_file'] else None
snvs = hnio.load_snvs(heat_parameters['snv_file'], genes, samples) if heat_parameters['snv_file'] else []
cnas = hnio.load_cnas(heat_parameters['cna_file'], genes, samples) if heat_parameters['cna_file'] else []
for cc in ccs:
output['mutation_matrices'][delta].append(viz.get_oncoprint_json(cc, snvs, cnas, d_name))
if heat_parameters.get('sample_type_file'):
with open(heat_parameters['sample_type_file']) as f:
output['sampleToTypes'] = dict(l.rstrip().split() for l in f if not l.startswith("#") )
output['typeToSamples'] = dict((t, []) for t in set(output['sampleToTypes'].values()))
for s, ty in output['sampleToTypes'].iteritems():
output['typeToSamples'][ty].append( s )
else:
output['sampleToTypes'] = dict( (s, "Cancer") for s in samples )
output['typeToSamples'] = dict(Cancer=list(samples))
output['stats'][delta] = results['statistics']
for k in sorted(map(int, results['statistics'].keys())):
ks.add(k)
continue
stats = results['statistics'][str(k)]
output['stats'][delta].append( dict(k=k, expected=stats['expected'], observed=stats['observed'], pval=stats['pval']))
output['ks'] = range(min(ks), max(ks)+1)
with open('%s/subnetworks.json' % outdir, 'w') as out:
json.dump(output, out, indent=4)
shutil.copy(subnetworks_file, '%s/%s' % (outdir, VIZ_INDEX))
def run(args):
# Load the network and heat files
assert( len(args.network_files) == len(args.permuted_network_paths) )
networks, graph_map = [], dict()
for network_file, pnp in zip(args.network_files, args.permuted_network_paths):
infmat, indexToGene, G, network_name = hnio.load_network(network_file, HN2_INFMAT_NAME)
graph_map[network_name] = G
networks.append( (infmat, indexToGene, G, network_name, pnp) )
heats, json_heat_map, heat_map, mutation_map, heat_file_map = [], dict(), dict(), dict(), dict()
for heat_file in args.heat_files:
json_heat = os.path.splitext(heat_file.lower())[1] == '.json'
heat, heat_name, mutations = hnio.load_heat_file(heat_file, json_heat)
json_heat_map[heat_name] = json_heat
heat_map[heat_name] = heat
heat_file_map[heat_name] = heat_file
mutation_map[heat_name] = mutations
heats.append( (heat, heat_name) )
# Run HotNet2 on each pair of network and heat files
if args.verbose > 0:
print('* Running HotNet2 in consensus mode...')
single_runs, consensus, linkers, auto_deltas, consensus_stats = consensus_with_stats(args, networks, heats)
# Output the single runs
if args.verbose > 0:
print('* Outputting results to file...')
params = vars(args)
result_dirs = []
for (network_name, heat_name, run) in single_runs:
# Set up the output directory and record for later
output_dir = '%s/%s-%s' % (args.output_directory, network_name.lower(), heat_name.lower())
result_dirs.append(output_dir)
hnio.setup_output_dir(output_dir)
# Output to file
hnio.output_hotnet2_run(run, params, network_name, heat_map[heat_name], heat_name, heat_file_map[heat_name], json_heat_map[heat_name], output_dir)
# create the hierarchy if necessary
if args.output_hierarchy:
hierarchy_out_dir = '{}/hierarchy/'.format(output_dir)
if not os.path.isdir(hierarchy_out_dir): os.mkdir(hierarchy_out_dir)
CD.createDendrogram( sim, list(index2gene.values()), hierarchy_out_dir, params, verbose=False)
# Output the consensus
hnio.output_consensus(consensus, linkers, auto_deltas, consensus_stats, params, args.output_directory)
# Create the visualization(s). This has to be after the consensus procedure
# is run because we want to default to the auto-selected deltas.
if args.verbose > 0:
print('* Generating and outputting visualization data...')
d_score = hnio.load_display_score_tsv(args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(args.display_name_file) if args.display_name_file else dict()
for (network_name, heat_name, run), result_dir, auto_delta in zip(single_runs, result_dirs, auto_deltas):
snvs, cnas, sampleToType = mutation_map[heat_name]
G = graph_map[network_name]
output = hnviz.generate_viz_json(run, G.edges(), network_name, heat_map[heat_name], snvs, cnas, sampleToType, d_score, d_name)
with open('{}/viz-data.json'.format(result_dir), 'w') as OUT:
output['params'] = dict(consensus=False, network_name=network_name, heat_name=heat_name, auto_delta=format(auto_delta, 'g'))
json.dump( output, OUT )
# Add the consensus visualization
snvs, cnas, sampleToType = mutations
consensus_ccs = [ d['core'] + d['expansion'] for d in consensus ]
consensus_auto_delta = 0
results = [[consensus_ccs, consensus_stats, consensus_auto_delta]]
with open('{}/consensus/viz-data.json'.format(args.output_directory), 'w') as OUT:
output = hnviz.generate_viz_json(results, G.edges(), network_name, heat, snvs, cnas, sampleToType, d_score, d_name)
output['params'] = dict(consensus=True, auto_delta=format(consensus_auto_delta, 'g'))
json.dump( output, OUT )
def run(args):
# Load the network and heat files
assert (len(args.network_files) == len(args.permuted_network_paths))
networks, graph_map = [], dict()
for network_file, pnp in zip(args.network_files,
args.permuted_network_paths):
infmat, indexToGene, G, network_name = hnio.load_network(
network_file, HN2_INFMAT_NAME)
graph_map[network_name] = G
networks.append((infmat, indexToGene, G, network_name, pnp))
heats, json_heat_map, heat_map, mutation_map, heat_file_map = [], dict(
), dict(), dict(), dict()
for heat_file in args.heat_files:
json_heat = os.path.splitext(heat_file.lower())[1] == '.json'
heat, heat_name, mutations = hnio.load_heat_file(heat_file, json_heat)
json_heat_map[heat_name] = json_heat
heat_map[heat_name] = heat
heat_file_map[heat_name] = heat_file
mutation_map[heat_name] = mutations
heats.append((heat, heat_name))
# Run HotNet2 on each pair of network and heat files
if args.verbose > 0:
print '* Running HotNet2 in consensus mode...'
single_runs, consensus, linkers, auto_deltas, consensus_stats = consensus_with_stats(
args, networks, heats)
# Output the single runs
if args.verbose > 0:
print '* Outputting results to file...'
params = vars(args)
result_dirs = []
for (network_name, heat_name, run) in single_runs:
# Set up the output directory and record for later
output_dir = '%s/%s-%s' % (args.output_directory, network_name.lower(),
heat_name.lower())
result_dirs.append(output_dir)
hnio.setup_output_dir(output_dir)
# Output to file
hnio.output_hotnet2_run(run, params, network_name, heat_map[heat_name],
heat_name, heat_file_map[heat_name],
json_heat_map[heat_name], output_dir)
# create the hierarchy if necessary
if args.output_hierarchy:
hierarchy_out_dir = '{}/hierarchy/'.format(output_dir)
if not os.path.isdir(hierarchy_out_dir):
os.mkdir(hierarchy_out_dir)
CD.createDendrogram(sim,
index2gene.values(),
hierarchy_out_dir,
params,
verbose=False)
# Output the consensus
hnio.output_consensus(consensus, linkers, auto_deltas, consensus_stats,
params, args.output_directory)
# Create the visualization(s). This has to be after the consensus procedure
# is run because we want to default to the auto-selected deltas.
if args.verbose > 0:
print '* Generating and outputting visualization data...'
d_score = hnio.load_display_score_tsv(
args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(
args.display_name_file) if args.display_name_file else dict()
for (network_name, heat_name,
run), result_dir, auto_delta in zip(single_runs, result_dirs,
auto_deltas):
snvs, cnas, sampleToType = mutation_map[heat_name]
G = graph_map[network_name]
output = hnviz.generate_viz_json(run, G.edges(), network_name,
heat_map[heat_name], snvs, cnas,
sampleToType, d_score, d_name)
with open('{}/viz-data.json'.format(result_dir), 'w') as OUT:
output['params'] = dict(consensus=False,
network_name=network_name,
heat_name=heat_name,
auto_delta=format(auto_delta, 'g'))
json.dump(output, OUT)
# Add the consensus visualization
snvs, cnas, sampleToType = mutations
consensus_ccs = [d['core'] + d['expansion'] for d in consensus]
consensus_auto_delta = 0
results = [[consensus_ccs, consensus_stats, consensus_auto_delta]]
with open('{}/consensus/viz-data.json'.format(args.output_directory),
'w') as OUT:
output = hnviz.generate_viz_json(results, G.edges(), network_name,
heat, snvs, cnas, sampleToType,
d_score, d_name)
output['params'] = dict(consensus=True,
auto_delta=format(consensus_auto_delta, 'g'))
json.dump(output, OUT)
def run(args):
subnetworks_file = '%s/viz_files/%s' % (str(hotnet2.__file__).rsplit('/', 1)[0], VIZ_SUBNETWORKS)
# create output directory if doesn't exist; warn if it exists and is not empty
outdir = args.output_directory
if not os.path.exists(outdir):
os.makedirs(outdir)
if len(os.listdir(outdir)) > 0:
print("WARNING: Output directory is not empty. Any conflicting files will be overwritten. "
"(Ctrl-c to cancel).")
ks = set()
output = dict(deltas=[], subnetworks=dict(), mutation_matrices=dict(), stats=dict())
predictions = set()
multipleHeatFiles = False
for results_file in args.results_files:
with open(results_file, 'r') as IN:
results = json.load(IN)
ccs = results['components']
heat_file = json.load(open(results['parameters']['heat_file']))
gene2heat = heat_file['heat']
heat_parameters = heat_file['parameters']
d_score = hnio.load_display_score_tsv(args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(args.display_name_file) if args.display_name_file else dict()
edges = hnio.load_ppi_edges(args.edge_file, hnio.load_index(results['parameters']['infmat_index_file']))
delta = format(results['parameters']['delta'], 'g')
output['deltas'].append(delta)
output["subnetworks"][delta] = []
predictions |= set( g for cc in ccs for g in cc )
for cc in ccs:
output['subnetworks'][delta].append(viz.get_component_json(cc, gene2heat, edges,
args.network_name, d_score, d_name))
# Record the heat scores
if 'geneToHeat' in output:
if any( output['geneToHeat'][g] != h for g, h in gene2heat.iteritems() ) or len(gene2heat.keys()) != len(output['geneToHeat'].keys()):
multipleHeatFiles = True
output['geneToHeat'] = gene2heat
# make oncoprints if heat file was generated from mutation data
if 'heat_fn' in heat_parameters and heat_parameters['heat_fn'] == 'load_mutation_heat':
output['mutation_matrices'][delta] = list()
samples = hnio.load_samples(heat_parameters['sample_file']) if heat_parameters['sample_file'] else None
genes = hnio.load_genes(heat_parameters['gene_file']) if heat_parameters['gene_file'] else None
snvs = hnio.load_snvs(heat_parameters['snv_file'], genes, samples) if heat_parameters['snv_file'] else []
cnas = hnio.load_cnas(heat_parameters['cna_file'], genes, samples) if heat_parameters['cna_file'] else []
# Get the samples and genes from the mutations directly if they weren't provided
if not samples:
samples = set( m.sample for m in snvs ) | set( m.sample for m in cnas )
if not genes:
genes = set( m.gene for m in snvs) | set( m.gene for m in cnas )
for cc in ccs:
output['mutation_matrices'][delta].append(viz.get_oncoprint_json(cc, snvs, cnas, d_name))
if heat_parameters.get('sample_type_file'):
with open(heat_parameters['sample_type_file']) as f:
output['sampleToTypes'] = dict(l.rstrip().split() for l in f if not l.startswith("#") )
output['typeToSamples'] = dict((t, []) for t in set(output['sampleToTypes'].values()))
for s, ty in output['sampleToTypes'].iteritems():
output['typeToSamples'][ty].append( s )
else:
if not samples:
samples = set( m.sample for m in snvs ) | set( m.sample for m in cnas )
output['sampleToTypes'] = dict( (s, "Cancer") for s in samples )
output['typeToSamples'] = dict(Cancer=list(samples))
output['stats'][delta] = results['statistics']
ks |= set(map(int, results['statistics'].keys()))
# Print a warning if there were multiple heat files referenced by
# the results files
if multipleHeatFiles:
sys.stderr.write('Warning: results files used multiple heat files. Only the last heat file will be used to tabulate scores.\n')
# Output to file
output['predictions'] = sorted(predictions) # list of nodes found in any run
output['ks'] = range(min(ks), max(ks)+1)
with open('%s/subnetworks.json' % outdir, 'w') as out:
json.dump(output, out, indent=4)
shutil.copy(subnetworks_file, '%s/%s' % (outdir, VIZ_INDEX))
def run(args):
subnetworks_file = '%s/viz_files/%s' % (str(hotnet2.__file__).rsplit(
'/', 1)[0], VIZ_SUBNETWORKS)
# create output directory if doesn't exist; warn if it exists and is not empty
outdir = args.output_directory
if not os.path.exists(outdir):
os.makedirs(outdir)
if len(os.listdir(outdir)) > 0:
print(
"WARNING: Output directory is not empty. Any conflicting files will be overwritten. "
"(Ctrl-c to cancel).")
ks = set()
output = dict(deltas=[],
subnetworks=dict(),
mutation_matrices=dict(),
stats=dict())
subnetworks = dict()
for results_file in args.results_files:
results = json.load(open(results_file))
ccs = results['components']
heat_file = json.load(open(results['parameters']['heat_file']))
gene2heat = heat_file['heat']
heat_parameters = heat_file['parameters']
d_score = hnio.load_display_score_tsv(
args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(
args.display_name_file) if args.display_name_file else dict()
edges = hnio.load_ppi_edges(
args.edge_file,
hnio.load_index(results['parameters']['infmat_index_file']))
delta = format(results['parameters']['delta'], 'g')
output['deltas'].append(delta)
subnetworks[delta] = ccs
output["subnetworks"][delta] = []
for cc in ccs:
output['subnetworks'][delta].append(
viz.get_component_json(cc, gene2heat, edges, args.network_name,
d_score, d_name))
# make oncoprints if heat file was generated from mutation data
if 'heat_fn' in heat_parameters and heat_parameters[
'heat_fn'] == 'load_mutation_heat':
output['mutation_matrices'][delta] = list()
samples = hnio.load_samples(
heat_parameters['sample_file']
) if heat_parameters['sample_file'] else None
genes = hnio.load_genes(heat_parameters['gene_file']
) if heat_parameters['gene_file'] else None
snvs = hnio.load_snvs(
heat_parameters['snv_file'], genes,
samples) if heat_parameters['snv_file'] else []
cnas = hnio.load_cnas(
heat_parameters['cna_file'], genes,
samples) if heat_parameters['cna_file'] else []
# Get the samples and genes from the mutations directly if they weren't provided
if not samples:
samples = set(m.sample for m in snvs) | set(m.sample
for m in cnas)
if not genes:
genes = set(m.gene for m in snvs) | set(m.gene for m in cnas)
for cc in ccs:
output['mutation_matrices'][delta].append(
viz.get_oncoprint_json(cc, snvs, cnas, d_name))
if heat_parameters.get('sample_type_file'):
with open(heat_parameters['sample_type_file']) as f:
output['sampleToTypes'] = dict(l.rstrip().split()
for l in f
if not l.startswith("#"))
output['typeToSamples'] = dict(
(t, []) for t in set(output['sampleToTypes'].values()))
for s, ty in output['sampleToTypes'].iteritems():
output['typeToSamples'][ty].append(s)
else:
output['sampleToTypes'] = dict((s, "Cancer") for s in samples)
output['typeToSamples'] = dict(Cancer=list(samples))
output['stats'][delta] = results['statistics']
ks |= set(map(int, results['statistics'].keys()))
output['ks'] = range(min(ks), max(ks) + 1)
with open('%s/subnetworks.json' % outdir, 'w') as out:
json.dump(output, out, indent=4)
shutil.copy(subnetworks_file, '%s/%s' % (outdir, VIZ_INDEX))
def run(args):
# Load the network and heat files
assert (len(args.network_files) == len(args.permuted_network_paths))
networks, graph_map = [], dict()
for network_file, pnp in zip(args.network_files,
args.permuted_network_paths):
infmat, indexToGene, G, network_name = hnio.load_network(
network_file, HN2_INFMAT_NAME)
graph_map[network_name] = G
networks.append((infmat, indexToGene, G, network_name, pnp))
"""Change from here"""
#print type(G)
"""End here """
heats, json_heat_map, heat_map, mutation_map, heat_file_map = [], dict(
), dict(), dict(), dict()
for heat_file in args.heat_files:
json_heat = os.path.splitext(heat_file.lower())[1] == '.json'
heat, heat_name, mutations = hnio.load_heat_file(heat_file, json_heat)
json_heat_map[heat_name] = json_heat
heat_map[heat_name] = heat
heat_file_map[heat_name] = heat_file
mutation_map[heat_name] = mutations
heats.append((heat, heat_name))
# Run HotNet2 on each pair of network and heat files
if args.verbose > 0:
print '* Running HotNet2 in consensus mode...'
single_runs, single_My_results, consensus, linkers, auto_deltas, consensus_stats = consensus_with_stats(
args, networks, heats)
#print single_sig_Count_delta
#print single_sig_Conp_delta
#print signle_Subnet_sig_delta
# Output the single runs
if args.verbose > 0:
print '* Outputting results to file...'
params = vars(args)
result_dirs = []
for (network_name, heat_name, run) in single_runs:
# Set up the output directory and record for later
output_dir = '%s/%s-%s' % (args.output_directory, network_name.lower(),
heat_name.lower())
result_dirs.append(output_dir)
hnio.setup_output_dir(output_dir)
# Output to file
hnio.output_hotnet2_run(run, params, network_name, heat_map[heat_name],
heat_name, heat_file_map[heat_name],
json_heat_map[heat_name], output_dir)
# create the hierarchy if necessary
if args.output_hierarchy:
hierarchy_out_dir = '{}/hierarchy/'.format(output_dir)
if not os.path.isdir(hierarchy_out_dir):
os.mkdir(hierarchy_out_dir)
CD.createDendrogram(sim,
index2gene.values(),
hierarchy_out_dir,
params,
verbose=False)
"""change my code here"""
# Set up the output directory and record for later
#for (network_name, heat_name, subnetScore) in single_subnet_score_delta:
#output_dir = '%s/%s-%s' % (args.output_directory, network_name.lower(), heat_name.lower())
# result_dirs.append(output_dir)
# hnio.setup_output_dir(output_dir)
# my.output_subnetscore_deltas(subnetScore,output_dir)
for (network_name, heat_name, Myresults) in single_My_results:
output_dir = '%s/%s-%s' % (args.output_directory, network_name.lower(),
heat_name.lower())
result_dirs.append(output_dir)
hnio.setup_output_dir(output_dir)
#(permuted_sub_score, subnet_geneScore, Alpha_sig, Subnetscore_sig, Conponet_sig, count, sig_conponet, degree, permuted_degree,cent_weighted_score, degree_weighted_score,delta)
my.output_Myresults(Myresults, output_dir)
###### permuted results store
"""end here"""
# Output the consensus
hnio.output_consensus(consensus, linkers, auto_deltas, consensus_stats,
params, args.output_directory)
"""changing here, I calculate the genes score"""
#####calculate all the gene score for gene list
#genelist, scoreVec = my.score_vec(infmat,indexToGene, heat)
#gene_socre = np.column_stack((genelist, scoreVec))
#subnet_geneScore = my.Matchscore_fun(genelist,scoreVec,consensus)
####output the subnetwork with score only
#my.output_scoreSubnet(subnet_geneScore,args.output_directory)
####output the result
#hnio.output_scoreVec(scoreVec, genelist,args.output_directory)
"""end here """
# Create the visualization(s). This has to be after the consensus procedure
# is run because we want to default to the auto-selected deltas.
if args.verbose > 0:
print '* Generating and outputting visualization data...'
d_score = hnio.load_display_score_tsv(
args.display_score_file) if args.display_score_file else None
d_name = hnio.load_display_name_tsv(
args.display_name_file) if args.display_name_file else dict()
for (network_name, heat_name,
run), result_dir, auto_delta in zip(single_runs, result_dirs,
auto_deltas):
snvs, cnas, sampleToType = mutation_map[heat_name]
G = graph_map[network_name]
output = hnviz.generate_viz_json(run, G.edges(), network_name,
heat_map[heat_name], snvs, cnas,
sampleToType, d_score, d_name)
with open('{}/viz-data.json'.format(result_dir), 'w') as OUT:
output['params'] = dict(consensus=False,
network_name=network_name,
heat_name=heat_name,
auto_delta=format(auto_delta, 'g'))
json.dump(output, OUT)
# Add the consensus visualization
snvs, cnas, sampleToType = mutations
consensus_ccs = [d['core'] + d['expansion'] for d in consensus]
consensus_auto_delta = 0
results = [[consensus_ccs, consensus_stats, consensus_auto_delta]]
with open('{}/consensus/viz-data.json'.format(args.output_directory),
'w') as OUT:
output = hnviz.generate_viz_json(results, G.edges(), network_name,
heat, snvs, cnas, sampleToType,
d_score, d_name)
output['params'] = dict(consensus=True,
auto_delta=format(consensus_auto_delta, 'g'))
json.dump(output, OUT)
