def main():
args = parser.parse_args()
beta = 0.5
#get mutational data
mvals = getMutationalData()
##load up interactome
gfile = args.graph
##TODO: replace this with Docker image call
g = pickle.load(open(gfile, 'rb'))
key = 'mpnstPDXmuts'
this_hyp = hyphalNetwork(mvals, g)
this_hyp._to_file(key + '_hypha.pkl')
##read from file
###this is all we need to do in a single eval, then we can do tests later
this_hyp.node_stats().to_csv(key + '_nodelist.csv')
for_e = hyEnrich.go_enrich_forests(this_hyp)
this_hyp.assign_enrichment(for_e, type='forest')
for_e.to_csv(key + 'enrichedForestGoTerms.csv')
com_e = hyEnrich.go_enrich_communities(this_hyp)
this_hyp.assign_enrichment(com_e, type='community')
this_hyp._to_file(key + '_hypha.pkl')
com_e.to_csv(key + 'enrichedCommunityGOterms.csv')
this_hyp.community_stats(prefix=key).to_csv(key + '_communityStats.csv')
res = hyStats.compute_all_distances({'mutations': this_hyp})
res.to_csv('panPDXDistances.csv')
nmi = hyStats.compute_all_nmi({'mutations': this_hyp}, g)
nmi.to_csv('panPDXNMI.csv')
def significant_genes(data_frame, group, subgroup, value):
data_frame['zscore'] = stats.zscore(data_frame[value])
#SG
#here you reference `data` and not `data frame`
#significant = data[abs(data['zscore']) >= 2.58]
#also altered to remove absolute value
significant = data_frame[data_frame['zscore'] >= 2.58]
#replaced with new function
#gene_dictionary = nested_dict(significant[[group, subgroup, value]])
#SG: updated this based on what i found on google....
gene_dictionary = (significant.groupby(group).apply(
lambda x: dict(zip(x[subgroup], x[value]))).to_dict())
#SG This is causing issue, it's old code g = hyp.make_graph_from_dict(gfile)
hyphae = dict()
beta = 0.5
#SG:
#this was the issue, it was an uncessary loop
#for key, val in gene_dictionary.items():
# hyphalNetwork is supposed to take a gene dictionary!
key = 'proteomics'
this_hyp = hyphalNetwork(gene_dictionary, g.copy(), beta)
hyphae[key] = this_hyp
this_hyp._to_file(key + '_hypha.pkl')
#print (hyphae)
return hyphae
def main():
args = parser.parse_args()
qval = args.quant
##first we run a helper function to make sure we have all cptac data
fdict = pdata.cptacData()
##first get proteomics measurements
allDat = pdata.getCancerData(fdict, qval, byType=False)
patientData = pdata.getCombinedClinicalData(fdict)
patientData.to_csv('clinicalData.csv')
mutationData = pdata.getCombinedMutationData(fdict)
mutationData.to_csv('mutationData.csv')
##make srue this file is built!
g = pickle.load(open('../odata/igraphPPI.pkl', 'rb'))
beta = .5
#build hyphal network of network communities
phyph = hyphalNetwork(allDat, g, beta)
phpyh._to_file(args.refName + '_hypNet.pkl')
#write out distances within communities
res = phyph.distVals
fname = args.refName + '_DistanceVals.csv'
res.to_csv(fname)
def main():
args = parser.parse_args()
beta = 0.5
#get mutational data
mdf = pd.read_csv(args.df)
key = args.output
mvals = df2dict(mdf)
##load up interactome
gfile = args.graph
##TODO: replace this with Docker image call
g = pickle.load(open(gfile, 'rb'))
this_hyp = hyphalNetwork(mvals, g)
this_hyp._to_file(key + '_hypha.pkl')
def significant_genes(data_frame, group, subgroup, value):
data_frame['zscore'] = stats.zscore(data_frame[value])
significant = data_frame[data_frame['zscore'] >= 2.58]
gene_dictionary = (significant.groupby(group).apply(
lambda x: dict(zip(x[subgroup], x[value]))).to_dict())
hyphae = dict()
beta = 0.5
key = 'proteomics'
this_hyp = hyphalNetwork(gene_dictionary, g.copy(), beta)
hyphae[key] = this_hyp
this_hyp._to_file(key + '_hypha.pkl')
return hyphae
def build_hyphae_from_data(qt, g, sample=False):
""" Temp function to load data from local directory"""
##this is the framework for the PDC data parser.
#now we want to build network communities for each
hyphae = dict()
patDiffs = loadCancerData(qt)
beta = 0.5
for key, vals in patDiffs.items():
if sample:
new_vals = {}
for v in random.sample(list(vals), 300):
new_vals[v] = vals[v]
vals = new_vals
print(len(vals))
this_hyp = hyphalNetwork(vals, g.copy(),beta=beta, g=3, do_forest=False, noComms=False)
hyphae[key+str(qt)] = this_hyp
this_hyp._to_file(key+str(qt)+'_hypha.pkl')
return hyphae
def build_hyphae_from_data():
""" Temp function to load data from local directory"""
##this is the framework for the PDC data parser.
norms = prot.normals_from_manifest('data/PDC_biospecimen_manifest_05112020_184928.csv')
# bcData = prot.parsePDCfile('data/TCGA_Breast_BI_Proteome.itraq.tsv')
bcData = prot.parsePDCfile('data/CPTAC2_Breast_Prospective_Collection_BI_Proteome.tmt10.tsv')
lungData = prot.parsePDCfile('data/CPTAC3_Lung_Adeno_Carcinoma_Proteome.tmt10.tsv')
colData = prot.parsePDCfile('data/CPTAC2_Colon_Prospective_Collection_PNNL_Proteome.tmt10.tsv')
gbmData = prot.parsePDCfile('data/CPTAC3_Glioblastoma_Multiforme_Proteome.tmt11.tsv')
normPats = {'brca': set([a for a in bcData['Patient'] if a in norms['Breast Invasive Carcinoma']]),\
'coad': set([a for a in colData['Patient'] if a in norms['Colon Adenocarcinoma']]),\
'luad': set([a for a in lungData['Patient'] if a in norms['Lung Adenocarcinoma']]),\
'gbm': set([a for a in gbmData['Patient'] if a in norms['Other']])}
gfile = '../../../OmicsIntegrator2/interactomes/inbiomap.9.12.2016.full.oi2'
g = hyp.make_graph(gfile)
namemapper = None #hyp.mapHGNCtoNetwork()
##here we get the top values for each patient
patVals = {'brca':prot.getProtsByPatient(bcData, namemapper),\
'luad':prot.getProtsByPatient(lungData, namemapper),\
'coad':prot.getProtsByPatient(colData, namemapper),\
'gbm':prot.getProtsByPatient(gbmData, namemapper)}
#here we get the top most distinguished from normals
patDiffs = {'brca': prot.getTumorNorm(bcData, normPats['brca'], namemapper),
'luad': prot.getTumorNorm(lungData, normPats['luad'], namemapper),
'coad': prot.getTumorNorm(colData, normPats['coad'], namemapper),
'gbm': prot.getTumorNorm(gbmData, normPats['gbm'], namemapper)}
#now we want to build network communities for each
hyphae = dict()
for key in patDiffs:
this_hyp = hyphalNetwork(patDiffs[key], g)
hyphae[key] = this_hyp
this_hyp._to_file(key+'_hypha.pkl')
return hyphae
def main():
args = parser.parse_args()
qval = args.quant
##first we run a helper function to make sure we have all cptac data
fdict = pdat.cptacData()
##first get proteomics measurements
allDat = pdat.getCancerData(fdict, qval, byType=True)
##make srue this file is built!
g = pickle.load(open('../odata/igraphPPI.pkl', 'rb'))
beta = .5
#build hyphal network of network communities
phyph = pickle.load(open(args.hyph, 'rb'))
hDict = {'panCan': phyph}
for ct, dat in allDat.items():
hDict[ct] = hyphalNetwork(dat, g, beta)
nmi = hyStats.compute_all_nmi(hDict, g)
nmi.to_csv(args.refName + '_nmi.csv')
def main():
gfile = '../../data/igraphPPI.pkl'
g = pickle.load(open(gfile, 'rb')) #hyp.make_graph_from_dict(gfile)
args = parser.parse_args()
beta = 0.5
proteomics_dictionary = significant_prots(data, 'AML sample', 'Gene',
'LogFoldChange')
gene_dictionary = tumor_genes(data, 'AML sample', 'Gene', 'Tumor VAF')
if args.fromFile is None:
hyphae = dict()
hyphae['mutations'] = hyphalNetwork(gene_dictionary, g.copy(), beta)
hyphae['proteomics'] = hyphalNetwork(proteomics_dictionary, g.copy(),
beta)
for key, this_hyp in hyphae.items():
this_hyp._to_file(key + '_amlPatientData_hypha.pkl')
else:
hyphae = loadFromFile(args.fromFile)
#now compute graph distances to ascertain fidelity
if args.getDist:
res = hyStats.compute_all_distances(hyphae)
res.to_csv('amlNetworkdistances.csv')
tab = table.build_table("AML Network Distances", 'syn22128879', res)
syn.store(tab)
nmi = hyStats.compute_all_nmi(hyphae, g)
nmi.to_csv('amlNMI.csv')
syn.store(File('amlNMI.csv', parent='syn22269875'))
#store distances
for key, this_hyp in hyphae.items():
node_stats = this_hyp.node_stats()
node_stats.to_csv(key + '_nodelist.csv')
tab = table.build_table("AML Network Nodes", 'syn22128879', node_stats)
syn.store(tab)
if args.doEnrich:
if len(this_hyp.forest_enrichment) == 0:
for_e = hyEnrich.go_enrich_forests(this_hyp) #SG, ncbi)
this_hyp.assign_enrichment(for_e, type='forest')
for_e.to_csv(key + 'enrichedForestGoTerms.csv')
syn.store(
File(key + 'enrichedForestGoTerms.csv',
parent='syn22269875'))
this_hyp._to_file(key + '_amlPatientData_hypha.pkl')
if len(this_hyp.community_enrichment) == 0:
com_e = hyEnrich.go_enrich_communities(this_hyp)
this_hyp.assign_enrichment(com_e, type='community')
com_e.to_csv(key + 'enrichedCommunityGOterms.csv')
syn.store(
File(key + 'enrichedCommunityGOterms.csv',
parent='syn22269875'))
this_hyp._to_file(key + '_amlPatientData_hypha.pkl')
##next: compare enrichment between patients mapped to communities
this_hyp.community_stats(prefix=key).to_csv(key +
'_communityStats.csv')
this_hyp.forest_stats().to_csv(key + '_TreeStats.csv')
for files in [
key + '_amlPatientData_hypha.pkl', key + '_communityStats.csv',
key + '_TreeStats.csv'
]:
syn.store(File(files, parent='syn22269875'))
