plt.subplot(212)
plt.title('Random KD-OE pairs - wtcs distribution')
plt.hist(csRandom, 30)
plt.show()
#plot ranks
plt.subplot(211)
plt.hist(rnkValues, 30)
plt.title('observed KD-OE - percent rank distribution')
plt.subplot(212)
plt.title('Random KD-OE pairs - percent rank distribution')
plt.hist(rnkRandom, 30)
plt.show()
reload(dgo)
dg = dgo.QueryTargetAnalysis(out=work_dir)
# dg.add_dictionary(targetDict=targetDict)
dg.make_result_frames(gp_type='KD', metric='wtcs')
##
cellDirs = [
f for f in os.listdir(work_dir) if os.path.isdir(work_dir + '/' + f)
]
prog = progress.DeterminateProgressBar('dataframe read')
df = pd.DataFrame()
dfRank = pd.DataFrame()
#loop through each cell line add to df
for icell, cell1 in enumerate(cellDirs):
#define directories and load in outputs
outdir = os.path.join(work_dir, cell1, 'sig_query_out')
if not glob.glob(outdir + '/result_*.gctx'):
for pert in targetDict:
for gene in targetDict[pert]:
if gene in DBcgsOverlap:
if targetDictCGS.has_key(pert):
targetDictCGS[pert].append(gene)
else:
targetDictCGS[pert] = [gene]
#list of targets with CGS:
OverlapTargets = [
item for sublist in targetDictCGS.values() for item in sublist
]
test1 = 'OEB001_A375_96H:BRDN0000399163:-666' #set random sig_id to initialize dgo object
test2 = 'OEB001_A375_96H:BRDN0000400484:-666'
reload(dgo)
dg = dgo.QueryTargetAnalysis(test1, test2, work_dir + '/drug_gene')
dg.add_dictionary(targetDict=targetDictCGS)
# dg.get_drug_kd_sig_ids()
# dg.run_drug_gene_query()
dg.make_result_frames()
dg.test_known_connections(pDescDict=pDescDict)
dg.FDR_correction(pDescDict=pDescDict)
### test rand vec time
# n_obs = 44
# n_rand = 1000000
# permMtrx = np.random.rand(n_obs,n_rand)
# nullDist = permMtrx.prod(axis=0)
# testStat = nullDist[4]
# exVals = nullDist[nullDist<testStat]
# plt.hist(nullDist,bins=np.logspace(-18, 0, 50))
targetDictCGS = {}
for pert in targetDict:
for gene in targetDict[pert]:
if gene in DBcgsOverlap:
if targetDictCGS.has_key(pert):
targetDictCGS[pert].append(gene)
else:
targetDictCGS[pert] = [gene]
#list of targets with CGS:
OverlapTargets = [
item for sublist in targetDictCGS.values() for item in sublist
]
### test KD - two sided
reload(dgo)
dg = dgo.QueryTargetAnalysis(out=work_dir + '/drug_KD_connection')
dg.add_dictionary(targetDict=targetDictCGS)
# dg.get_sig_ids(genomic_pert='KD',is_gold=True)
# dg.run_drug_gene_query(metric='spearman',max_processes=10)
# #wait until queries finish
dg.make_result_frames(gp_type='KD', metric='spearman')
# dg.test_known_connections(gp_type='KD',metric='spearman',pDescDict=pDescDict,make_graphs=True)
# dg.FDR_correction(pDescDict=pDescDict,metric='spearman',outName='apriori_connections_pass_FDR',alpha=0.2,make_graphs=False)
dg.test_known_connections(gp_type='KD',
metric='spearman',
pDescDict=pDescDict,
outName='two_sided_dg_graphs',
make_graphs=False,
n_rand=1000000,
connection_test='two_sided')
dg.FDR_correction(pDescDict=pDescDict,
CM = mutil.CMapMongo()
# pert_List = CM.find({'sig_id':{'$regex':'DOSBIO'},'pert_iname':{'$regex':'BRD'}},{'sig_id':True,'pert_id':True,'pert_iname':True})
pert_List = CM.find(
{
'sig_id': {
'$regex': 'DOSBIO'
},
'pert_iname': {
'$regex': 'BRD'
}
}, {'pert_id': True})
dosbioSet = set(pert_List)
# check to make sure the brds are DOS compounds and don't represent known compounds
inameDict = {}
for brd in dosbioSet:
inames = CM.find({'pert_id': brd}, {'pert_iname': True})
inameSet = set(inames)
inameDict[brd] = inameSet
### test KD
reload(dgo)
dg = dgo.QueryTargetAnalysis(out=work_dir + '/drug_KD_spearman')
# dg.add_dictionary(targetDict=targetDict)
dg.get_sig_ids(genomic_pert='KD',
targetDict_loaded=False,
pert_list=list(dosbioSet),
is_gold=True)
dg.run_drug_gene_query(metric='spearman', max_processes=10)
# #wait until queries finish
dg.make_result_frames(gp_type='KD', metric='spearman')
# Query instances of BMS cps
CM = mu.CMapMongo()
pert_Q = CM.find({'pert_iname':{'$regex':'BMS'},'pert_type':'trt_cp'},{'sig_id':True,'pert_iname':True,'pert_id':True,})
#make brd-iname dictionary
pDescDict = {}
for sig in pert_Q:
pDescDict[sig['pert_id']] = sig['pert_iname']
pert_list = pDescDict.keys()
### run dgo object
test1 = 'OEB001_A375_96H:BRDN0000399163:-666' #set random sig_id to initialize dgo object
test2 = 'OEB001_A375_96H:BRDN0000400484:-666'
reload(dgo)
dg = dgo.QueryTargetAnalysis(test1,test2,work_dir + '/drug_KD_connection')
dg.add_dictionary(targetDict=targetDict)
# dg.get_sig_ids(genomic_pert='KD',targetDict_loaded=False,pert_list=pert_list)
# dg.run_drug_gene_query(max_processes=10)
# #wait until queries finish
dg.make_result_frames()
dg.test_unknown_connections(gp_type='KD',pDescDict=pDescDict)
# dg.FDR_correction(pDescDict=pDescDict)
### run OE analysis
reload(dgo)
dg = dgo.QueryTargetAnalysis(test1,test2,work_dir + '/drug_OE_connection')
# dg.add_dictionary(targetDict=targetDict)
dg.get_sig_ids(genomic_pert='OE',targetDict_loaded=False,pert_list=pert_list)
dg.run_drug_gene_query(max_processes=10)
# #wait until queries finish
for i, line in enumerate(splt):
splt2 = line.split('\t')
pID = splt2[0] #the pert_id listed the line
pDesc = splt2[1]
targets = splt2[2]
targets = targets.split(';')
targets = [x for x in targets if x != '']
if targets[0] == '' or targets[0] == '?' or targets[0] == '-666':
continue
else:
targetDict[pID] = targets
pDescDict[pID] = pDesc
### test KD
reload(dgo)
dg = dgo.QueryTargetAnalysis(out=work_dir + '/drug_KD_spearman')
dg.add_dictionary(targetDict=targetDict)
# dg.get_sig_ids(genomic_pert='KD',is_gold=True)
# dg.run_drug_gene_query(metric='spearman',max_processes=10)
# #wait until queries finish
dg.make_result_frames(gp_type='KD', metric='spearman')
# dg.test_known_connections(gp_type='KD',metric='spearman',pDescDict=pDescDict,make_graphs=True)
# dg.FDR_correction(pDescDict=pDescDict,metric='spearman',outName='apriori_connections_pass_FDR',alpha=0.2,make_graphs=False)
dg.test_known_connections(gp_type='KD',
metric='spearman',
pDescDict=pDescDict,
outName='test_dg_graphs2',
conn_thresh=.05,
make_graphs=True,
n_rand=100000,
connection_test='two_sided')
'BRD-A69592287': 'oxetane',
'BRD-A70150975': 'hydroxyl'
}
goiTested = [
'PIK3CA', 'PIK3CB', 'GAPDH', 'AKT1', 'AKT2', 'MTOR', 'ALDOA', 'NFKB1',
'MYC'
]
# create target dictionary for genes of interest
targetDict = {}
for brd in avicinsBrds:
targetDict[brd] = goiTested
### test KD
dg = dgo.QueryTargetAnalysis(out=work_dir + '/KD_spearman_all_doses')
dg.add_dictionary(targetDict=targetDict)
dg.get_sig_ids(genomic_pert='KD', is_gold=True)
dg.run_drug_gene_query(metric='spearman', max_processes=10)
# #wait until queries finish
dg.make_result_frames(gp_type='KD', metric='spearman')
dg.test_known_connections(gp_type='KD',
metric='spearman',
pDescDict=pDescDict,
outName='test_dg_graphs2',
conn_thresh=.05,
make_graphs=True,
n_rand=100000,
connection_test='two_sided')
dg.FDR_correction(pDescDict=pDescDict,
gp_type='KD',
