def test_volcano_plot():
an = ANOVA(ic50_test)
an.features.df = an.features.df[an.features.df.columns[0:10]]
an = ANOVA(ic50_test, genomic_features=an.features.df)
results = an.anova_all()
# try the constructors
v = VolcanoANOVA(results.df)
v = VolcanoANOVA(results)
# the selector metho
v.df = v.selector(v.df)
v.settings.savefig = False
# some of the plotting
v.volcano_plot_all_drugs()
v.volcano_plot_all_features()
v.volcano_plot_all()
v._get_fdr_from_pvalue_interp(1e-10)
v._get_pvalue_from_fdr(50)
v._get_pvalue_from_fdr([50,60])
def test_get_boxplot_data():
an = ANOVA(ic50_test)
odof = an._get_one_drug_one_feature_data('Drug_1047_IC50','TP53_mut')
bb = BoxPlots(odof)
data = bb._get_boxplot_data(mode='msi')
assert data[1] == ['***MSI-stable neg', '***MSI-stable pos',
'**MSI-unstable neg', '**MSI-unstable pos']
expected = [2.0108071495663922e-47, 0.0012564798887037905]
assert_list_almost_equal([data[2][0], data[2][1]], expected)
def test_get_boxplot_data():
an = ANOVA(ic50_test)
odof = an._get_one_drug_one_feature_data(1047, 'TP53_mut')
bb = BoxPlots(odof)
data = bb._get_boxplot_data(mode='msi')
assert data[1] == [
'***MSI-stable neg', '***MSI-stable pos', '**MSI-unstable neg',
'**MSI-unstable pos'
]
expected = [2.0108071495663922e-47, 0.0012564798887037905]
assert_list_almost_equal([data[2][0], data[2][1]], expected)
def test_IC50Cluster():
dataset = gdsctools_data("test_v18_clustering.tsv")
ic50 = IC50Cluster(dataset)
# In this data set, a drug is reported 3 times (1211) and should appear
# as follows:
assert 1211 in ic50.df.columns
assert 11211 in ic50.df.columns
assert 21211 in ic50.df.columns
assert len(ic50.drugIds) == 860
assert len(ic50.df) == 50
an = ANOVA(ic50, dataset)
an.diagnostics()['feasible_tests'] == 65026
def test_IC50Cluster():
dataset = gdsctools_data("test_v18_clustering.tsv")
ic50 = IC50Cluster(dataset)
# In this data set, a drug is reported 3 times (1211) and should appear
# as follows:
assert 1211 in ic50.df.columns
assert 11211 in ic50.df.columns
assert 21211 in ic50.df.columns
assert len(ic50.drugIds) == 860
assert len(ic50.df) == 50
an = ANOVA(ic50, dataset)
an.diagnostics()['feasible_tests'] == 65026
def anova_pipeline(args=None):
"""This function is used by the standalone application called
**gdsctools_anova**
Type::
gdsctools_anova --help
to get some help.
"""
msg = "Welcome to GDSCTools standalone"
print_color(msg, purple, underline=True)
# Keep the argument args as None by default to
# allow testing e.g., in nosetests
if args is None:
args = sys.argv[:]
elif len(args) == 1:
args += ['--help']
user_options = ANOVAOptions(prog="gdsctools_anova")
try:
options = user_options.parse_args(args[1:])
except SystemExit:
return
# -----------------------------------------------------------------
# ---------------------------------------- options without analysis
# -----------------------------------------------------------------
if options.version is True:
print("This is version %s of gdsctools_anova" % gdsctools.version)
return
if options.testing is True:
print('Testing mode:')
from gdsctools import ANOVA, ic50_test
an = ANOVA(ic50_test)
df = an.anova_one_drug_one_feature('Drug_1047_IC50', 'TP53_mut')
assert df.loc[1,'N_FEATURE_pos'] == 554, \
"N_feature_pos must be equal to 554"
print(df.T)
print(darkgreen("\nGDSCTools seems to be installed properly"))
return
if options.save_settings:
from gdsctools import ANOVA, ic50_test
an = ANOVA(ic50_test)
an.settings.to_json(options.save_settings)
print('Save a default parameter set in %s' % options.save_settings)
return
if options.license is True:
print(gdsctools.license)
return
if options.summary is True:
from gdsctools import anova
an = anova.ANOVA(options.input_ic50, options.input_features)
print(an)
return
if options.print_tissues is True:
from gdsctools import anova
an = anova.ANOVA(options.input_ic50, options.input_features)
tissues = an.tissue_factor
try:
tissues = tissues.sort_values('Tissue Factor').unique()
except:
tissues = tissues.sort(inplace=False).unique()
for name in tissues:
print(name)
return
if options.print_drugs is True:
from gdsctools import anova
gdsc = anova.ANOVA(options.input_ic50, options.input_features)
import textwrap
print("\n".join(textwrap.wrap(" , ".join(gdsc.drugIds))))
return
if options.print_features is True:
from gdsctools import anova
gdsc = anova.ANOVA(options.input_ic50, options.input_features)
import textwrap
print("\n".join(textwrap.wrap(" , ".join(gdsc.feature_names))))
return
# -----------------------------------------------------------------
# --------------------------------------------------- real analysis
# -----------------------------------------------------------------
# dispatcher to the functions according to the user parameters
from gdsctools import ANOVA, ANOVAReport
anova = ANOVA(options.input_ic50, options.input_features,
options.input_drug,
low_memory=not options.fast)
anova = _set_settings(anova, options)
if options.drug and options.drug not in anova.ic50.df.columns:
print(red("Invalid Drug. Try --print-drug-names"))
sys.exit(1)
if options.drug is not None and options.feature is not None:
print_color("ODOF mode", purple)
anova_one_drug_one_feature(anova, options)
elif options.drug is not None:
print_color("ODAF mode", purple)
anova_one_drug(anova, options)
else: # analyse everything
if options.feature is None:
print_color("ADAF mode", purple)
else:
print_color("ADOF mode", purple)
anova_all(anova, options)
if options.onweb is False and options.no_html is False:
msg = "\nNote that a directory {} was created and files saved into it"
print(purple(msg.format(options.directory)))
return
""" Analyse all associations (drug/feature) ========================================= Volcano plot (all associations) """ ##################################################### # from gdsctools import ANOVA, ic50_test gdsc = ANOVA(ic50_test) results = gdsc.anova_all() results.volcano()
def test_anova_brca():
an1 = ANOVA(gdsctools_data('IC50_v17.csv.gz'))
an1.set_cancer_type('breast')
an = ANOVA(an1.ic50, gdsctools_data('GF_BRCA_v17.csv.gz'))
dfori = an.anova_all()
df = dfori.df.sum()
df = df.drop(['DRUG_TARGET', 'DRUG_NAME', 'DRUG_ID', 'FEATURE'])
df = df.fillna(0)
totest = df.to_dict()
exact = {'ANOVA_FEATURE_FDR': 1133416.7761055394,
'ANOVA_FEATURE_pval': 5824.8201538614458,
'FEATURE_IC50_T_pval': 5824.8201538614449,
'FEATURE_IC50_effect_size': 4408.511449781573,
'FEATURE_delta_MEAN_IC50': 261.11373729866705,
'FEATURE_neg_Glass_delta': 4487.7401723134735,
'FEATURE_neg_IC50_sd': 14701.868130868914,
'FEATURE_neg_logIC50_MEAN': 28701.510736736222,
'FEATURE_pos_Glass_delta': 6536.8938399490198,
'FEATURE_pos_IC50_sd': 13362.588398939894,
'FEATURE_pos_logIC50_MEAN': 28962.624474034845,
'ANOVA_MSI_pval': 0.0,
'N_FEATURE_neg': 439196.0,
'N_FEATURE_pos': 92140.0,
'ANOVA_TISSUE_pval': 0.0,
'ASSOC_ID': 68509365.0,
'index': 68497660.0}
for k, v in totest.items():
if k in ['ANOVA_MEDIA_pval']:
continue
assert_almost_equal(v, exact[k])
# test part of the report (summary section)
r = ANOVAReport(an, dfori)
totest = r.diagnostics().to_dict()
exact = {'text': {0: 'Type of analysis',
1: 'Total number of possible drug/feature associations',
2: 'Total number of ANOVA tests performed',
3: 'Percentage of tests performed',
4: '',
5: 'Total number of tested drugs',
6: 'Total number of genomic features used',
7: 'Total number of screened cell lines',
8: 'MicroSatellite instability included as factor',
9: '',
10: 'Total number of significant associations',
11: ' - sensitive',
12: ' - resistant',
13: 'p-value significance threshold',
14: 'FDR significance threshold',
15: 'Range of significant p-values',
16: 'Range of significant % FDRs'},
'value': {0: 'breast',
1: 13780,
2: 11705,
3: 84.94,
4: '',
5: 265,
6: 52,
7: 51,
8: False,
9: '',
10: 27,
11: 17,
12: 10,
13: 0.001,
14: 25,
15: '[2.098e-09, 0.0004356]',
16: '[0.002456 18.89]'}}
assert totest == exact
import shutil
shutil.rmtree('breast')
""" Analyse all associations (drug/feature) ========================================= Volcano plot (all associations) """ ##################################################### # from gdsctools import ANOVA, ic50_test gdsc = ANOVA(ic50_test) results = gdsc.anova_all() results.volcano()
def test_anova_brca():
an1 = ANOVA(gdsctools_data('IC50_v17.csv.gz'))
an1.set_cancer_type('breast')
an = ANOVA(an1.ic50, gdsctools_data('GF_BRCA_v17.csv.gz'))
dfori = an.anova_all()
df = dfori.df.sum()
df = df.drop(['DRUG_TARGET', 'DRUG_NAME', 'DRUG_ID', 'FEATURE'])
df = df.fillna(0)
totest = df.to_dict()
exact = {
'ANOVA_FEATURE_FDR': 1133416.7761055394,
'ANOVA_FEATURE_pval': 5824.8201538614458,
'FEATURE_IC50_T_pval': 5824.8201538614449,
'FEATURE_IC50_effect_size': 4408.511449781573,
'FEATURE_delta_MEAN_IC50': 261.11373729866705,
'FEATURE_neg_Glass_delta': 4487.7401723134735,
'FEATURE_neg_IC50_sd': 14701.868130868914,
'FEATURE_neg_logIC50_MEAN': 28701.510736736222,
'FEATURE_pos_Glass_delta': 6536.8938399490198,
'FEATURE_pos_IC50_sd': 13362.588398939894,
'FEATURE_pos_logIC50_MEAN': 28962.624474034845,
'ANOVA_MSI_pval': 0.0,
'N_FEATURE_neg': 439196.0,
'N_FEATURE_pos': 92140.0,
'ANOVA_TISSUE_pval': 0.0,
'ASSOC_ID': 68509365.0,
'index': 68497660.0
}
for k, v in totest.items():
if k in ['ANOVA_MEDIA_pval']:
continue
assert_almost_equal(v, exact[k])
# test part of the report (summary section)
r = ANOVAReport(an, dfori)
totest = r.diagnostics().to_dict()
exact = {
'text': {
0: 'Type of analysis',
1: 'Total number of possible drug/feature associations',
2: 'Total number of ANOVA tests performed',
3: 'Percentage of tests performed',
4: '',
5: 'Total number of tested drugs',
6: 'Total number of genomic features used',
7: 'Total number of screened cell lines',
8: 'MicroSatellite instability included as factor',
9: '',
10: 'Total number of significant associations',
11: ' - sensitive',
12: ' - resistant',
13: 'p-value significance threshold',
14: 'FDR significance threshold',
15: 'Range of significant p-values',
16: 'Range of significant % FDRs'
},
'value': {
0: 'breast',
1: 13780,
2: 11705,
3: 84.94,
4: '',
5: 265,
6: 52,
7: 51,
8: False,
9: '',
10: 27,
11: 17,
12: 10,
13: 0.001,
14: 25,
15: '[2.098e-09, 0.0004356]',
16: '[0.002456 18.89]'
}
}
assert totest == exact
"""
Association between a Drug and a Feature
=========================================
Boxplot association
"""
#####################################################
#
from gdsctools import ANOVA, ic50_test
gdsc = ANOVA(ic50_test)
gdsc.set_cancer_type('breast')
df = gdsc.anova_one_drug_one_feature(1047, 'TP53_mut', show=True)
def anova_pipeline(args=None):
"""This function is used by the standalone application called
**gdsctools_anova**
Type::
gdsctools_anova --help
to get some help.
"""
msg = "Welcome to GDSCTools standalone"
print_color(msg, purple, underline=True)
# Keep the argument args as None by default to
# allow testing e.g., in nosetests
if args is None:
args = sys.argv[:]
elif len(args) == 1:
args += ['--help']
user_options = ANOVAOptions(prog="gdsctools_anova")
try:
options = user_options.parse_args(args[1:])
except SystemExit:
return
# -----------------------------------------------------------------
# ---------------------------------------- options without analysis
# -----------------------------------------------------------------
if options.version is True:
print("This is version %s of gdsctools_anova" % gdsctools.version)
return
if options.testing is True:
print('Testing mode:')
from gdsctools import ANOVA, ic50_test
an = ANOVA(ic50_test)
df = an.anova_one_drug_one_feature('Drug_1047_IC50', 'TP53_mut')
assert df.loc[1,'N_FEATURE_pos'] == 554, \
"N_feature_pos must be equal to 554"
print(df.T)
print(darkgreen("\nGDSCTools seems to be installed properly"))
return
if options.save_settings:
from gdsctools import ANOVA, ic50_test
an = ANOVA(ic50_test)
an.settings.to_json(options.save_settings)
print('Save a default parameter set in %s' % options.save_settings)
return
if options.license is True:
print(gdsctools.license)
return
if options.summary is True:
from gdsctools import anova
an = anova.ANOVA(options.input_ic50, options.input_features)
print(an)
return
if options.print_tissues is True:
from gdsctools import anova
an = anova.ANOVA(options.input_ic50, options.input_features)
tissues = an.tissue_factor
try:
tissues = tissues.sort_values('Tissue Factor').unique()
except:
tissues = tissues.sort(inplace=False).unique()
for name in tissues:
print(name)
return
if options.print_drugs is True:
from gdsctools import anova
gdsc = anova.ANOVA(options.input_ic50, options.input_features)
import textwrap
print("\n".join(textwrap.wrap(" , ".join(gdsc.drugIds))))
return
if options.print_features is True:
from gdsctools import anova
gdsc = anova.ANOVA(options.input_ic50, options.input_features)
import textwrap
print("\n".join(textwrap.wrap(" , ".join(gdsc.feature_names))))
return
# -----------------------------------------------------------------
# --------------------------------------------------- real analysis
# -----------------------------------------------------------------
# dispatcher to the functions according to the user parameters
from gdsctools import ANOVA, ANOVAReport
anova = ANOVA(options.input_ic50,
options.input_features,
options.input_drug,
low_memory=not options.fast)
anova = _set_settings(anova, options)
if options.drug and options.drug not in anova.ic50.df.columns:
print(red("Invalid Drug. Try --print-drug-names"))
sys.exit(1)
if options.drug is not None and options.feature is not None:
print_color("ODOF mode", purple)
anova_one_drug_one_feature(anova, options)
elif options.drug is not None:
print_color("ODAF mode", purple)
anova_one_drug(anova, options)
else: # analyse everything
if options.feature is None:
print_color("ADAF mode", purple)
else:
print_color("ADOF mode", purple)
anova_all(anova, options)
if options.onweb is False and options.no_html is False:
msg = "\nNote that a directory {} was created and files saved into it"
print(purple(msg.format(options.directory)))
return
"""
Association between a Drug and a Feature
=========================================
Boxplot association
"""
#####################################################
#
from gdsctools import ANOVA, ic50_test
gdsc = ANOVA(ic50_test)
gdsc.set_cancer_type('breast')
df = gdsc.anova_one_drug_one_feature(1047, 'TP53_mut', show=True)