def plotter(df, option, DURATION, EVENT, CategoricalDtype):
kmf = KaplanMeierFitter()
fig, ax = plt.subplots(figsize=(10, 5))
T = df[DURATION]
E = df[EVENT]
if isinstance(df[option].dtype, CategoricalDtype):
unique_codes = list(set(df[option].cat.codes.values))
unique_codes.sort()
mapping = dict(zip(df[option].cat.codes.values, df[option].values))
kmf.fit(T, E)
for code in unique_codes:
subset = (df[option] == mapping[code])
kmf.fit(T[subset], event_observed=E[subset], label=mapping[code])
kmf.plot_survival_function(ax=ax)
else:
unique_codes = list(set(df[option].values))
unique_codes.sort()
kmf.fit(T, E)
for code in unique_codes:
subset = (df[option] == code)
kmf.fit(T[subset], event_observed=E[subset], label=code)
kmf.plot_survival_function(ax=ax)
# plt.title("Lifespans by " + option)
return plt
def kmftest(data, part):
kmf = KaplanMeierFitter()
kmf.fit(df['time'], event_observed=df['status'])
#filter dataset on part
dataframe = pd.read_json(data)
dataframe = dataframe[dataframe["sex"] == part]
#dataframe = dataframe[dataframe['part'] == part]
#TODO: change to duration later
T = dataframe['time']
#TODO: chagne this key to the right dataset.
E = dataframe['censored']
kmf.fit(data['time'], event_observed=data['status'])
kmf.plot_survival_function()
def plot(self):
"""
Plot side-by-side kaplan-meier of input datasets
"""
figsize = (10, 5)
fig, ax = plt.subplots(1, 2, figsize=figsize, sharey=True)
# sns.set(font_scale=1.5)
# sns.despine()
palette = ['#0d3d56', '#006887', '#0098b5', '#00cbde', '#00ffff']
datasets = [self.stats_original_, self.stats_synthetic_]
for data, label, ax_cur in zip(datasets, self.labels, ax):
t = data['time']
e = data['event']
kmf = KaplanMeierFitter()
groups = np.sort(data['group'].unique())
for g, color in zip(groups, palette):
mask = (data['group'] == g)
kmf.fit(t[mask], event_observed=e[mask], label=g)
ax_cur = kmf.plot_survival_function(ax=ax_cur, color=color)
ax_cur.legend(title=self.group_column)
ax_cur.set_title('Kaplan-Meier - {} data'.format(label))
ax_cur.set_ylim(0, 1)
plt.tight_layout()
def estimate_survival(df, plot=True, censoring='right', fig_path=None):
if censoring not in {'right', 'left'}:
raise ValueError(f"unknown fit type: {censoring},"
f" use one of {{'left', 'right'}}")
kmf = KaplanMeierFitter(alpha=1.0) # disable confidence interval
if plot:
fig = plt.figure(figsize=(20, 10))
ax = fig.add_subplot(111)
medians = {}
for system in sorted(df.domain_system.unique()):
if censoring == 'right':
kmf.fit(
df.loc[df['domain_system'] == system].time,
df.loc[df['domain_system'] == system].spotted,
label=system,
)
elif censoring == 'left':
kmf.fit_left_censoring(
df.loc[df['domain_system'] == system].time,
~df.loc[df['domain_system'] == system].spotted,
label=system,
)
else:
raise ValueError(f"unknown fit type: {censoring},"
f" use one of {{'left', 'right'}}")
if plot:
kmf.plot_survival_function(ax=ax)
medians[system] = kmf.median_survival_time_
if plot:
plt.ylim(0.0, 1.0)
plt.xlabel("Turns")
plt.ylabel("Survival Probability")
plt.title("Estimated Survival Function of different systems")
save_path = fig_path or "survival.png"
print(f'saving plot of estimated survival functions to: {save_path}')
plt.savefig(save_path)
return medians
def _PlotFigure(self, s_mean, s_std, x, time):
## Plot KM curve
from lifelines import CoxPHFitter, KaplanMeierFitter
fname = 'Data_Clinical_Encoded.csv'
df = pd.read_csv(fname)
df.dropna(axis=0, inplace=True)
df['event'] = 1 - (df['Count_as_OS'] == 'N')
print(df.head(5))
age_bool = (df['Age_@_Dx'] > self.params[0] - 7) & (df['Age_@_Dx'] <
self.params[0] + 7)
T_bool = (df['T' + str(self.params[1])] == 1)
N_bool = (df['N' + str(self.params[2])] == 1)
M_bool = (df['M' + str(self.params[3])] == 1)
grade_bool = (df['Grade' + str(self.params[4])] == 1)
ER_bool = (df['ER'] == self.params[5])
PR_bool = (df['PR'] == self.params[6])
HER2_bool = (df['Her2'] == self.params[7])
invade_bool = (df['Invasion'] == self.params[8])
size_bool = (df['size_precise'] > self.params[9] - 5) & (
df['size_precise'] < self.params[9] + 5)
nodes_bool = (df['nodespos'] > self.params[10] - 5) & (
df['nodespos'] < self.params[10] + 5)
df_subset = df[age_bool & T_bool & N_bool & M_bool & grade_bool
& ER_bool & PR_bool & HER2_bool & invade_bool]
print(len(df_subset), len(df))
kmf = KaplanMeierFitter()
kmf.fit(df_subset['Time_OS'], df_subset['event'], label="Train")
kmf.plot_survival_function(show_censors=False,
ci_show=True,
at_risk_counts=True)
titleString ='Age=' + str(self.params[0]) + ', T' + str(self.params[1]) + ', N'\
+ str(self.params[2]) + ', M' + str(self.params[3]) + ', Grade' + str(self.params[4])\
+ ', ER=' + str(self.params[5]) + ', PR=' + str(self.params[6]) + ', Her2=' + str(self.params[7])\
+ ', Invasion=' + str(self.params[8])
plt.title(titleString)
plt.plot(time, s_mean, 'r-', label='DeepSurv')
plt.fill_between(time, s_mean - s_std, s_mean + s_std, alpha=0.4)
plt.legend()
plt.tight_layout()
plt.savefig('survival_comparison')
plt.show()
def SevenPlot(stimes,Np,aps):
#ax = plt.subplot(111)
for x in range(3,10,1):
stime = stimes[0,x,:]
E = np.zeros(Np).astype(int)
for i in range(0,30,1):
if((stime[0,i]) > 62831):
E[i] = 0
else:
E[i] = 1
data1 = {'T':stime[0], 'E':E}
df = pd.DataFrame(data=data1)
T = df['T']
E = df['E']
kmf = KaplanMeierFitter()
kmf.fit(T,E, label = "ap = {}".format(aps[x]))
kmf.plot_survival_function(at_risk_counts = True)
plt.tight_layout()
plt.title("Survival functions as a function of Planetary Semimajor Axis (ap)")
from lifelines.datasets import load_waltons from lifelines import KaplanMeierFitter from lifelines.utils import median_survival_times df = load_waltons() print(df.head(),'\n') print(df['T'].min(), df['T'].max(),'\n') print(df['E'].value_counts(),'\n') print(df['group'].value_counts(),'\n') kmf = KaplanMeierFitter() kmf.fit(df['T'], event_observed=df['E']) kmf.plot_survival_function() median_ = kmf.median_survival_time_ median_confidence_interval_ = median_survival_times(kmf.confidence_interval_) print(median_confidence_interval_)
# %% [markdown]
# #### lifelines
# %%
from lifelines import KaplanMeierFitter
from lifelines.plotting import add_at_risk_counts
ax = plt.subplot(111)
kmfs = []
for status in dat['ER Status'].unique():
mask = dat['ER Status']==status
kmf = KaplanMeierFitter()
kmf.fit(dat["OS Time"][mask], event_observed = dat['OS event'][mask], label = status)
ax = kmf.plot_survival_function(ax=ax,ci_show=False)
kmfs.append(kmf)
add_at_risk_counts(*kmfs, ax=ax) # *kmfs expands to the elements of kmfs
plt.tight_layout()
# %% [markdown]
# Note that `lifelines` and `scikit-survival` leverage `matplotlib` and so the layers are having to be drawn as part of a for loop. This is typical `matplotlib` behavior.
#
# The `kaplanmeier` function basically packages up the `lifeline` Kaplan-Meier curve and makes some choices for you while creating the graph.
# %% [markdown]
# ## Question 1b
# %%
ax = plt.subplot(111)
SurvEst['S_est']) / (SurvEst['dt'] * SurvEst['S_est'])
SurvEst['cumhazard'] = SurvEst['hazard'].cumsum()
#dfC_KM['survived'] = len(dfC) - dfC_KM['E'] #total_events - dfC_KM['E']
#dfC_KM['KMfactor'] = 1 - dfC_KM['E'] / dfC_KM['survived']
# S_est = np.empty((len(dfC_KM)+1,))
# S_est[0] = 1
# for i,val in enumerate(dfC_KM.KMfactor.values):
# S_est[i+1] = S_est[i] * val
# Inspect result: compare to lifelines values
S_est = SurvEst['S_est'].values
print(np.allclose(kmf.survival_function_.KM_estimate.values, S_est))
t_KM = np.hstack((0, dfC_KM.index.values))
#plt.figure(figsize=(7,6))
#plt.plot(t_KM, kmf.survival_function_.KM_estimate.values, 'm-')
kmf.plot_survival_function(figsize=(7, 6))
#plt.plot(dfC_KM['T'].values, S_est, 'r.', alpha=.5, label='manual')
plt.plot(SurvEst['T'].values,
SurvEst['S_est'].values,
'r.:',
alpha=.6,
label='manual')
plt.legend()
kmf.cumulative_density_.plot(figsize=(7, 6))
plt.plot(SurvEst['T'].values,
(SurvEst['hazard'] * SurvEst['S_est']).cumsum().values,
'r.:',
alpha=.6,
label='manual')
plt.legend()
kmf_test = KaplanMeierFitter()
# In[26]:
figure = plt.figure(figsize=(12, 8), tight_layout=False)
ax = figure.add_subplot(111)
t = np.linspace(0, 84, 85)
kmf_train.fit(train['PFS'],
event_observed=train['disease_progress'],
timeline=t,
label='Train set')
ax = kmf_train.plot_survival_function(show_censors=True,
ci_force_lines=False,
ci_show=False,
ax=ax)
kmf_test.fit(test['PFS'],
event_observed=test['disease_progress'],
timeline=t,
label='Test set')
ax = kmf_test.plot_survival_function(show_censors=True,
ci_force_lines=False,
ci_show=False,
ax=ax)
add_at_risk_counts(kmf_train, kmf_test, ax=ax, fontsize=12)
ax.set_xlabel('Time (months)', fontsize=12, fontweight='bold')
ax.set_ylabel('Survival probability', fontsize=12, fontweight='bold')
using lifeline package
'''
# single category
from lifelines.datasets import load_waltons
df = load_waltons() # returns a Pandas DataFrame
T = df['T']
E = df['E']
from lifelines import KaplanMeierFitter
kmf = KaplanMeierFitter()
kmf.fit(T, E)
a = kmf.survival_function_
b = kmf.cumulative_density_
c = kmf.plot_survival_function(at_risk_counts=True)
# two categories
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
kmf = KaplanMeierFitter()
T, E = [], []
for name, grouped_df in df.groupby('group'):
T.append(grouped_df['T'].values)
E.append(grouped_df['E'].values)
kmf.fit(grouped_df["T"], grouped_df["E"], label=name)
kmf.plot_survival_function(ax=ax)
from lifelines.statistics import logrank_test
results = logrank_test(T[0], T[1], E[0], E[1])
ax.text(x=0.05,
y=0.05,
def CoxRegressionModel(stimes, N, ap_s, ap_s2, ap_s3, ap_s5, ebb_, eap_, ebs,
aps_0, mu, stime, Np):
#ebs = singular value = 0, 0.175, 0.35, 0.525, 0.7
#stimes = 1050 values ( 5 eb values x 7 ap values x 30 survival times)
#stime = 30 values for each aps_) value
#N = 1050
#Np = 30
#ap_s, ap_s2, ap_s3 = array of 1050 values used for the data frame
#*************************************EVENT OBSERVATION**************************************************#
E = np.zeros(N).astype(int)
for i, time in enumerate(stimes):
if (time > 62831):
E[i] = 0
else:
E[i] = 1
#**************************************MAKING A DATA FRAME************************************************#
data1 = {
'T': stimes,
'E': E,
'aps': ap_s,
'aps2': ap_s2,
'aps3': ap_s3,
'aps5': ap_s5,
'eap': eap_,
'eb': ebb_
}
df = pd.DataFrame(data=data1)
T = df['T']
E = df['E']
aps = df['aps']
aps2 = df['aps2']
aps3 = df['aps3']
aps5 = df['aps5']
eap = df['eap']
eb = df['eb']
#print(df)
#************************************COX PH FITTER*************************************#
fig, axes = plt.subplots()
axes.set_xscale('log')
axes.set_ylabel("S(t)")
KT, KE, Kdf = PlottingLL.PlottingLL(ebs, aps_0, mu, stime, Np)
kmf = KaplanMeierFitter().fit(KT, KE, label='KaplanMeierFitter')
kmf.plot_survival_function(ax=axes)
cph = CoxPHFitter()
#cph.fit(df,duration_col = 'T', event_col = 'E', formula = "eap")
#cph.fit(df,duration_col = 'T', event_col = 'E')
cph.fit(df, duration_col='T', event_col='E', formula="aps + I(aps**3)")
#cph.print_summary()
cph.plot_partial_effects_on_outcome(plot_baseline=False,
ax=axes,
cmap="coolwarm")
#cph.plot_partial_effects_on_outcome(covariates = ['aps'], values = [round(aps_0,3)], plot_baseline = False, ax = axes, cmap = "coolwarm")
cph.baseline_survival_.plot(ax=axes, ls=":", color=f"C{i}")
#cph.fit(df,duration_col = 'T', event_col = 'E', formula = "eb + aps + I(aps**3)")
#cph.print_summary()
#cph.plot_partial_effects_on_outcome(covariates = ['aps'], values = [round(aps_0,3)], ax = axes)
plt.title('Formula(aps vs. aps3 (1050 values)): (eb,ap,mu)={}'.format(
(round(ebs, 3), round(aps_0, 3), mu)),
fontsize=12)
def KM_age_groups(features, outcomes):
T = outcomes['duration_mortality']
E = outcomes['event_mortality']
features['Leeftijd (jaren)'] = features['Leeftijd (jaren)'].astype(float)
age_70_plus = features['Leeftijd (jaren)'] >= 70.
print('Alle patiënten: n=', len(age_70_plus))
print('Leeftijd ≥ 70: n=', sum(age_70_plus))
print('Leeftijd < 70: n=', sum(~age_70_plus))
was_not_on_icu = data[[
'Levend ontslagen en niet heropgenomen - waarvan niet opgenomen geweest op IC',
'Levend dag 21 maar nog in het ziekenhuis - niet op IC geweest',
'Dood op dag 21 - niet op IC geweest'
]].any(axis=1)
was_on_icu = data[[
'Levend ontslagen en niet heropgenomen - waarvan opgenomen geweest op IC',
'Levend dag 21 maar nog in het ziekenhuis - op IC geweest',
'Levend dag 21 maar nog in het ziekenhuis - waarvan nu nog op IC',
'Dood op dag 21 - op IC geweest'
]].any(axis=1)
print('Alle patiënten: n=',
(np.nansum(was_on_icu) + np.nansum(was_not_on_icu)))
print('ICU yes: n=', np.nansum(was_on_icu))
print('ICU no: n=', np.nansum(was_not_on_icu))
icu = False
fig, axes = plt.subplots(1, 1)
if icu:
kmf1 = KaplanMeierFitter().fit(T[was_on_icu],
E[was_on_icu],
label='Wel op ICU geweest')
kmf2 = KaplanMeierFitter().fit(T[was_not_on_icu],
E[was_not_on_icu],
label='Niet op ICU geweest')
else: # age 70
kmf1 = KaplanMeierFitter().fit(T[age_70_plus],
E[age_70_plus],
label='Leeftijd ≥ 70 jaar')
kmf2 = KaplanMeierFitter().fit(T[~age_70_plus],
E[~age_70_plus],
label='Leeftijd < 70 jaar')
kmf3 = KaplanMeierFitter().fit(T, E, label='Alle patienten')
if icu:
kmf1.plot_survival_function(color=c4)
kmf2.plot_survival_function(color=c5)
kmf3.plot_survival_function(color=c3)
else:
kmf1.plot_survival_function(color=c1)
kmf2.plot_survival_function(color=c2)
kmf3.plot_survival_function(color=c3)
axes.set_xticks([1, 5, 9, 13, 17, 21])
axes.set_xticklabels(['1', '5', '9', '13', '17', '21'])
axes.set_xlabel('Aantal dagen sinds opnamedag')
axes.set_ylabel('Proportie overlevend')
axes.set_xlim(0, 21)
axes.set_ylim(0, 1)
titledict = {
'fontsize': 18,
'fontweight': 'bold',
'verticalalignment': 'baseline',
'horizontalalignment': 'center'
}
# plt.title('COVID-PREDICT survival functie tot t = 21 dagen (n='+str(len(T))+')',fontdict=titledict)
plt.tight_layout()
if icu:
plt.savefig('KM_survival_curve_ICU.png',
format='png',
dpi=300,
figsize=(20, 20),
pad_inches=0,
bbox_inches='tight')
else:
plt.savefig('KM_survival_curve_DEATH.png',
format='png',
dpi=300,
figsize=(20, 20),
pad_inches=0,
bbox_inches='tight')
plt.show()
return kmf1, kmf2, kmf3
ofile.write("Marginal PGS Effect on Hematuria Risk P-value (LR Test): {0:.2e}\n".format(test_stats_PGS[1]))
ofile.write("Marginal PGSxP/LP Effect on Hematuria Risk P-value (LR Test): {0:.2e}\n".format(test_stats_pgs_int[1]))
ofile.write('\n\n')
ofile.write("#"*5+" Final Model "+'#'*5+'\n')
ofile.write(full_model.summary.to_string())
f, axis = plt.subplots(1, 1,figsize=(10,8))
f.tight_layout(pad=2)
axis.spines['right'].set_visible(False)
axis.spines['top'].set_visible(False)
kmf=KaplanMeierFitter()
kmf.fit(cph_table.loc[cph_table.Genotype=='Control']['ObsWindow'],event_observed=cph_table.loc[cph_table.Genotype=='Control']['Dx'], label='Control')
kmf.plot_survival_function(ax=axis,color=color_list[0],lw=2.0)
kmf.fit(cph_table.loc[cph_table.Genotype=='P/LP Carrier']['ObsWindow'],event_observed=cph_table.loc[cph_table.Genotype=='P/LP Carrier']['Dx'], label='P/LP Carrier')
kmf.plot_survival_function(ax=axis,color=color_list[4],lw=2.0)
axis.text(axis.get_xlim()[0]+0.25*(axis.get_xlim()[1]-axis.get_xlim()[0]),axis.get_ylim()[0]+0.35*(axis.get_ylim()[1]-axis.get_ylim()[0]),'P/LP P-value={0:.2e}'.format(test_stats_geno[1]),fontsize=20,fontweight='bold')
axis.set_title('Pack-Year Test Dataset (N={0:d})'.format(cph_table.shape[0]),fontsize=20,fontweight='bold')
axis.set_ylabel('Fraction of Patients\nUnaffected by\nPersistent Hematuria')
axis.set_xlabel('Patient Age')
plt.savefig(fig_direc+'Genotype_Hematuria_WithSmoking_KaplanMeier.svg')
plt.close()
ax.set_yticks([0.5, 0.6, 0.7, 0.8, 0.9, 1.0])
ax.set_xticklabels([
'4_feat.', "Prediction 1", 'Prediction 2', 'Prediction 3', 'Prediction 4'
],
ha='center')
#ax.set_xticklabels(['4_feat.', "Prediction 3", 'vd = 2%', 'vd = 2.5%',
# 'vd = 3%', 'vd = 3.8%', 'vd = 5%'], ha = 'center')
#%%
kmf = KaplanMeierFitter()
T = data['bio_rec_6_delay']
E = data['bio_rec_6']
kmf.fit(T, event_observed=E)
fig, ax = plt.subplots()
kmf.plot_survival_function(at_risk_counts=True)
plt.tight_layout()
#%%
#cov = 'age'
#nameCov = 'Age'
#unitCov = 'years'
#cov = 'tum_vol'
#nameCov = 'Tumor volume'
#unitCov = 'mm$^3$'
#cov = 'ADC_ave'
#nameCov = 'Average ADC'
#unitCov = ''
delimiter="\t")
regression_dataset['Class'] = regression_dataset['Class'].map({
'YES': 1,
'NO': 0
})
regression_dataset = regression_dataset.join(
pd.get_dummies(regression_dataset['clusters']))
regression_dataset = regression_dataset.drop(['id', 'clusters'], axis=1)
print(regression_dataset.head(100))
print(regression_dataset.dtypes)
kmf = KaplanMeierFitter()
kmf.fit(regression_dataset['Time'], event_observed=regression_dataset['Class'])
kmf.survival_function_
survival_function = kmf.plot_survival_function() # or just kmf.plot()
survival_function.get_figure().savefig("KaplanMeierFitter.png")
# Using Cox Proportional Hazards model
cph = CoxPHFitter(penalizer=0.1)
cph.fit(regression_dataset, 'Time', event_col='Class')
cph.print_summary()
# predict=cph.predict_survival_function(regression_dataset).plot()
# predict.get_figure().savefig("predicted.png")
ax = cph.plot()
ax.get_figure().savefig("CoxPHFitter.png")
# %% Grouped km using lifelines
from lifelines.plotting import add_at_risk_counts
T1 = df.query('group==1')['time']
T2 = df.query('group==2')['time']
C1 = df.query('group==1')['Died']
C2 = df.query('group==2')['Died']
fig, ax = plt.subplots()
kmf1 = KaplanMeierFitter()
kmf1.fit(T1, C1,label='Group 1',)
kmf2 = KaplanMeierFitter()
kmf2.fit(T2, C2,label = 'Group 2',)
ax = kmf1.plot_survival_function(ax = ax, ci_show=False)
ax = kmf2.plot_survival_function(ax = ax,
label = 'Group 2', ci_show=False)
add_at_risk_counts(kmf1, kmf2, ax=ax)
plt.tight_layout();
# %% [markdown]
# ## Kaplan-Meier curve using _lifelines_
#
# We can also format the axes to show percentages rather than proportions.
#
# We'll demo this with a single KM curve
# %%
import matplotlib.ticker as mtick
ax = kmf.plot()
def pipeline(rem_zeros,
compute,
make_plot,
seed,
dir,
method='ttest',
test_size=0.15):
# loads all data
os.chdir(dir)
labels = pd.read_csv('Label_selected.csv')
ndx = labels.index[
labels['label'] > -1].tolist() #gets indices of patients to use
lbl = labels.iloc[ndx, [
0, 4
]] #makes a new dataframe of patients to use (their IDs and survival response)
surv = labels.iloc[ndx, :]
genes = pd.read_csv('pr_coding_feats.csv')
genes = genes.set_index('ensembl_gene_id')
os.chdir(dir)
gene = pd.read_csv('gene.csv', index_col=0)
miRNA = pd.read_csv('miRNA.csv', index_col=0)
meth = pd.read_csv('meth.csv', index_col=0)
CNV = pd.read_csv('CNV.csv', index_col=0)
os.chdir(dir)
# optionally removes rows (features) that are all 0 across patients
if rem_zeros == True:
gene = gene.loc[~(gene == 0).all(axis=1)]
miRNA = miRNA.loc[~(miRNA == 0).all(axis=1)]
# splitting labels into train set and validation set
train_labels, test_labels, train_class, test_class = train_test_split(
lbl['case_id'], lbl, test_size=test_size, random_state=seed)
# removes features (rows) that have any na in them
meth = meth.dropna(axis='rows')
miRNA = miRNA.dropna(axis='rows')
gene = gene.dropna(axis='rows')
CNV = CNV.dropna(axis='rows')
# divides individual modalities into train and test sets based on same patient splits and transposes
miRNA_train = miRNA[train_labels].T
miRNA_test = miRNA[test_labels].T
gene_train = gene[train_labels].T
gene_test = gene[test_labels].T
CNV_train = CNV[train_labels].T
CNV_test = CNV[test_labels].T
meth_train = meth[train_labels].T
meth_test = meth[test_labels].T
# normalizing gene expression and miRNA datasets
miRNA_train_copy = pd.DataFrame(miRNA_train,
copy=True) # copies the original dataframe
miRNA_scaler = preprocessing.MinMaxScaler().fit(miRNA_train)
miRNA_train = miRNA_scaler.transform(miRNA_train)
miRNA_train = pd.DataFrame(miRNA_train,
columns=list(miRNA_train_copy)).set_index(
miRNA_train_copy.index.values)
miRNA_test_copy = pd.DataFrame(miRNA_test,
copy=True) # copies the original dataframe
miRNA_test = miRNA_scaler.transform(miRNA_test)
miRNA_test = pd.DataFrame(miRNA_test,
columns=list(miRNA_test_copy)).set_index(
miRNA_test_copy.index.values)
gene_train_copy = pd.DataFrame(gene_train,
copy=True) # copies the original dataframe
gene_scaler = preprocessing.MinMaxScaler().fit(gene_train)
gene_train = gene_scaler.transform(gene_train)
gene_train = pd.DataFrame(gene_train,
columns=list(gene_train_copy)).set_index(
gene_train_copy.index.values)
gene_test_copy = pd.DataFrame(gene_test,
copy=True) # copies the original dataframe
gene_test = gene_scaler.transform(gene_test)
gene_test = pd.DataFrame(gene_test,
columns=list(gene_test_copy)).set_index(
gene_test_copy.index.values)
train_class = train_class.set_index(
'case_id') # changes first column to be indices
test_class = test_class.set_index(
'case_id') # changes first column to be indices
# makes copies of the y dataframe because tr_ind alters it
train_class_copy1,train_class_copy2,train_class_copy3,train_class_copy4,train_class_copy5 = pd.DataFrame(train_class, copy=True),\
pd.DataFrame(train_class, copy=True),\
pd.DataFrame(train_class, copy=True),\
pd.DataFrame(train_class, copy=True),\
pd.DataFrame(train_class, copy=True)
if make_plot == 'valplot':
gen_curve(gene, lbl, 'gene', 10)
gen_curve(miRNA, lbl, 'miRNA', 10)
gen_curve(meth, lbl, 'meth', 10)
gen_curve(CNV, lbl, 'CNV', 10)
return
# makes copies of training data
miRNA_train_copy2 = pd.DataFrame(miRNA_train, copy=True)
meth_train_copy2 = pd.DataFrame(meth_train, copy=True)
CNV_train_copy2 = pd.DataFrame(CNV_train, copy=True)
gene_train_copy2 = pd.DataFrame(gene_train, copy=True)
if make_plot == 'fplot':
make_fplot = True
else:
make_fplot = False
if compute == 'recompute':
# Runs CV script to generate clfs w best parameters
clf_gene, fea_gene, _ = do_cv(gene_train, train_class_copy1, gene_test,
test_class, 'ttest', 'gene', 40, 2,
make_fplot)
clf_miRNA, fea_miRNA, _ = do_cv(miRNA_train, train_class_copy2,
miRNA_test, test_class, 'minfo',
'miRNA', 24, 2, make_fplot)
clf_meth, fea_meth, _ = do_cv(meth_train, train_class_copy3, meth_test,
test_class, 'minfo', 'meth', 60, 2,
make_fplot)
clf_CNV, fea_CNV, _ = do_cv(CNV_train, train_class_copy4, CNV_test,
test_class, 'minfo', 'CNV', 50, 2,
make_fplot)
elif compute == 'custom':
clf_gene, fea_gene, _ = do_cv(gene_train, train_class_copy1, gene_test,
test_class, method, 'gene', 100, 2,
make_fplot)
clf_miRNA, fea_miRNA, _ = do_cv(miRNA_train, train_class_copy2,
miRNA_test, test_class, method,
'miRNA', 100, 2, make_fplot)
clf_meth, fea_meth, _ = do_cv(meth_train, train_class_copy3, meth_test,
test_class, method, 'meth', 100, 2,
make_fplot)
clf_CNV, fea_CNV, _ = do_cv(CNV_train, train_class_copy4, CNV_test,
test_class, method, 'CNV', 100, 2,
make_fplot)
elif compute == 'precomputed':
# loads up the precomputed best classifiers and selected features
clf_gene = load('clf_gene4.joblib')
fea_gene = load('fea_gene4.joblib')
clf_meth = load('clf_meth4.joblib')
fea_meth = load('fea_meth4.joblib')
clf_CNV = load('clf_CNV4.joblib')
fea_CNV = load('fea_CNV4.joblib')
clf_miRNA = load('clf_miRNA4.joblib')
fea_miRNA = load('fea_miRNA4.joblib')
else:
return "enter a valid parameter for compute"
# shrinks test feature matrix to contain only selected features
miRNA_test = miRNA_test[fea_miRNA]
gene_test = gene_test[fea_gene]
meth_test = meth_test[fea_meth]
CNV_test = CNV_test[fea_CNV]
# gets acc results from predicting on test set with individual modalities
gene_ind_res = clf_gene.score(gene_test, test_class)
meth_ind_res = clf_meth.score(meth_test, test_class)
CNV_ind_res = clf_CNV.score(CNV_test, test_class)
miRNA_ind_res = clf_miRNA.score(miRNA_test, test_class)
# calculates auc for each modality
c1_gene, c2_gene, _ = roc_curve(
test_class.values.ravel(),
clf_gene.decision_function(gene_test).ravel())
c1_miRNA, c2_miRNA, _ = roc_curve(
test_class.values.ravel(),
clf_miRNA.decision_function(miRNA_test).ravel())
c1_CNV, c2_CNV, _ = roc_curve(test_class.values.ravel(),
clf_CNV.decision_function(CNV_test).ravel())
c1_meth, c2_meth, _ = roc_curve(
test_class.values.ravel(),
clf_meth.decision_function(meth_test).ravel())
area_gene = auc(c1_gene, c2_gene)
area_miRNA = auc(c1_miRNA, c2_miRNA)
area_CNV = auc(c1_CNV, c2_CNV)
area_meth = auc(c1_meth, c2_meth)
if make_plot == 'ROC_gene' or make_plot == 'ROC_miRNA' or make_plot == 'ROC_meth' or make_plot == 'ROC_CNV':
if make_plot == 'ROC_gene':
c1, c2, area = c1_gene, c2_gene, area_gene
elif make_plot == 'ROC_miRNA':
c1, c2, area = c1_miRNA, c2_miRNA, area_miRNA
elif make_plot == 'ROC_meth':
c1, c2, area = c1_meth, c2_meth, area_meth
elif make_plot == 'ROC_CNV':
c1, c2, area = c1_CNV, c2_CNV, area_CNV
plt.title('Receiver Operating Characteristic')
plt.plot(c1, c2, 'b', label='AUC = %0.2f' %
area) # change params to change modality
plt.legend(loc='lower right')
plt.plot([0, 1], [0, 1], 'r--')
plt.xlim([0, 1])
plt.ylim([0, 1])
plt.ylabel('True Positive Rate')
plt.xlabel('False Positive Rate')
plt.show()
# START OF DATA INTEGRATION
fins = []
areas = []
fin1s = []
combinations = [["meth"], ["miRNA"], ["gene"], ["CNV"], ["meth", "miRNA"],
["meth", "gene"], ["meth", "CNV"], ["miRNA", "gene"],
["miRNA", "CNV"], ["gene", "CNV"],
["meth", "miRNA", "gene"], ["meth", "miRNA", "CNV"],
["miRNA", "gene", "CNV"], ["meth", "gene", "CNV"],
["meth", "miRNA", "gene", "CNV"]]
# shrink training feature matricies to selected features
miRNA_train_copy2 = miRNA_train_copy2[fea_miRNA]
gene_train_copy2 = gene_train_copy2[fea_gene]
meth_train_copy2 = meth_train_copy2[fea_meth]
CNV_train_copy2 = CNV_train_copy2[fea_CNV]
# gets prediction probabilities for samples in the training data
pred_miRNA = clf_miRNA.predict_proba(miRNA_train_copy2)[:, 0]
pred_gene = clf_gene.predict_proba(gene_train_copy2)[:, 0]
pred_CNV = clf_CNV.predict_proba(CNV_train_copy2)[:, 0]
pred_meth = clf_meth.predict_proba(meth_train_copy2)[:, 0]
# creates dataframe with the prediction probabilities
new_feats = {
'sample': miRNA_train.index.values,
'miRNA': pred_miRNA,
'gene': pred_gene,
'meth': pred_meth,
'CNV': pred_CNV
}
new_feats = pd.DataFrame(data=new_feats)
new_feats = new_feats.set_index('sample')
clfs = []
cvals = []
for com in combinations:
print(com)
clf = tr_comb_grid(new_feats[com], train_class_copy5)
# clf = bayes(new_feats[com],train_class_copy5)
clfs.append(clf)
pred_miRNA = clf_miRNA.predict_proba(miRNA_test)[:, 0]
pred_gene = clf_gene.predict_proba(gene_test)[:, 0]
pred_CNV = clf_CNV.predict_proba(CNV_test)[:, 0]
pred_meth = clf_meth.predict_proba(meth_test)[:, 0]
# creates new feature matrix for test predictions
new_feats_val = {
'sample': miRNA_test.index.values,
'miRNA': pred_miRNA,
'gene': pred_gene,
'meth': pred_meth,
'CNV': pred_CNV
}
new_feats_val = pd.DataFrame(data=new_feats_val)
new_feats_val = new_feats_val.set_index('sample')
fin = clf.score(new_feats_val[com], test_class)
pred = clf.decision_function(new_feats_val[com])
c1, c2, _ = roc_curve(test_class.values.ravel(), pred.ravel())
area = auc(c1, c2)
cvals.append([c1, c2, area])
print('auc: ', area)
print('acc: ', fin)
areas.append(area)
fins.append(fin)
# substitutes list entries for single modalities that were changed during integration
fins[0] = meth_ind_res
fins[1] = miRNA_ind_res
fins[2] = gene_ind_res
fins[3] = CNV_ind_res
areas[0] = area_meth
areas[1] = area_miRNA
areas[2] = area_gene
areas[3] = area_CNV
clfs[0] = clf_meth
clfs[1] = clf_miRNA
clfs[2] = clf_gene
clfs[3] = clf_CNV
indx = np.argmax(fins)
tr_score = clfs[indx].score(new_feats[combinations[indx]],
train_class_copy5)
te_score = clfs[indx].score(new_feats_val[combinations[indx]], test_class)
clf = clfs[indx]
if make_plot == 'barplot':
n_groups = 15
fig, ax = plt.subplots()
index = np.arange(n_groups)
bar_width = 0.35
opacity = 0.4
rects2 = ax.bar(index,
areas,
bar_width,
alpha=opacity,
color='r',
label='auc')
rects3 = ax.bar(index + bar_width,
fins,
bar_width,
alpha=opacity,
color='b',
label='score')
ax.set_xlabel('Combination')
ax.set_ylabel('Scores')
ax.set_xticks(index + bar_width / 2)
ax.set_xticklabels([
'meth', 'miRNA', 'gene', 'CNV', 'meth\nmiRNA', 'meth\ngene',
'meth\nCNV', 'miRNA\ngene', 'miRNA\nCNV', 'gene\nCNV',
'meth\nmiRNA\ngene', 'meth\nmiRNA\nCNV', 'miRNA\ngene\nCNV',
'meth\ngene\nCNV', 'meth\nmiRNA\ngene\nCNV'
])
ax.legend()
fig.tight_layout()
plt.show()
elif make_plot == 'ROC_int':
plt.title('Receiver Operating Characteristic')
plt.plot(cvals[indx][0],
cvals[indx][1],
'b',
label='AUC = %0.2f' % cvals[indx][2])
plt.legend(loc='lower right')
plt.plot([0, 1], [0, 1], 'r--')
plt.xlim([0, 1])
plt.ylim([0, 1])
plt.ylabel('True Positive Rate')
plt.xlabel('False Positive Rate')
plt.show()
elif make_plot == 'KM' or make_plot == 'KM Gene' or make_plot == 'KM miRNA'\
or make_plot == 'KM Meth' or make_plot == 'KM CNV' or make_plot == 'KM Integrated':
ax = plt.subplot(111)
kmf = KaplanMeierFitter()
m = max(surv["days_to_death"])
fill_max = {"days_to_death": m}
surv = surv.fillna(value=fill_max)
T = surv["days_to_death"]
surv = surv.replace("alive", False)
surv = surv.replace("dead", True)
E = surv["vital_status"]
if make_plot == 'KM':
kmf.fit(T, event_observed=E)
kmf.plot()
class0 = (surv["label"] == 0)
kmf.fit(T[class0],
event_observed=E[class0],
label="Low Survival (<5 years)")
kmf.plot_survival_function(ax=ax, ci_show=False, fontsize=20)
kmf.fit(T[~class0],
event_observed=E[~class0],
label="High Survival (>= 5 years)")
kmf.plot_survival_function(ax=ax, ci_show=False, fontsize=20)
ax.set_xlabel("Duration (days)", fontsize=20)
ax.set_ylabel("Percent Alive", fontsize=20)
ax.set_title("Breast Cancer Kaplan Meier Survival Curve",
fontsize=32)
elif make_plot == "Kaplan Meier Integrated":
surv2 = surv.copy(True)
surv2 = surv2.loc[surv2["case_id"].isin(test_labels.values)]
surv2 = surv2.set_index("case_id")
surv2 = surv2.reindex(test_class.index.values)
prd = clf.predict(new_feats_val[["meth", "miRNA", "gene"]])
surv2["label_new"] = prd
count = 0
preds = clf.predict(new_feats_val[["meth", "miRNA", "gene"]])
for i in range(test_class.shape[0]):
if preds[i] == test_class.iloc[i, 0]:
count += 1
print(count)
T2 = surv2["days_to_death"]
E2 = surv2["vital_status"]
class02 = (surv2["label_new"] == 0)
kmf.fit(T2[class02],
event_observed=E2[class02],
label="Low Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
kmf.fit(T2[~class02],
event_observed=E2[~class02],
label="High Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
ax.set_xlabel("Duration (days)")
ax.set_ylabel("Percent Alive")
ax.set_title("Integrated Classifier Kaplan Meier Survival Curve")
ax.set_ylim((0, 1))
plt.show()
elif make_plot == 'KM Gene':
surv2 = surv.copy(True)
surv2 = surv2.loc[surv2["case_id"].isin(test_labels.values)]
surv2 = surv2.set_index("case_id")
surv2 = surv2.reindex(test_class.index.values)
prd = clf_gene.predict(gene_test.loc[:, fea_gene])
surv2["label_new"] = prd
count = 0
preds = clf_gene.predict(gene_test.loc[:, fea_gene])
for i in range(test_class.shape[0]):
if preds[i] == test_class.iloc[i, 0]:
count += 1
print(count)
T2 = surv2["days_to_death"]
E2 = surv2["vital_status"]
class02 = (surv2["label_new"] == 0)
kmf.fit(T2[class02],
event_observed=E2[class02],
label="Low Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
kmf.fit(T2[~class02],
event_observed=E2[~class02],
label="High Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
ax.set_xlabel("Duration (days)")
ax.set_ylabel("Percent Alive")
ax.set_title("Gene Expression Kaplan Meier Survival Curve")
ax.set_ylim((0, 1))
plt.show()
elif make_plot == 'KM miRNA':
surv2 = surv.copy(True)
surv2 = surv2.loc[surv2["case_id"].isin(test_labels.values)]
surv2 = surv2.set_index("case_id")
surv2 = surv2.reindex(test_class.index.values)
prd = clf_miRNA.predict(miRNA_test.loc[:, fea_miRNA])
surv2["label_new"] = prd
count = 0
preds = clf_miRNA.predict(miRNA_test.loc[:, fea_miRNA])
for i in range(test_class.shape[0]):
if preds[i] == test_class.iloc[i, 0]:
count += 1
print(count)
T2 = surv2["days_to_death"]
E2 = surv2["vital_status"]
class02 = (surv2["label_new"] == 0)
kmf.fit(T2[class02],
event_observed=E2[class02],
label="Low Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
kmf.fit(T2[~class02],
event_observed=E2[~class02],
label="High Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
ax.set_xlabel("Duration (days)")
ax.set_ylabel("Percent Alive")
ax.set_title("miRNA Expression Kaplan Meier Survival Curve")
ax.set_ylim((0, 1))
plt.show()
elif make_plot == 'KM Meth':
surv2 = surv.copy(True)
surv2 = surv2.loc[surv2["case_id"].isin(test_labels.values)]
surv2 = surv2.set_index("case_id")
surv2 = surv2.reindex(test_class.index.values)
prd = clf_meth.predict(meth_test.loc[:, fea_meth])
surv2["label_new"] = prd
count = 0
preds = clf_meth.predict(meth_test.loc[:, fea_meth])
for i in range(test_class.shape[0]):
if preds[i] == test_class.iloc[i, 0]:
count += 1
print(count)
T2 = surv2["days_to_death"]
E2 = surv2["vital_status"]
class02 = (surv2["label_new"] == 0)
kmf.fit(T2[class02],
event_observed=E2[class02],
label="Low Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
kmf.fit(T2[~class02],
event_observed=E2[~class02],
label="High Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
ax.set_xlabel("Duration (days)")
ax.set_ylabel("Percent Alive")
ax.set_title("DNA Methylation Kaplan Meier Survival Curve")
ax.set_ylim((0, 1))
plt.show()
elif make_plot == 'KM CNV':
surv2 = surv.copy(True)
surv2 = surv2.loc[surv2["case_id"].isin(test_labels.values)]
surv2 = surv2.set_index("case_id")
surv2 = surv2.reindex(test_class.index.values)
prd = clf_CNV.predict(CNV_test.loc[:, fea_CNV])
surv2["label_new"] = prd
count = 0
preds = clf_CNV.predict(CNV_test.loc[:, fea_CNV])
for i in range(test_class.shape[0]):
if preds[i] == test_class.iloc[i, 0]:
count += 1
print(count)
T2 = surv2["days_to_death"]
E2 = surv2["vital_status"]
class02 = (surv2["label_new"] == 0)
kmf.fit(T2[class02],
event_observed=E2[class02],
label="Low Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
kmf.fit(T2[~class02],
event_observed=E2[~class02],
label="High Survival")
kmf.plot_survival_function(ax=ax, ci_show=False)
ax.set_xlabel("Duration (days)")
ax.set_ylabel("Percent Alive")
ax.set_title("CNV Kaplan Meier Survival Curve")
ax.set_ylim((0, 1))
plt.show()
elif make_plot == 'Gene Distribution' or make_plot == 'miRNA Distribution' or make_plot == 'Meth Distribution' \
or make_plot == 'CNV Distribution':
if make_plot == 'Gene Distribution':
lbl = lbl.set_index("case_id")
boxframe = gene.loc[fea_gene, :].T
boxframe = boxframe.iloc[:, 0:10]
cols = boxframe.columns
g = genes.loc[cols, :]
boxframe = pd.concat([boxframe, lbl[["label"]]], axis=1)
c1_gene = boxframe[boxframe["label"] == 1]
c1_gene = c1_gene.iloc[:, 0:10]
c2_gene = boxframe[boxframe["label"] == 0]
c2_gene = c2_gene.iloc[:, 0:10]
for i in range(0, 10):
cl = ['High', 'Low']
ind = [1, 2]
plt.subplot(2, 5, i + 1)
p1 = c1_gene.iloc[:, i]
p2 = c2_gene.iloc[:, i]
genes_to_plot = (p1, p2)
plt.boxplot(genes_to_plot)
plt.ylabel('Gene Expression (FPKM)', fontsize=16)
if i > 4:
plt.xlabel('Survival Class', fontsize=16)
plt.xticks(ind, cl, fontsize=16)
max = 0
max1 = (p1.values.max())
max2 = (p2.values.max())
if max1 > max2:
max = max1
else:
max = max2
plt.yticks([0, max], rotation='vertical')
# print(cols[i])
# title_str = 'Survival Breakdown of CNV Features by Gain/Loss: {}'
# title_str = title_str.format(cols[i])
plt.title(genes.loc[cols[i], 'hgnc_symbol'], fontsize=18)
plt.suptitle('Top 10 Gene Feature Distributions by Survival Class',
fontsize=22)
plt.show()
elif make_plot == 'miRNA Distribution':
lbl = lbl.set_index("case_id")
boxframe = miRNA.loc[fea_miRNA, :].T
boxframe = boxframe.iloc[:, 0:10]
cols = boxframe.columns
boxframe = pd.concat([boxframe, lbl[["label"]]], axis=1)
c1_miRNA = boxframe[boxframe["label"] == 1]
c1_miRNA = c1_miRNA.iloc[:, 0:10]
c2_miRNA = boxframe[boxframe["label"] == 0]
c2_miRNA = c2_miRNA.iloc[:, 0:10]
for i in range(0, 10):
cl = ['High', 'Low']
ind = [1, 2]
plt.subplot(2, 5, i + 1)
p1 = c1_miRNA.iloc[:, i]
p2 = c2_miRNA.iloc[:, i]
genes_to_plot = (p1, p2)
plt.boxplot(genes_to_plot)
plt.ylabel('miRNA Expression (RPM)', fontsize=16)
if i > 4:
plt.xlabel('Survival Class', fontsize=16)
plt.xticks(ind, cl, fontsize=16)
max = 0
max1 = (p1.values.max())
max2 = (p2.values.max())
if max1 > max2:
max = max1
else:
max = max2
plt.yticks([0, max], rotation='vertical')
plt.title(cols[i], fontsize=18)
plt.suptitle(
'Top 10 miRNA Feature Distributions by Survival Class',
fontsize=22)
plt.show()
elif make_plot == 'Meth Distribution':
lbl = lbl.set_index("case_id")
boxframe = meth.loc[fea_meth, :].T
boxframe = boxframe.iloc[:, 0:10]
cols = boxframe.columns
boxframe = pd.concat([boxframe, lbl[["label"]]], axis=1)
c1_meth = boxframe[boxframe["label"] == 1]
c1_mmeth = c1_meth.iloc[:, 0:10]
c2_meth = boxframe[boxframe["label"] == 0]
c2_meth = c2_meth.iloc[:, 0:10]
for i in range(0, 10):
cl = ['High', 'Low']
ind = [1, 2]
plt.subplot(2, 5, i + 1)
p1 = c1_meth.iloc[:, i]
p2 = c2_meth.iloc[:, i]
genes_to_plot = (p1, p2)
plt.boxplot(genes_to_plot)
plt.ylabel('DNA Methylation (Beta Value)', fontsize=16)
if i > 4:
plt.xlabel('Survival Class', fontsize=16)
plt.xticks(ind, cl, fontsize=16)
max = 0
max1 = (p1.values.max())
max2 = (p2.values.max())
if max1 > max2:
max = max1
else:
max = max2
plt.yticks([0, max], rotation='vertical')
plt.title(cols[i], fontsize=18)
plt.suptitle(
'Top 10 DNA Methylation Feature Distributions by Survival Class',
fontsize=22)
plt.show()
elif make_plot == 'CNV Distribution':
lbl = lbl.set_index("case_id")
cnv_lbl = pd.concat([CNV.loc[fea_CNV, :].T, lbl["label"]], axis=1)
c1_CNV = cnv_lbl[cnv_lbl["label"] == 1]
c1_CNV = c1_CNV.iloc[:, 0:16]
cols = c1_CNV.columns.values
c2_CNV = cnv_lbl[cnv_lbl["label"] == 0]
c2_CNV = c2_CNV.iloc[:, 0:16]
for i in range(0, 10):
c1_CNV_1 = c1_CNV.iloc[:, i][c1_CNV.iloc[:,
i] == 1].sum().sum()
c1_CNV_neg1 = -1 * (
c1_CNV.iloc[:, i][c1_CNV.iloc[:, i] == -1].sum().sum())
c1_CNV_0 = (247 * 1) - (c1_CNV_1 + c1_CNV_neg1)
c1_tot = (247 * 1)
c1_pct_1 = c1_CNV_1 / c1_tot
c1_pct_neg1 = c1_CNV_neg1 / c1_tot
c1_pct_0 = c1_CNV_0 / c1_tot
c2_CNV_1 = c2_CNV.iloc[:, i][c2_CNV.iloc[:,
i] == 1].sum().sum()
c2_CNV_neg1 = -1 * (
c2_CNV.iloc[:, i][c2_CNV.iloc[:, i] == -1].sum().sum())
c2_CNV_0 = (95 * 1) - (c2_CNV_1 + c2_CNV_neg1)
c2_tot = (95 * 1)
c2_pct_1 = c2_CNV_1 / c2_tot
c2_pct_neg1 = c2_CNV_neg1 / c2_tot
c2_pct_0 = c2_CNV_0 / c2_tot
ones = (c1_pct_1, c2_pct_1)
negones = (c1_pct_neg1, c2_pct_neg1)
zeros = (c1_pct_0, c2_pct_0)
hold = tuple(map(operator.add, zeros, negones))
cl = ['High', 'Low']
ind = [1, 2]
width = 0.5
plt.subplot(2, 5, i + 1)
p1 = plt.bar(ind, negones, width)
p2 = plt.bar(ind, zeros, width, bottom=negones)
p3 = plt.bar(ind, ones, width, bottom=hold)
if i == 0 or i == 5:
plt.ylabel('Percent of Top Features', fontsize=16)
plt.xticks(ind, cl, fontsize=16)
if i > 4:
plt.xlabel('Survival Class', fontsize=16)
plt.yticks(np.arange(0, 1, 0.1))
plt.legend((p1[0], p2[0], p3[0]), ('-1', '0', '1'))
plt.title(genes.loc[cols[i][0:15], 'hgnc_symbol'], fontsize=18)
plt.suptitle('Top 10 CNV Feature Distributions by Survival Class',
fontsize=22)
plt.show()
elif make_plot == 'Heat Map Gene' or make_plot == 'Heat Map miRNA' or make_plot == 'Heat Map Meth' \
or make_plot == 'Heat Map CNV':
if make_plot == 'Heat Map Gene':
clf_gene_h, fea_gene_h, heat_gene = do_cv(gene_train, train_class,
gene_test, test_class,
'ttest', 'gene', 100, 2)
n_c = 27
n_feat = 49
up_bound = 100
lin_kern = np.empty(
[n_c,
n_feat]) # c vals ( see in cv_2 ), # feats (n_max - 2 / 2)
poly_kern = np.empty([n_c, n_feat])
rbf_kern = np.empty([n_c, n_feat])
sig_kern = np.empty([n_c, n_feat])
linvec = []
polvec = []
rbfvec = []
sigvec = []
heat_gene_out = heat_gene[1:len(heat_gene), :]
for n, i in enumerate(heat_gene_out):
if n % 4 == 0:
linvec.append(i[3])
elif n % 4 == 1:
polvec.append(i[3])
elif n % 4 == 2:
rbfvec.append(i[3])
elif n % 4 == 3:
sigvec.append(i[3])
for n in range(0, n_feat):
for c in range(0, n_c):
lin_kern[c][n] = linvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
poly_kern[c][n] = polvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
rbf_kern[c][n] = rbfvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
sig_kern[c][n] = sigvec[c + (27 * n)]
ax = plt.subplot(111)
chosen = []
out_kern = None
if clf_gene_h.get_params(False)['kernel'] == 'linear':
im = plt.imshow(lin_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = linvec
out_kern = lin_kern.copy(True)
elif clf_gene_h.get_params(False)['kernel'] == 'rbf':
im = plt.imshow(rbf_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = rbfvec
out_kern = rbf_kern.copy(True)
elif clf_gene_h.get_params(False)['kernel'] == 'sigmoid':
im = plt.imshow(sig_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = sigvec
out_kern = sig_kern.copy(True)
elif clf_gene_h.get_params(False)['kernel'] == 'poly':
im = plt.imshow(poly_kern,
cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = polvec
out_kern = poly_kern.copy(True)
plt.colorbar()
plt.clim(min(chosen), max(chosen))
ax.set_xticks(np.arange(n_feat))
ax.set_yticks(np.arange(n_c))
ax.set_xticklabels(range(2, up_bound, 2), fontsize=12)
ax.set_yticklabels([
.001, .005, 0.1, .5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 15, 20, 25, 50, 75, 100, 150, 200, 250
],
fontsize=14)
ax.set_xlabel("Number of features", fontsize=20)
ax.set_ylabel("C value", fontsize=20)
plt.setp(ax.get_xticklabels(), rotation=90, rotation_mode="anchor")
for j in range(0, n_c):
for k in range(0, n_feat):
text = ax.text(k,
j,
'%.2f' % out_kern[j, k],
ha="center",
va="center",
color="w",
fontsize=6)
ax.set_title(
"Gene Expression Parameter Sensitivity (Accuracy) - %s kernel"
% clf_gene_h.get_params(False)['kernel'],
fontsize=24)
plt.show()
elif make_plot == 'Heat Map miRNA':
clf_miRNA_h, fea_miRNA_h, heat_miRNA = do_cv(
miRNA_train, train_class, miRNA_test, test_class, 'minfo',
'miRNA', 100, 2)
n_c = 27
n_feat = 49
up_bound = 100
lin_kern = np.empty(
[n_c,
n_feat]) # c vals ( see in cv_2 ), # feats (n_max - 2 / 2)
poly_kern = np.empty([n_c, n_feat])
rbf_kern = np.empty([n_c, n_feat])
sig_kern = np.empty([n_c, n_feat])
linvec = []
polvec = []
rbfvec = []
sigvec = []
heat_miRNA_out = heat_miRNA[1:len(heat_miRNA), :]
for n, i in enumerate(heat_miRNA_out):
if n % 4 == 0:
linvec.append(i[3])
elif n % 4 == 1:
polvec.append(i[3])
elif n % 4 == 2:
rbfvec.append(i[3])
elif n % 4 == 3:
sigvec.append(i[3])
for n in range(0, n_feat):
for c in range(0, n_c):
lin_kern[c][n] = linvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
poly_kern[c][n] = polvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
rbf_kern[c][n] = rbfvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
sig_kern[c][n] = sigvec[c + (27 * n)]
ax = plt.subplot(111)
chosen = []
out_kern = None
if clf_miRNA_h.get_params(False)['kernel'] == 'linear':
im = plt.imshow(lin_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = linvec
out_kern = lin_kern.copy(True)
elif clf_miRNA_h.get_params(False)['kernel'] == 'rbf':
im = plt.imshow(rbf_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = rbfvec
out_kern = rbf_kern.copy(True)
elif clf_miRNA_h.get_params(False)['kernel'] == 'sigmoid':
im = plt.imshow(sig_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = sigvec
out_kern = sig_kern.copy(True)
elif clf_miRNA_h.get_params(False)['kernel'] == 'poly':
im = plt.imshow(poly_kern,
cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = polvec
out_kern = poly_kern.copy(True)
plt.colorbar()
plt.clim(min(chosen), max(chosen))
ax.set_xticks(np.arange(n_feat))
ax.set_yticks(np.arange(n_c))
ax.set_xticklabels(range(2, up_bound, 2), fontsize=12)
ax.set_yticklabels([
.001, .005, 0.1, .5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 15, 20, 25, 50, 75, 100, 150, 200, 250
],
fontsize=14)
ax.set_xlabel("Number of features", fontsize=20)
ax.set_ylabel("C value", fontsize=20)
plt.setp(ax.get_xticklabels(), rotation=90, rotation_mode="anchor")
for j in range(0, n_c):
for k in range(0, n_feat):
text = ax.text(k,
j,
'%.2f' % out_kern[j, k],
ha="center",
va="center",
color="w",
fontsize=6)
ax.set_title(
"miRNA Expression Parameter Sensitivity (Accuracy) - %s kernel"
% clf_miRNA_h.get_params(False)['kernel'],
fontsize=24)
plt.show()
elif make_plot == 'Heat Map Meth':
clf_meth_h, fea_meth_h, heat_meth = do_cv(meth_train, train_class,
meth_test, test_class,
'minfo', 'meth', 100, 2)
n_c = 27
n_feat = 49
up_bound = 100
lin_kern = np.empty(
[n_c,
n_feat]) # c vals ( see in cv_2 ), # feats (n_max - 2 / 2)
poly_kern = np.empty([n_c, n_feat])
rbf_kern = np.empty([n_c, n_feat])
sig_kern = np.empty([n_c, n_feat])
linvec = []
polvec = []
rbfvec = []
sigvec = []
heat_meth_out = heat_meth[1:len(heat_meth), :]
for n, i in enumerate(heat_meth_out):
if n % 4 == 0:
linvec.append(i[3])
elif n % 4 == 1:
polvec.append(i[3])
elif n % 4 == 2:
rbfvec.append(i[3])
elif n % 4 == 3:
sigvec.append(i[3])
for n in range(0, n_feat):
for c in range(0, n_c):
lin_kern[c][n] = linvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
poly_kern[c][n] = polvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
rbf_kern[c][n] = rbfvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
sig_kern[c][n] = sigvec[c + (27 * n)]
ax = plt.subplot(111)
chosen = []
out_kern = None
if clf_meth_h.get_params(False)['kernel'] == 'linear':
im = plt.imshow(lin_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = linvec
out_kern = lin_kern.copy(True)
elif clf_meth_h.get_params(False)['kernel'] == 'rbf':
im = plt.imshow(rbf_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = rbfvec
out_kern = rbf_kern.copy(True)
elif clf_meth_h.get_params(False)['kernel'] == 'sigmoid':
im = plt.imshow(sig_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = sigvec
out_kern = sig_kern.copy(True)
elif clf_meth_h.get_params(False)['kernel'] == 'poly':
im = plt.imshow(poly_kern,
cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = polvec
out_kern = poly_kern.copy(True)
plt.colorbar()
plt.clim(min(chosen), max(chosen))
ax.set_xticks(np.arange(n_feat))
ax.set_yticks(np.arange(n_c))
ax.set_xticklabels(range(2, up_bound, 2), fontsize=12)
ax.set_yticklabels([
.001, .005, 0.1, .5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 15, 20, 25, 50, 75, 100, 150, 200, 250
],
fontsize=14)
ax.set_xlabel("Number of features", fontsize=20)
ax.set_ylabel("C value", fontsize=20)
plt.setp(ax.get_xticklabels(), rotation=90, rotation_mode="anchor")
for j in range(0, n_c):
for k in range(0, n_feat):
text = ax.text(k,
j,
'%.2f' % out_kern[j, k],
ha="center",
va="center",
color="w",
fontsize=6)
ax.set_title(
"DNA Methylation Parameter Sensitivity (Accuracy) - %s kernel"
% clf_meth_h.get_params(False)['kernel'],
fontsize=24)
plt.show()
elif make_plot == 'Heat Map CNV':
clf_CNV_h, fea_CNV_h, heat_CNV = do_cv(CNV_train, train_class,
CNV_test, test_class,
'minfo', 'CNV', 100, 2)
n_c = 27
n_feat = 49
up_bound = 100
lin_kern = np.empty(
[n_c,
n_feat]) # c vals ( see in cv_2 ), # feats (n_max - 2 / 2)
poly_kern = np.empty([n_c, n_feat])
rbf_kern = np.empty([n_c, n_feat])
sig_kern = np.empty([n_c, n_feat])
linvec = []
polvec = []
rbfvec = []
sigvec = []
heat_CNV_out = heat_CNV[1:len(heat_CNV), :]
for n, i in enumerate(heat_CNV_out):
if n % 4 == 0:
linvec.append(i[3])
elif n % 4 == 1:
polvec.append(i[3])
elif n % 4 == 2:
rbfvec.append(i[3])
elif n % 4 == 3:
sigvec.append(i[3])
for n in range(0, n_feat):
for c in range(0, n_c):
lin_kern[c][n] = linvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
poly_kern[c][n] = polvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
rbf_kern[c][n] = rbfvec[c + (27 * n)]
for n in range(0, n_feat):
for c in range(0, n_c):
sig_kern[c][n] = sigvec[c + (27 * n)]
ax = plt.subplot(111)
chosen = []
out_kern = None
if clf_CNV_h.get_params(False)['kernel'] == 'linear':
im = plt.imshow(lin_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = linvec
out_kern = lin_kern.copy(True)
elif clf_CNV_h.get_params(False)['kernel'] == 'rbf':
im = plt.imshow(rbf_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = rbfvec
out_kern = rbf_kern.copy(True)
elif clf_CNV_h.get_params(False)['kernel'] == 'sigmoid':
im = plt.imshow(sig_kern, cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = sigvec
out_kern = sig_kern.copy(True)
elif clf_CNV_h.get_params(False)['kernel'] == 'poly':
im = plt.imshow(poly_kern,
cmap=plt.cm.get_cmap('coolwarm', 50))
chosen = polvec
out_kern = poly_kern.copy(True)
plt.colorbar()
plt.clim(min(chosen), max(chosen))
ax.set_xticks(np.arange(n_feat))
ax.set_yticks(np.arange(n_c))
ax.set_xticklabels(range(2, up_bound, 2), fontsize=12)
ax.set_yticklabels([
.001, .005, 0.1, .5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10,
11, 12, 15, 20, 25, 50, 75, 100, 150, 200, 250
],
fontsize=14)
ax.set_xlabel("Number of features", fontsize=20)
ax.set_ylabel("C value", fontsize=20)
plt.setp(ax.get_xticklabels(), rotation=90, rotation_mode="anchor")
for j in range(0, n_c):
for k in range(0, n_feat):
text = ax.text(k,
j,
'%.2f' % out_kern[j, k],
ha="center",
va="center",
color="w",
fontsize=6)
ax.set_title(
"Copy Number Variation Parameter Sensitivity (Accuracy) - %s kernel"
% clf_CNV_h.get_params(False)['kernel'],
fontsize=24)
plt.show()
elif make_plot == None:
pass
if compute == 'custom':
return clf_gene, fea_gene, clf_miRNA, fea_miRNA, clf_meth, fea_meth, clf_CNV, fea_CNV, clf
return tr_score, te_score
