def extract_features(df):
df_n = df.xs('01', level=1, axis=1)
binary = df_n > -1
binary = binary[binary.sum(1).isin(range(20, df.shape[1]/2))]
rr = df.ix[binary.index].apply(exp_change, 1)
binary = binary.ix[true_index(rr.p < .05)]
real = df_n.ix[df_n.index.diff(binary.index)]
singles = real[((real.max(1) - real.min(1)) > 1)]
singles = singles[(singles.std(1) > .25)]
ch = df.ix[singles.index].apply(exp_change, 1)
singles = df_n.ix[true_index(ch.p < .01)]
return binary, singles, real
def extract_features(df):
df_n = df.xs('01', level=1, axis=1)
binary = df_n > -1
binary = binary[binary.sum(1).isin(range(20, df.shape[1] / 2))]
rr = df.ix[binary.index].apply(exp_change, 1)
binary = binary.ix[true_index(rr.p < .05)]
real = df_n.ix[df_n.index.diff(binary.index)]
singles = real[((real.max(1) - real.min(1)) > 1)]
singles = singles[(singles.std(1) > .25)]
ch = df.ix[singles.index].apply(exp_change, 1)
singles = df_n.ix[true_index(ch.p < .01)]
return binary, singles, real
def extract_pc_filtered(df, pc_threshold=.2, filter_down=True):
'''
First pre-filters for patients with no tumor/normal change.
Then normalizes by normals.
'''
if ('11' in df.columns.levels[1]) and filter_down:
tt = df.xs('11', axis=1, level=1)
rr = df.apply(exp_change, 1).sort('p')
m, s = tt.mean(1), tt.std(1)
df_n = df.xs('01', axis=1, level=1)
df_n = ((df_n.T - m) / s).T
df_n = df_n.ix[true_index(rr.p < .05)]
else: #No matched normals
df_n = df.xs('01', axis=1, level=1)
df_n = ((df_n.T - df_n.mean(1)) / df_n.std(1)).T
pc = extract_pc(df_n, pc_threshold, standardize=False)
return pc
def rna_filter(cn, val, rna):
'''
Filter copy number events with rna expression data.
Here we test whether the event is associated with a subsequent
change in expression in those patients.
cn: copy number matrix, should have a MultiIndex, with the gene name
in the last level
val: value of the copy number to test in [-2, -1, 1, 2]
'''
assert val in [-2, -1, 1, 2]
change = pd.DataFrame({g: kruskal_pandas(vec == val, rna.ix[g[-1]])
for g, vec in cn.iterrows()
if g[-1] in rna.index}).T
q_vals = bhCorrection(change.p)
filtered = cn.ix[true_index(q_vals < .1)]
return filtered
def extract_pc_filtered(df, pc_threshold=.2, filter_down=True):
'''
First pre-filters for patients with no tumor/normal change.
Then normalizes by normals.
'''
if ('11' in df.columns.levels[1]) and filter_down:
tt = df.xs('11', axis=1, level=1)
rr = df.apply(exp_change, 1).sort('p')
m, s = tt.mean(1), tt.std(1)
df_n = df.xs('01', axis=1, level=1)
df_n = ((df_n.T - m) / s).T
df_n = df_n.ix[true_index(rr.p < .05)]
else: # No matched normals
df_n = df.xs('01', axis=1, level=1)
df_n = ((df_n.T - df_n.mean(1)) / df_n.std(1)).T
pc = extract_pc(df_n, pc_threshold, standardize=False)
return pc
def rna_filter(cn, val, rna):
'''
Filter copy number events with rna expression data.
Here we test whether the event is associated with a subsequent
change in expression in those patients.
cn: copy number matrix, should have a MultiIndex, with the gene name
in the last level
val: value of the copy number to test in [-2, -1, 1, 2]
'''
assert val in [-2, -1, 1, 2]
change = pd.DataFrame({
g: kruskal_pandas(vec == val, rna.ix[g[-1]])
for g, vec in cn.iterrows() if g[-1] in rna.index
}).T
q_vals = bhCorrection(change.p)
filtered = cn.ix[true_index(q_vals < .1)]
return filtered
