def create_info(self):
"""
Creates several info dataframe for the input data.
self.peaks_per_feature .... top peaks per feature for features that
are hit by top peaks
self.top_peaks ... top peaks with the features that are close
"""
value_s = self.value_s.sort(ascending=self.ascending, inplace=False)
top_s = value_s.iloc[: self.top_n]
self.candidate_features = hp.get_features(
top_s, self.feature_df, feature_name=self.feature_name, max_dist=self.max_dist
)
sub_feature_df = (
self.feature_df.reset_index()
.set_index(self.feature_name)
.ix[self.candidate_features]
.reset_index()
.set_index(["chrom", "start"])
)
self.peaks_per_feature = get_peaks(sub_feature_df, top_s, self.max_dist, feature_name=self.feature_name)
features_sort_by_max = self.peaks_per_feature["peak_height"].groupby(lambda i: i[0]).max()
features_sort_by_max.sort(ascending=False, inplace=True)
self.peaks_per_feature = self.peaks_per_feature.ix[features_sort_by_max.index]
self.top_peaks = self.get_peak_info(top_s, self.peaks_per_feature)
# super(SummaryEnrichment, self).create_info()
CandidateEnrichment.create_info(self)
def get_association(self, value_s):
value_s = value_s.sort(ascending=self.ascending, inplace=False)
top_s = value_s.iloc[:self.top_n]
del value_s
candidate_features = hp.get_features(top_s, self.feature_df, feature_name=self.feature_name,
max_dist=self.max_dist)
assoc = CandidateEnrichment.get_association(self, candidate_features)
#assoc = self.feature_to_category.set_index(self.feature_name).ix[cand_genes].groupby(self.category_name).apply(len)
#assoc.name = "n_" + self.feature_name
#assoc.index.name = self.category_name
return assoc