def getMutData(self, category, study, plot, cutoff):
if category == 'ontology':
df = pd.crosstab(self.ontologyDF[category],
self.ontologyDF['mutation_type'])
else:
df = pd.crosstab(self.rawMutationDF[category],
self.rawMutationDF['mutation_type'])
df['sum'] = df.sum(axis=1)
df = getMutationSupertype(df)
df.sort_values(by=['sum'], ascending=False, inplace=True)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/')
df.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/' + category + '_mutation_data.txt')
if plot == True:
#take only the top 50 rows... only applies to gene mutation data but doesn't effect other categories
df = df.head(50).drop('sum', axis=1)
uplot.mutPlot(df, study, category, cutoff=0, weighted=False)
uplot.mutSupertypeSumPlot(df, study, category)
uplot.mutSupertypeIndividualPlot(df, study, category)
def genePlot(df, study, category, subcat, cutoff, weighted):
if weighted == True:
metric = 'normalized mutation count'
m = 'weight'
else:
metric = 'mutation count'
m = ''
plot = df.drop(df.columns[df.apply(lambda col: col.sum() < cutoff)],
axis=1)[:50].plot(kind='bar',
stacked=True,
figsize=(35, 20),
fontsize=15,
color=colorList)
plot.set_title(category + '\n' + subcat + '\n' + 'gene vs ' + metric,
fontsize=15)
plot.set_ylabel(metric, fontsize=15)
plot.set_xlabel('gene', fontsize=15)
plot.legend(loc='center left', bbox_to_anchor=(1, 0.5), prop={'size': 15})
fig = plot.get_figure()
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/' + category + '_gene' + m + '_images/')
fig.savefig('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/' + category + '_gene' + m + '_images/' + subcat +
'_gene' + m + '_image.png',
bbox_inches='tight')
plt.close()
def mutSupertypeSumPlot(df, study, category):
mutSupertypePlot = df.reset_index().plot(
x=category,
y=[
'base_substitution', 'small_indel', 'chromosomal_instability',
'fusion'
],
kind='bar',
stacked=True,
figsize=(25, 15),
title=category + ' vs #mutations',
fontsize=15,
color=['yellow', 'green', 'red', 'pink'])
mutSupertypePlot.set_ylabel('#mutations')
mutSupertypePlot.set_xlabel(category, size=15)
mutSupertypePlot.legend(loc='center left',
bbox_to_anchor=(1, 0.5),
prop={'size': 15})
mutSupertypeFig = mutSupertypePlot.get_figure()
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/')
mutSupertypeFig.savefig('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/' + category +
'MutSupertypeSumData.png',
bbox_inches='tight')
plt.clf()
def mutSupertypeIndividualPlot(df, study, category):
baseSubCols = df.columns[df.columns.isin(baseSubList)]
smallIndelCols = df.columns[df.columns.isin(smallIndelList)]
chromInstabCols = df.columns[df.columns.isin(chromInstabilityList)]
supertypeDict = {
'base_substitution': baseSubCols,
'small_indel': smallIndelCols,
'chromosomal_instability': chromInstabCols
}
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/')
for key, value in supertypeDict.items():
if df[value].empty:
break
else:
df['sum'] = df[value].sum()
df.sort_values(by=['sum'], ascending=False, inplace=True)
df.drop('sum', axis=1, inplace=True)
supertypePlot = df.reset_index().plot(x=category,
y=value,
kind='bar',
stacked=True,
figsize=(25, 15),
title=category + ' vs #' + key +
' mutations',
fontsize=15)
supertypePlot.set_ylabel('#mutations', size=15)
supertypePlot.set_xlabel(category, size=15)
supertypePlot.legend(loc='center left',
bbox_to_anchor=(1, 0.5),
prop={'size': 15})
supertypeFig = supertypePlot.get_figure()
supertypeFig.savefig('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/' + category + '_' + key +
'.png',
bbox_inches='tight')
plt.clf()
def assimilateRawMutationDF(self, study):
self.rawMutationDF = self.pointDF.append(self.cnaDF, ignore_index=True)
self.rawMutationDF = self.rawMutationDF[[
'gene_symbol', 'pointORcna', 'mutation_type', 'amino_acid_change',
'functional_impact_score', 'sample_id', 'study_id', 'ontology'
]]
self.rawMutationDF.sort_values('gene_symbol', inplace=True)
self.rawMutationDF = combineRepeatMutations(self.rawMutationDF)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/')
def getGeneData(self, category, study, plot, cutoff):
if category == 'ontology':
subcategoryList = self.ontologyDF[category].unique()
else:
subcategoryList = self.rawMutationDF[category].unique()
for subcat in subcategoryList:
if category == 'ontology':
df = self.ontologyDF.loc[self.ontologyDF[category] == subcat]
else:
df = self.rawMutationDF.loc[self.rawMutationDF[category] ==
subcat]
geneDF = pd.crosstab(df['gene_symbol'], df['mutation_type'])
geneDF['sum'] = geneDF.sum(axis=1)
geneDF.sort_values(by='sum', ascending=False, inplace=True)
geneWeight = df.groupby('gene_symbol').mutation_type.value_counts(
normalize=True)
geneWeightDF = geneWeight.unstack().fillna(0.0)
geneWeightDF = geneWeightDF.reindex(geneDF.index)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/')
geneDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/' + category + '_gene_data.txt')
geneWeightDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/' + category +
'_gene_weight_data.txt')
if plot == True:
geneDF.drop('sum', axis=1, inplace=True)
uplot.genePlot(geneDF,
study,
category,
subcat,
cutoff,
weighted=False)
uplot.genePlot(geneWeightDF,
study,
category,
subcat,
cutoff,
weighted=True)
def groupByOntology(self, study):
self.ontologyDF = self.rawMutationDF
#drops and resets index
self.ontologyDF = self.ontologyDF.reset_index(drop=True)
#maps a list of ontologies to each row based on the gene_symbol
self.ontologyDF['ontology'] = self.ontologyDF['gene_symbol'].map(
ontologyDict)
#if any rows do not have a value for 'ontology' set that value to a list containing 'no_ontology_found'
self.ontologyDF.loc[self.ontologyDF['ontology'].isnull(),
['ontology']] = self.ontologyDF.loc[
self.ontologyDF['ontology'].isnull(),
'ontology'].apply(
lambda x: ['no_ontology_found'])
#get number of ontologies for each row
dfLength = self.ontologyDF['ontology'].str.len()
#convert ontology list to numpy array
dfValues = self.ontologyDF['ontology'].values
#get an array with the indices repeated by the number of ontologies for each row
dfIndex = np.repeat(self.ontologyDF.index, dfLength)
#expand rows by duplicated index values
self.ontologyDF = self.ontologyDF.loc[dfIndex]
#add ontology values, flattened from the list
self.ontologyDF['ontology'] = np.concatenate(dfValues)
#reset the index again, since it currently contains repeat values
self.ontologyDF = self.ontologyDF.reset_index(drop=True)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/')
self.ontologyDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study +
'/mappedDF.txt',
index=False)
def getOntologyData(self, category, study,
plot): #DEPRECATED... NO LONGER IN USE (TENTATIVE)
for subcat in self.rawMutationDF[category].unique():
df = self.ontologyDF.loc[self.ontologyDF[category] == subcat]
ontologyDF = pd.crosstab(df['ontology'], df['mutation_type'])
ontologyDF['sum'] = ontologyDF.sum(axis=1)
ontologyDF.sort_values(by='sum', ascending=False, inplace=True)
ontologyGrouped = df.groupby('ontology')
ontologyWeight = ontologyGrouped.mutation_type.value_counts(
) / ontologyGrouped.gene_symbol.nunique()
ontologyWeightDF = ontologyWeight.unstack().fillna(0.0)
ontologyWeightDF['sum'] = ontologyWeightDF.sum(axis=1)
ontologyWeightDF.sort_values(by='sum',
ascending=False,
inplace=True)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/')
ontologyDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/' + category +
'_ontology_data.txt')
ontologyWeightDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' +
study + '/' + category + '/' + category +
'_ontology_weight_data.txt')
if plot == True:
ontologyDF.drop('sum', axis=1, inplace=True)
ontologyWeightDF.drop('sum', axis=1, inplace=True)
uplot.ontologyPlot(ontologyDF, study, category, subcat)
uplot.ontologyWeightPlot(ontologyWeightDF, study, category,
subcat)
def getMutWeightData(self, category, study, plot, cutoff):
#get a groupby object containing all the mutations grouped by the category
#then gets the avg mutations per sample or gene
if category == 'ontology':
grouped = self.ontologyDF.groupby(category)
mutWeight = grouped.mutation_type.value_counts(
) / grouped.gene_symbol.nunique()
else:
grouped = self.rawMutationDF.groupby(category)
mutWeight = grouped.mutation_type.value_counts(
) / grouped.sample_id.nunique()
#unstacks groupby object and converts to dataframe to plot
#fills NaN values with 0... NaN values here mean that the category did not have that type of mutation
mutWeightDF = mutWeight.unstack().fillna(0.0)
mutWeightDF['sum'] = mutWeightDF.sum(axis=1)
mutWeightDF.sort_values(by='sum', ascending=False, inplace=True)
ensureDirectory('/Users/Rohil/Documents/Young Dawgs/' + study + '/' +
category + '/')
mutWeightDF.to_csv('/Users/Rohil/Documents/Young Dawgs/' + study +
'/' + category + '/' + category +
'_mut_weight_data.txt')
if plot == True:
mutWeightDF.drop('sum', axis=1, inplace=True)
uplot.mutPlot(mutWeightDF, study, category, cutoff, weighted=True)